In August 2026 I asked the same AI model (claude-opus-5) eight carbon removal research questions in four setups: with my plugin connected, and without. A judge scored each answer against a checklist of checkable facts, never knowing which setup wrote it. The full run is on this page, with the prompts, answers, and verdicts unedited.
Connecting the database recovered nearly every checkable fact, and the connected setups kept searching the web just as much as the baseline (7.5 vs 7.5 findings beyond the database per question).
On its first run, the database-only control failed the Frontier funding question, scoring 2 of 5: my query tools could not reach a data layer that existed. I fixed the tools and re-ran the question: 5 of 5. Both results are preserved in the scoreboard below.
Each opens to what web search alone missed, then every setup’s full answer and verdict.
Asked, verbatim: “What offtake deals exist for pumpable biomass carbon removal (bio-oil or slurry injection), and what do they tell us about pricing?”
Judge’s note — Three of four gold deals are captured with correct volumes, values and implied prices, and the analysis of the Frontier–Charm price schedule (37–75% decline, ~$600 spot, milestone conditions) is exactly the kind of interpretation an expert wants. Structure is excellent — deal table, explicit implied $/t, benchmarking against market averages, and unusually candid caveats about inference vs. reporting, unverified recall, and taxonomy errors in secondary sources (coverage honesty is a standout). The major flaw is the omission of the Frontier–NULIFE GreenTech deal (Dec 2025, 122,000 t, $44.2M, ~$362/t). This is not merely a missing row: the answer builds an entire argument on the claim that 'only two deals in this entire category have a public price,' and that the category is essentially Charm + Vaulted. A third priced, third-party-verifiable deal at a lower price point would materially strengthen (and slightly reframe) the pricing trajectory argument, so the omission propagates into a false structural claim. Beyond-corpus material is substantial: Charm–Google biochar (~100k t, Jan 2025), Charm–TD Bank (44,000 t, Jan 2026), Vaulted's >5,000 t first-year delivery, Charm's ~$600/t spot price and $100/t-by-2040 target, S&P biochar credit pricing (~$150/t), the $320/t weighted-average durable CDR price, CDR.fyi/OPIS bid-ask spread data, and the CO280/JPMorgan sub-$200/t comparator. Sources are named (Frontier, Canary Media, S&P Global, CDR.fyi/OPIS) though not linked, so claims are checkable but require legwork. Net: strong, well-reasoned, honest, but incomplete on the deal record in a way that undercuts a headline assertion.
I'll research this across the main players in this category — bio-oil injection and slurry injection — and pull together the deal record. ## The category, and who's actually in it "Pumpable" biomass CDR covers two distinct approaches that share one economic insight: if you can turn biomass into something that flows through a pump, you can use existing oilfield/waste-injection infrastructure instead of building new plants. - **Charm Industrial** — fast pyrolysis of ag/forestry residues into bio-oil, injected into disposal wells. It produces roughly 0.2 tons of biochar and 0.5 tons of bio-oil per ton of biomass processed , and converts residues into two outputs, with bio-oil injected underground and biochar applied to soil . - **Vaulted Deep** — takes wet organic waste and slurry-injects it. Its feedstock is sludgy waste like biosolids, manure, food and ag waste, and papermill sludge — material typically incinerated or land-applied . It spun out of injection-well waste management company Advantek in 2023 . These two account for essentially all disclosed offtake volume in the category. ## The deal record | Deal | Announced | Volume | Value | Delivery window | Implied $/t | |---|---|---|---|---|---| | Charm ← Frontier | May 2023 | 112,000 t | $53M | 2024–2030 | **~$473** | | Vaulted ← Frontier | May 2024 | 152,480 t | $58.3M | 2024–2027 | **~$382** | | Charm ← Google (biochar) | Jan 2025 | ~100,000 t | undisclosed | n/d | n/d | | Vaulted ← Microsoft | Jul 2025 | up to 4.9 Mt | undisclosed | through 2038 | n/d | | Charm ← TD Bank | Jan 2026 | 44,000 t | undisclosed | 10 yrs from 2029 | n/d | **Charm/Frontier (May 2023).** Frontier's first offtake agreements totaled $53M for 112,000 tons removed between 2024 and 2030, on behalf of Stripe, Alphabet, Shopify, Meta, McKinsey, plus Autodesk, H&M Group and Workday , with Aledade, Boom Supersonic, Canva, SKIMS, Wise and Zendesk contributing via Watershed . **Vaulted/Frontier (May 2024).** A $58.3M offtake for 152,480 tonnes between 2024 and 2027, of which 18,321 tonnes were expected in 2024 . Vaulted delivered its first batch against the agreement within four months of signing, issuing over 5,000 tonnes in its first year of operation . **Vaulted/Microsoft (July 2025).** Up to 4.9 million tonnes over 12 years, running through 2038 — described at the time as the second-largest carbon removal agreement to date . **Charm/TD Bank (January 2026).** A 10-year agreement for 44,000 tonnes, delivered through both biochar removals and bio-oil sequestration, with the term beginning in 2029 — Charm's first step into Canada . ## What the pricing actually tells us **1. Only two deals in this entire category have a public price.** Everything else — including the largest deal by a factor of 30 — is undisclosed. Any statement about "the price of pumpable BiCRS" rests on two data points from 2023 and 2024. **2. Headline $/t is a contract average across a steeply declining curve, not a spot price.** Frontier is explicit: buyers pay a price per ton that declines by at least 37% between 2024 and 2030, and could decline by as much as 75% based on Charm's scaling plans and potential expanded government incentives, with the price covering both removal and MRV . Charm's spot price at the time was around $600 per ton , and Canary Media reported that buyers could pay roughly $470 per ton on average over the contract — conditional on federal incentives and Charm's scaling plans materializing . So the $473/t figure is best read as a path from roughly $600 in 2024 toward roughly $380 by 2030 (my arithmetic, not a disclosed schedule) — and the low end of that range only holds if 45Q-type incentives arrive. **3. Slurry priced ~19% below bio-oil, one year later.** $382/t vs $473/t. That is a meaningful spread, and the stated cost trajectories diverge much further: Frontier says Vaulted's approach offers 10,000+ year permanence, near-term scalability, local co-benefits and a line of sight to under $100 cost per ton , and Vaulted has argued it has a "believable path" to driving costs well under $100 per ton . Charm's stated trajectory is slower: $600 per tonne with an expectation of $100 per tonne by 2040 . The economic reason is structural — slurry injection skips the energy-intensive pyrolysis step and earns tipping-fee revenue from waste generators, so the CDR credit only has to cover part of the cost stack. **4. Contracted price ≠ revenue.** Frontier's Charm agreement includes milestones that must be met prior to delivery, including pyrolyzer performance and community engagement . These are offtakes with conditions, not guaranteed cash. **5. The product mix is quietly shifting, which muddies the price signal.** Charm has moved biochar from byproduct to co-product — the company describes biochar as now playing "a key supporting role" in its production model , and the TD deal is explicitly a bio-oil/biochar blend. That matters because biochar is priced in a different market: S&P Global found U.S. biochar credits at around $150/tonne for 2025 delivery in October 2025, with 2026 delivery near $148 . A blended bio-oil/biochar contract should therefore price below a pure bio-oil contract, and none of the disclosed blended deals reveal the mix. **6. Benchmarking.** Against a market where the weighted average durable CDR price was $320/tonne at end-2024, down from $490 in 2023 , both priced deals sit above the market average but far below DAC. And the buyer/supplier gap is closing: CDR.fyi and OPIS found the average price spread between durable CDR buyers and suppliers fell from $107/tonne to $98/tonne, with respondents expecting $48/tonne by 2030 . ## The Microsoft deal is the biggest gap No price was disclosed for the 4.9 Mt Vaulted agreement. Simple arithmetic frames the range: at Frontier's $382/t it would be ~$1.9bn; at $200/t, ~$980M; at $100/t, ~$490M. Given the volume and Microsoft's portfolio behavior, the realized price is almost certainly far below the Frontier level — but I want to be clear that **this is inference, not reporting**. For calibration, JPMorgan's CO280 pulp-mill deal was priced under $200 per ton, reported as one of the lowest-ever for an engineered solution . Note also that it is "up to" 4.9 Mt — an option-like ceiling contingent on Vaulted building sites, not a firm commitment. ## Caveats and what I couldn't verify - **Sample size.** Two priced deals, both facilitated by the same buyer (Frontier), both partly funded by the same underlying founding members. This is not a market price; it's a bilateral price from an advance-market-commitment buyer explicitly paying above cost to build supply. - **Delivery vs. contract.** Vaulted's early delivery record is documented above; I was not able to verify in this session how Charm's cumulative deliveries track against its 112,000-ton Frontier schedule, and its shift toward biochar suggests bio-oil scale-up has been harder than modeled. Treat Charm's remaining contracted volume as uncertain. - **Sources conflict on taxonomy.** At least one outlet describes Vaulted Deep as a "bio-oil sequestration company" in JPMorgan's portfolio — that's a mislabel; Vaulted injects slurried waste, not pyrolysis oil. Be careful with secondary databases in this category. - **Unverified in this session** (my recall, flagged as such): Stripe's original 2020 Charm purchase at ~$600/t, JPMorgan's 2022 Charm purchase, Microsoft's earlier Charm purchases, and the status of any U.S. DOE CDR Purchase Pilot Prize awards to either company. I hit my search limit before confirming these — don't rely on them without checking. - **Policy dependence.** Charm's declining price schedule is explicitly conditioned on government incentives. If 45Q eligibility for bio-oil injection or comparable support doesn't materialize, the realized price path will land at the top of the disclosed range, not the bottom.
Judge’s note — All four gold checklist deals are captured with correct volumes, values and derived $/t, presented in a clean comparative table with feedstock/storage/window detail. The answer goes well beyond the corpus: JPMorganChase–Charm (June 2026, 61,500 t plus $20M venture debt, cumulative 90,000 t), TD Bank–Charm (Jan 2026, 44,000 t blended bio-oil/biochar from 2029), actual delivery tallies from Frontier's portfolio page (~100.6k/112k Charm, ~52.1k/154k Vaulted) benchmarked against CDR.fyi's ~3% market delivery rate, Frontier's disclosure that the Charm contract price declines 37–75% across 2024–2030, Frontier's 'line of sight to under $100/ton' comment on Vaulted, the 2026 CDR.fyi×OPIS buyer–supplier gap survey, Frontier's >$360M across eight BiCRS companies, plus Isometric's CORSIA approval and the DACCS/BECCS-only scoping of EU Art. 9c and BSI Flex — a genuinely useful policy-floor observation. Analysis is sharp rather than a list: it correctly refuses to read a learning curve from three single-buyer deals, flags that ~93% of contracted tonnes (Microsoft–Vaulted) carry no disclosed price, notes bundled financing/product-blending distorting $/t, and labels the '~$1B' figure as speculation. Sources are named and checkable (Charm blog, ESG Dive, Frontier portfolio page, CDR.fyi, ICAO) though without links or precise dates in places, and it also correctly disqualifies the Google–Charm biochar deal as a common miscount. Honesty is exemplary: derived-not-quoted prices, thin evidence base (one paper), corpus threshold effects, and an explicit statement that post-Aug-2026 deals could not be checked because search budget ran out. Minor deductions only for citation granularity and the inherent unverifiability of the two press-reported 2026 bank deals as presented.
I'll start by checking what the corpus covers, then dig into the specific pathway. Now let me pull the supplier entities and search for deals the pathway profile may have missed. Let me check the corpus for related media/analysis and recent changes, and search the web for what the corpus can't see. ## The short answer There are only four publicly recorded offtakes in this pathway that I can verify, and three of them come from a single buyer — Frontier — which is the only party that publishes deal values. Everything we "know" about pricing for pumpable biomass therefore comes from ~7% of the contracted tonnes. ## The deals **In the Carbon Curve corpus (capture pathway `biomass-injection`, 4 deals, 5,286,480 t contracted):** | Announced | Buyer | Supplier | Tonnes | Value | Derived $/t | Window | Form | |---|---|---|---|---|---|---|---| | 2023-05-18 | Frontier | Charm Industrial | 112,000 | $53.0M | **~$473** | 2024–2030 | Bio-oil (fast pyrolysis), EPA-regulated wells | | 2024-05-01 | Frontier | Vaulted Deep | 152,480 | $58.3M | **~$382** | 2024–2027 | Organic-waste slurry, Class V wells (Kansas, LA) | | 2025-07-17 | Microsoft | Vaulted Deep | up to 4,900,000 | **undisclosed** | — | 12 yrs, through 2038 | Slurry (biosolids, manure, paper sludge, food/ag residues) | | 2025-12-18 | Frontier | NULIFE GreenTech | 122,000 | $44.2M | **~$362** | 2026–2030 | Bio-oil via hydrothermal liquefaction, salt caverns >1,000 m, Saskatoon; Isometric-verified | All four are firm commitments classed as durable-1000y with geological storage. Note the corpus flags the per-tonne figures as *derived* (value ÷ tonnes), not quoted prices. **Found on the web, not in the corpus** (mostly because they fall under the corpus's ≥50kt / ≥$10M non-Frontier inclusion threshold): - **JPMorganChase ↔ Charm**, announced 4 June 2026: 61,500 tons of bio-oil removals over multiple years, plus a separate **$20M venture debt facility** for Charm's Colorado expansion. It follows a 28,500-ton 2023 purchase, bringing JPMorgan's total to 90,000 tons (Charm press release; ESG Dive, 5 June 2026). No price disclosed. - **TD Bank ↔ Charm**, announced 21–23 Jan 2026: 44,000 tonnes over 10 years starting 2029, spanning **both bio-oil and biochar**, tied to Charm's Canadian expansion (Charm blog; ESG Dive). No price disclosed. - **Frontier prepurchase round, Sept 2023** ($7M across 12 recipients) included Vaulted Deep — a cohort funding commitment, not an offtake. - Google's 100kt Charm deal (Jan 2025) is **biochar**, not injection — a common miscount when people tally this pathway. ## What the prices actually tell us **1. The disclosed band is $362–473/tonne, volume-weighted ~$402/t.** That places pumpable biomass roughly an order of magnitude above biochar and well below DAC — consistent with its pitch: geological-grade permanence without the energy penalty of air capture. **2. The apparent 24% decline (2023 → 2025) is softer evidence than it looks.** Three deals, one buyer, different suppliers, geographies, feedstocks and delivery windows. Frontier's Charm page states the contract price *within* that deal declines by **at least 37% and possibly 75% between 2024 and 2030** — so $473 is a portfolio average across a seven-year ramp, not a 2024 spot price. The 2024 price was higher and the 2030 price materially lower. Comparing it to NULIFE's $362 average over 2026–2030 conflates two different points on two different learning curves. **3. The price that matters most is the one nobody published.** Microsoft's up-to-4.9 Mt Vaulted Deep contract is 93% of contracted tonnes in this pathway and carries no disclosed value. The corpus explicitly labels the circulating "~$1B" third-party estimates as **speculation**. Frontier's own 2024 Vaulted write-up said the company had "line of sight to under $100 cost per ton" — if that is anywhere near right, the at-scale clearing price is a different market from the $362–473 Frontier band, and we are reasoning about premium first-of-a-kind tonnes. **4. Deal structure is becoming part of the price.** JPMorgan's June 2026 deal bundles an offtake with $20M of venture debt; TD Bank contracted a blended bio-oil/biochar portfolio with deliveries starting 2029. Both make a clean $/t read impossible even if values were published — cost of capital and product blending are being negotiated alongside price. **5. Delivery track record is unusually good for durable CDR, which supports the pricing.** Frontier's portfolio page shows Charm at ~100.6k of 112,003 tons delivered and Vaulted at ~52,100 of 154,146 as of Aug 2026. Against CDR.fyi's estimate that the market-wide delivery rate is roughly 3% (13% excluding Microsoft and Frontier), these are genuinely delivering contracts, not paper ones — Charm booked the field's first verified closed-system tons in 2023 and Vaulted in 2024. **6. Context from the market surveys** (flagged as uncorroborated source claims in the corpus): the 2026 CDR.fyi × OPIS pricing survey found the buyer–supplier price gap across durable CDR narrowed from $107 to $98/tonne, with respondents expecting ~$48 by 2030 and judging that durable CDR is unlikely to broadly reach $100/tonne by 2030. Frontier's own August 2026 BiCRS post says its members have committed **over $360M in offtakes with eight BiCRS companies**. ## The policy vacuum behind the price No tracked policy instrument names this pathway. That matters for pricing: the EU ETS Article 9c proposal (17 July 2026) would centrally purchase 250 Mt of CRCF-certified removals over 2031–2040 but is scoped to **DACCS and BioCCS**; the UK's interim quantification standards are **BSI Flex 2006 (BECCS) and 2007 (DACCS)**. Bio-oil and slurry injection sit outside both. So there is no compliance floor, no strike price, and no obvious CfD — every tonne above is voluntary demand. The nearest supports are US Class V/Class VI well permitting and **Isometric**, which became the first durable-CDR-only registry approved for CORSIA units (ICAO, 26 Nov 2025) and is the registry for both the NULIFE and Microsoft–Vaulted contracts. ## Caveats worth being explicit about - **Four deals is a thin base.** The corpus is a curated floor, not a census; CDR.fyi is the fuller transaction scoreboard, and smaller offtakes below the threshold are invisible here. - **All per-tonne figures are derived**, and Frontier's deals may embed milestone structures, uncertainty discounts (Frontier applies "Verification Confidence Levels" and an explicit uncertainty discount to delivered tons) and MRV costs that are not separable. - **Evidence base is nearly empty**: the corpus holds exactly **one** classified paper on this pathway (hydrochar subsurface storage in the Netherlands, *IJGGC*, Dec 2025). Pricing is running well ahead of published literature. - I could not verify Vaulted's current cost claims or check for deals after ~August 2026 — corpus freshness ends 2026-08-15 and my web search budget was exhausted mid-check. Treat the JPMorgan/TD figures as press-reported, and assume there may be 2026 deals neither source captured.
Judge’s note — All four gold deals are surfaced with correct volumes, values, and derived $/t, plus correct dates and suppliers' technology distinctions (fast pyrolysis vs HTL vs waste slurry). The answer goes well beyond the deal-record corpus: Charm and Vaulted delivered-tonnage figures, the sub-threshold corporate buyer tier (JPMorganChase 61,500 t plus $20M debt facility, Boeing 100 kt, TD Bank 44 kt 10-year, Capgemini), Vaulted's 2025 throughput/delivery figures and $32.3M Series A, Frontier's 37–75% price-decline language, Canary Media's ~$470/t commentary, the CDR.fyi×OPIS buyer–supplier gap narrowing, and Frontier's $360M/eight-BiCRS-company claim. The pricing interpretation is genuinely expert: it refuses to read $473→$382→$362 as a market curve, correctly identifies contract-average vs spot, the declining intra-contract ramp, and the structural reason Vaulted prices lower (gate fees / waste-management revenue rather than learning curve). Honesty is the strongest dimension — it quantifies that 92.7% of pathway tonnes are price-dark, flags derived-not-quoted prices, flags the 'up to' framing on Microsoft's 4.9 Mt and dismisses third-party ~$1B estimates as speculation, notes pathway-boundary blurring (Google/Charm classified biochar; TD deal spans biochar+bio-oil), and openly states the web-search quota cut off verification of post-August-2026 announcements. Deductions: sources are named but not linked/dated precisely enough to check quickly (Carbon Herald, 'corpus-adjacent CDR.fyi analysis'), a couple of the sub-threshold deals rest on unverifiable single mentions, and there is an internal arithmetic inconsistency ('scale gap is 32:1' vs '12.7× the combined tonnage'). Leading process narration is minor noise. Overall a near-complete, decision-useful answer.
I'll start by understanding the corpus coverage, then dig into the pathway. Now let me join the entity records and check for anything the pathway filter might have missed. Now the required step — searching the web for what the corpus says it cannot cover: deals below thresholds, non-Frontier buyers, and anything after the 2026-08-15 ingest. Web search quota is exhausted for this session; let me pull the remaining corpus detail on NULIFE and recent changes. ## Short answer The corpus holds **four offtakes** in the `biomass-injection` pathway (pumpable bio-oil or organic-waste slurry to geological storage), totalling **5,286,480 tonnes contracted**, of which **three disclose a value**. Every disclosed price comes from a single buyer — Frontier — and covers just **7.3% of the pathway's contracted tonnes**. The other 92.7% is one undisclosed Microsoft contract. So the pathway has a *price series*, but only for the buyer that publishes. --- ## 1. The deals in the database | Announced | Buyer | Supplier | Tonnes | Value | Derived $/t | Form | Delivery | |---|---|---|---|---|---|---|---| | 2023-05-18 | Frontier | Charm Industrial (US) | 112,000 | $53.0M | **$473** | bio-oil (fast pyrolysis) | 2024–2030 | | 2024-05-01 | Frontier | Vaulted Deep (US) | 152,480 | $58.3M | **$382** | organic-waste slurry | 2024–2027 | | 2025-07-17 | Microsoft | Vaulted Deep (US) | up to 4,900,000 | undisclosed | — | slurry | 12 yrs, to 2038 | | 2025-12-18 | Frontier | NULIFE GreenTech (CA) | 122,000 | $44.2M | **$362** | bio-oil (hydrothermal liquefaction) | 2026–2030 | Detail worth keeping: - **Charm** — Frontier's *first-ever* offtake. Bio-oil from fast pyrolysis of waste biomass, injected into EPA-regulated wells. Frontier's portfolio page lists 112,003 contracted tons with ~100.6k delivered as of Aug 2026 — by far the most complete delivery record in the pathway. Frontier's own page states the price per ton will decline by at least 37% between 2024 and 2030, and could decline by as much as 75% based on Charm's scaling plans and potential expanded government incentives . Canary Media reported buyers could pay roughly $470 per ton on average over the contract if federal incentives and Charm's scaling plans materialise . - **Vaulted Deep** — spinout of Advantek Waste Management; injects carbon-rich organic waste (biosolids, manure, paper sludge, food/ag residues) into Class V wells in Kansas and Los Angeles. Frontier stated the approach offers 10K+ year permanence, near-term scalability, local co-benefits and a line of sight to under $100 cost per ton . Portfolio: 154,146 contracted tons, ~52,100 delivered as of Aug 2026. - **Microsoft ↔ Vaulted** — up to 4.9 million tonnes of durable removal over 12 years, running through 2038 , with Isometric selected as the registry . The corpus flags the "up to" explicitly and notes that **third-party ~$1B estimates are speculation**, not disclosure. - **NULIFE** — hydrothermal liquefaction of wet and dry waste biomass into bio-oil, injected into licensed salt caverns >1,000 m near Saskatoon; Isometric-verified. It follows a 2024 Frontier prepurchase, more than half delivered. All four records are `review_status: approved`. **Volume-weighted average of the three priced deals: $402/t** ($155.5M ÷ 386,480 t). --- ## 2. Deals the corpus structurally cannot hold — found by search The corpus only ingests Frontier plus a handful of mega-buyers at ≥50kt or ≥$10M. Searching for what it excludes surfaced a materially larger commercial base for Charm in particular: - **JPMorganChase ↔ Charm (June 2026)**: JPMorganChase will purchase 61,500 tons of carbon removal credits from Charm over multiple years, bringing its total purchases from Charm to 90,000 tCO₂e . The bank previously purchased 28,500 tons in 2023; the price per ton was not disclosed . Paired with a $20 million debt facility to expand Charm's Colorado facility . - **Boeing ↔ Charm (Nov 2025)**: a deal to remove 100,000 metric tons, with financial terms not disclosed . - **TD Bank ↔ Charm (Jan 2026)**: a 10-year offtake for 44,000 metric tons of biochar and bio-oil removal credits, marking Charm's expansion into Canada , with deliveries beginning in 2029 . - **Capgemini** has also signed with Charm (alongside Climeworks), per Carbon Herald. - **Vaulted operations, 2025**: processed more than 90,000 tonnes of organic waste at its Great Plains facility and delivered more than 26,000 tonnes , and launched a methane quantification initiative with Google and Isometric . It raised $32.3M in Series A funding led by Prelude Ventures in Nov 2024 . None of these disclose a price. That is the finding: **the sub-threshold, corporate-buyer tier of this pathway is entirely price-dark.** --- ## 3. What the prices actually tell us — and what they don't **The trend line, read literally:** $473 (2023) → $382 (2024) → $362 (2025), a 23% decline across three Frontier contracts in 31 months. I would not present that as a market price curve, for four reasons: 1. **One buyer.** All three are Frontier. Corpus-wide, 21 of 59 deals disclose value and those are 4.5% of contracted tonnes; the concentration is worse in this pathway. The two largest buyers here by volume (Microsoft, Boeing) published nothing. 2. **These are contract averages over multi-year windows, not spot prices.** The corpus flags all three as `price_derivation: usd_amount/tonnes, derived not quoted`. Charm's contract price *falls 37–75% within its own delivery window* — so $473 is the midpoint of a declining ramp, not a 2023 clearing price. Comparing it to NULIFE's $362 (a 2026–2030 window) partly compares contract shapes, not cost levels. 3. **The suppliers are not doing the same thing.** Charm and NULIFE convert biomass thermochemically (pyrolysis vs. hydrothermal liquefaction) and pay for feedstock and conversion. Vaulted injects waste that generators *pay it* to take — a waste-management revenue line that pyrolysis routes don't have. That, not learning-curve progress, is the most plausible reason Vaulted priced $91/t below Charm a year later and could credibly claim line of sight below $100/t. It is also the most plausible reason Microsoft could contract 4.9 Mt from Vaulted and no one else. 4. **The scale gap is 32:1.** Microsoft's single undisclosed contract is 12.7× the combined tonnage of all three priced deals. Any pathway-level average price you compute from the corpus describes 7.3% of the tonnes. **Delivery, as a price check.** Charm has delivered ~100.6k of 112,003 tons and Vaulted ~52.1k of 154,146 — genuinely unusual in durable CDR, where the corpus-adjacent CDR.fyi analysis puts the overall delivery rate at low single digits. A price paid against tonnes actually delivered is worth more than a headline price against a 2038 promise. --- ## 4. Context, flagged as uncorroborated These are **signals** in the corpus — source claims, not verified records: - Frontier's Aug 2026 BiCRS post claims Frontier members have committed over $360M in offtake agreements with eight BiCRS companies , which would put the three deals above (~$155.5M) at under half of Frontier's BiCRS book. - The 2026 CDR.fyi × OPIS pricing survey reports the average buyer–supplier price gap narrowing from $107/t to $98/t, with respondents expecting ~$48/t by 2030 and judging durable CDR unlikely to broadly reach $100/t by 2030. - CDR.fyi's 2026 market update reports BiCRS methods at 96% of 2025 purchase volume, and that Microsoft accounted for 78.5% of all disclosed durable CDR tonnes as of April 2026. Treat all three as claims by interested parties. --- ## 5. Gaps neither source closes - **Price for 92.7% of the pathway's tonnes is simply unknown.** Microsoft, Boeing, JPMorgan and TD all declined to disclose. Anyone quoting a bio-oil "market price" is quoting Frontier. - **The literature is close to non-existent.** The evidence profile returns **one paper** for this pathway — a Dec 2025 study on subsurface hydrochar storage in the Netherlands (*Int. J. Greenhouse Gas Control*), study type unclassified. Zero field trials in the corpus. That is a striking asymmetry against >5 Mt contracted, and it means durability and leakage claims rest largely on registry protocols (Isometric) rather than peer-reviewed field evidence. Papers were fetched 2026-08-14; nothing later is held. - **No tracked policy instrument names this pathway** — the corpus says so explicitly. Bio-oil and slurry injection sit awkwardly relative to CO₂-specific instruments like 45Q, and neither the EU CRCF nor the July 2026 EU ETS review proposal (which contemplates 250 Mt of DACCS/BioCCS purchases over 2031–2040) obviously reaches it. I flag that as my reading, not a corpus finding. - **Pathway boundaries blur in practice.** Google's Jan 2025 100kt Charm deal is classified `biochar`, and the TD deal is explicitly biochar *and* bio-oil. Suppliers are selling across pathway codes, so pathway-level tonnage totals understate these companies' books. - **Freshness:** deals held to 2026-08-04, ingest 2026-08-15. My web searches ran past that but hit a quota limit; I did not verify NULIFE's post-offtake activity, any Q3 2026 announcements, or whether Charm/Vaulted have signed since August. CDR.fyi remains the fuller transaction scoreboard for this question.
Judge’s note — All four gold deals are surfaced with correct volumes, values, dates and correctly derived $/t ($473 / $382 / $362), plus form-of-storage and delivery-window detail. Beyond the core deal records the answer adds: (1) actual Frontier portfolio delivery figures for Charm (~100.6k of 112k) and Vaulted (~52.1k of 154k); (2) CDR.fyi market-wide ~3% (13% ex-Microsoft/Frontier) delivery-rate benchmark; (3) Frontier's disclosure that the Charm contract price declines 37–75% over 2024–2030 — genuinely the most informative pricing datapoint here; (4) the sub-threshold JPMorgan–Charm 28,500 t bio-oil deal and correction of the misattributed '$75M'; (5) explicit disambiguation of the Google–Charm 100k t deal as biochar, not pumpable biomass. It also benchmarks the pathway against Frontier's full offtake book, which is decision-useful context. Citations are named and linkable, and uncorroborated/supplier claims are flagged as such; the '~$1B' Microsoft estimate is correctly identified as speculation rather than used to derive an implied price. Coverage honesty is strong: it states that only 3 of 4 deals disclose value, that 93% of tonnes are unpriced, that the price series describes a disclosing minority of one buyer, and it gives an explicit freshness cutoff plus named gaps (no non-US/EU deals, no policy instrument, thin literature). Analyst inference is clearly fenced off from corpus fact. Minor deductions only: the three-point 'downward trend' framing is arguably over-read even with its own caveat (three different technologies), and there is a typo ('pyrolypsis'). Otherwise near-exemplary for an expert reader.
I'll start by checking what the corpus actually covers, then pull the pathway profile. Now let me pull the supplier entities and search for deals that may sit outside the pathway tag. ## Short answer The corpus holds **four offtake deals in the `biomass-injection` pathway** (pumpable biomass — bio-oil or organic slurry — pumped to geological storage), totalling **5,286,480 tonnes contracted, of which only three deals (386,480 t, 7.3% of tonnes) disclose a value**. Every disclosed price comes from a single buyer, Frontier. That one fact governs everything you can say about pricing here. --- ## The deals | Announced | Buyer | Supplier | Tonnes | Value | Derived $/t | Form & storage | Delivery | |---|---|---|---|---|---|---|---| | 2023-05-18 | Frontier | Charm Industrial (US) | 112,000 | $53.0M | **$473** | Bio-oil, fast pyrolysis → EPA-regulated wells | 2024–2030 | | 2024-05-01 | Frontier | Vaulted Deep (US) | 152,480 | $58.3M | **$382** | Organic-waste slurry → Class V wells (Kansas, Los Angeles) | 2024–2027 | | 2025-07-17 | Microsoft | Vaulted Deep (US) | up to 4,900,000 | **undisclosed** | — | Slurry of biosolids, manure, paper sludge, food/ag residues | 12 years, through 2038 | | 2025-12-18 | Frontier | NULIFE GreenTech (CA) | 122,000 | $44.2M | **$362** | Bio-oil via hydrothermal liquefaction → licensed salt caverns >1,000 m, Saskatoon; verified by Isometric | 2026–2030 | Sources: [Frontier/Charm](https://frontierclimate.com/writing/first-offtake); [Frontier/Vaulted](https://frontierclimate.com/writing/vaulted); [Microsoft/Vaulted PR Newswire](https://www.prnewswire.com/news-releases/vaulted-deep-signs-deal-with-microsoft-to-remove-carbon-via-waste-management-infrastructure-302507387.html); [Frontier/NULIFE](https://frontierclimate.com/writing/nulife). **Two adjacent items that are easy to miscount:** - **JPMorgan Chase ↔ Charm Industrial**, May 2023: 28,500 t of bio-oil injection over five years, value undisclosed. It sits below the corpus's inclusion threshold and is recorded only as a note on the [JPMorgan/Climeworks record](https://www.jpmorganchase.com/newsroom/press-releases/2023/jpmorgan-chase-seeks-to-scale-investment-in-emerging-carbon-removal-technologies). The "$75M" often attached to JPMorgan and Charm was actually JPMorgan's Frontier commitment. - **Google ↔ Charm Industrial**, 2025-01-16, 100,000 t: this is **biochar**, not bio-oil injection ([Axios](https://www.axios.com/2025/01/16/google-carbon-removal-charm-industrial-biochar)). Don't fold it into pumpable-biomass volumes. --- ## What the prices actually tell us **1. The headline trend is downward, but it's three points from one buyer.** $473 (2023) → $382 (2024) → $362 (2025) is a ~23% decline in derived price across three years. Treat it cautiously: these are *different companies with different processes* (fast pyrolysis, waste-slurry injection, hydrothermal liquefaction), not one supplier repricing. All three figures are **derived by dividing total contract value by contracted tonnes — none was quoted as a per-tonne price**, and each is a blended average across a multi-year delivery window. **2. The contracts themselves decline steeply within their term.** Frontier disclosed that the Charm contract price **falls by at least 37% and up to 75% between 2024 and 2030**. So the $473 average masks a schedule that starts well above it and ends far below — plausibly in the low hundreds by 2030. This is the single most informative pricing datapoint in the pathway, and it says the contracted *trajectory* matters more than the contracted *average*. **3. Pumpable biomass prices mid-pack against Frontier's other offtakes.** Ranked by derived price across Frontier's book: CO280 BECCS $214, Planetary OAE $272, Terradot ERW $300, Hafslund Celsio BECCS $316, **NULIFE $362**, Lithos ERW $370, **Vaulted $382**, Eion ERW $419, Reverion BECCS $427, CarbonCapture DAC $440, CREW $447, CarbonRun $458, **Charm $473**, 280 Earth DAC $650, Phlair DAC $651, Heirloom DAC $989. Pumpable biomass is materially cheaper than DAC and roughly level with enhanced weathering, but above point-source biogenic capture retrofits. **4. The largest deal by far is unpriced, and the circulating estimate is not evidence.** Microsoft's up-to-4.9 Mt agreement with Vaulted Deep — 93% of all contracted tonnes in this pathway — carries no disclosed value. The corpus explicitly flags third-party "~$1B" estimates as **speculation**. If that number were right it would imply roughly $204/t, i.e. a volume discount of ~45% against Frontier's 2024 price with the same supplier. That is a plausible-sounding story with no verified basis, and I would not use it. **5. These are prices for tonnes that are actually being delivered.** Frontier's portfolio (as of 2026-08) lists Charm at ~100,600 t delivered against 112,003 contracted (~90%) and Vaulted at ~52,100 against 154,146 (~34%). Frontier separately describes Charm (2023) and Vaulted (2024) as the field's first closed-system deliveries ([Frontier innovation-program post](https://frontierclimate.com/writing/innovation-program-2026), recorded as an uncorroborated source claim). Against a market where CDR.fyi puts the overall delivery rate at ~3%, rising to ~13% excluding Microsoft and Frontier ([CDR.fyi market update, 12 June 2026](https://www.cdr.fyi/blog/durable-cdr-market-update-carbon-unbound-east-coast-2026) — supplier claim, uncorroborated), these are unusually well-collateralised prices. --- ## Analyst read (my inference, not a corpus fact) The $360–475 band looks like an **AMC price for early, physically-delivering tonnes, inclusive of MRV**, not a market-clearing price. Two structural features argue the pathway's cost floor is lower than DAC's: the feedstock is a waste with disposal-cost credit (biosolids, manure, paper sludge), and Vaulted injects into **Class V** wells rather than Class VI — which, given that the corpus records Louisiana suspending Class VI review in October 2025 ([EO JML 25-119](https://www.doa.la.gov/media/elrfv1zx/jml-25-119-co2-capture-and-storage-limits-and-moratorium.pdf)) and Texas only receiving primacy in November 2025, is a real permitting-timeline advantage. Against that, HTL and pyrolypsis routes carry conversion capex that waste-slurry injection does not, which may explain why NULIFE prices near Vaulted despite being a first-of-a-kind Canadian salt-cavern project. --- ## Gaps and caveats you should carry forward - **Disclosure is the binding constraint.** Corpus-wide, 21 of 59 deals carry a value, covering 4.5% of contracted tonnes; Frontier publishes on every deal, the largest-volume buyers publish on none. Any price series here describes the disclosing minority. - **No policy instrument in the corpus names this pathway.** No procurement programme, tax credit or certification protocol is recorded as covering pumpable biomass specifically — a notable absence given that BECCS and DACCS have the UK GGR Business Model, the Swedish reverse auction and the proposed EU ETS 250 Mt central purchase (2031–2040). Policy absence shapes prospects as much as policy presence. - **The evidence base is thin.** One paper is classified to this pathway ([hydrochar subsurface storage in the Netherlands](https://doi.org/10.1016/j.ijggc.2025.104539), *Int. J. Greenhouse Gas Control*, Dec 2025) — and this is a curated, not exhaustive, paper set. - **Freshness.** Deals held through 2026-08-04, policy events through 2026-07-17. Anything after that — including any repricing, new Charm or Vaulted offtakes, or disclosure of the Microsoft contract value — is outside this dataset. CDR.fyi is the fuller transaction scoreboard for a census. - **No corpus record** of European or Asian pumpable-biomass offtakes, of Charm's non-Frontier bio-oil buyers beyond the sub-threshold JPMorgan deal, or of any government purchase in this pathway.
Asked, verbatim: “How much money has Frontier actually committed to carbon removal, across its AMCs, prepurchases, and offtakes? Break it down.”
It found every checklist fact on this question — not every question favors the database, and this page reports those too.
Judge’s note — All five checklist items are substantively present: the 2022 $1B/nine-year launch intent alongside the $925M initial funded figure, the 2023 ~$100M expansion with the four named joiners, the June 2026 $915M Growth AMC, three named small prepurchase rounds ($7M, $4.5M, $1.75M) with company lists, and a detailed offtake ledger naming Charm, Vaulted, NULIFE, CO280, CREW, Terradot, CarbonRun, Planetary, Hafslund, Phlair, with Arbor/Graphyte/Lithos flagged as undisclosed. The one weakness on item 5 is that it does not clearly separate Frontier-facilitated offtakes for non-AMC buyers from AMC-pool spend — it folds everything into one ~$698M contracted figure. Beyond-corpus value is real and unusually good: contracted-vs-pledged reconciliation ($698M across 53 projects / 1.8Mt), annual offtake run-rate ($166M 2023, $279M 2024, ~$170M to June 2026), 2025 and 2024 delivery volumes, per-deal values for recent offtakes, the 2030→2040 horizon shift, and a Microsoft scale comparison. The pledged/contracted/delivered-and-paid framing is exactly the decomposition an expert wants, and the flagged $1.94B-vs-$1.8B arithmetic gap is a genuinely sharp catch. Coverage honesty is exemplary — estimates are labelled as such, ranges given, missing round totals admitted, contract-enforceability caveats surfaced. Citation quality is the weak point: many quoted snippets float without a named source (only CTVC, ESG Dive and 'Frontier's annual letter' are attributed), and several offtake dollar figures (Stockholm Exergi, 280 Earth, Phlair) are not independently traceable and are plausible-but-unverified. Clarity and decision-usefulness are high.
I'll research Frontier's commitments across its facilities and deal types. # Frontier's money: pledged vs. contracted vs. actually paid The single most important distinction: **Frontier's headline number is a pledge, not a bank account.** Frontier is a purchasing facilitator — it operates through Delaware public benefit LLCs that are subsidiaries of Stripe — and its team "facilitates purchases from high-potential carbon removal companies on behalf of buyers," with buyers deciding how much they want to spend each year and Frontier aggregating those commitments into an annual demand pool . Contracts sit between individual buyers and suppliers. ## 1. The AMC pledge stack ($1.8B headline) | Date | Event | Amount | Cumulative | |---|---|---|---| | Apr 12, 2022 | Launch by Stripe, Alphabet, Shopify, Meta, McKinsey Sustainability | $925M | ~$925M | | Apr 12, 2023 | Autodesk, H&M Group, JPMorgan Chase, Workday join | $100M | ">$1B" | | Jun 17, 2026 | "Growth AMC" second tranche | $915M | **$1.8B** | At launch the founding companies planned to commit $1B over nine years to purchase permanent carbon removal, aiming to send a demand signal to researchers, entrepreneurs and investors ; the funded figure was an initial $925M between 2022 and 2030, funded by Stripe, Alphabet, Shopify, Meta, McKinsey and tens of thousands of businesses using Stripe Climate . The four 2023 members added a combined $100M over eight years, taking the AMC over $1B . In June 2026, Frontier committed an additional $915 million, with participating buyers including Stripe, Google, Shopify, Salesforce, H&M Group, McKinsey Sustainability, Workday and Anthropic — taking Frontier's total commitment to $1.8 billion . Frontier now describes itself as an advance market commitment to buy $1.8B of permanent carbon removal by 2040 — note the deadline has moved from 2030 to 2040, which materially slows the implied annual purchasing pace. **Gap to flag:** $925M + $100M + $915M = $1.94B, but Frontier states $1.8B. The ~$140M discrepancy is unexplained in public materials — plausible causes are member churn (Meta is conspicuously absent from the 2026 buyer list) or restatement of the original pledge. I could not verify which; treat the $1.8B as the authoritative figure and the arithmetic as unreconciled. ## 2. What's actually under contract: ~$698M As of the June 2026 announcement, Frontier's portfolio includes over 1.8 million tons of contracted carbon removal across 53 projects, with $698 million contracted, of which 463,700 tons for $169.8 million were contracted since June 2025 . CTVC's read is consistent: to date it has contracted ~$700m across 50+ projects . That implies: - **~39% of the $1.8B pledge** is contracted; **~68% of the original ~$1.03B** tranche. - **~$388/ton** blended across the portfolio; **~$366/ton** for deals signed since June 2025. Trajectory by year (Frontier's own accounting): - **2023 offtakes: $166M.** In 2024 Frontier signed a record $279M of offtakes, up from $166M in 2023, across seven companies — Vaulted, Stockholm Exergi, 280 Earth, CarbonRun, Terradot, CO280 and CREW . - **2024 offtakes: $279M** (the seven named deals sum to $279.4M — see table below). - **~$170M in the twelve months to June 2026**, per the ESG Dive figure above. ## 3. Offtakes (the bulk of the money — paid on delivery) Offtakes are legally binding contracts to purchase future tons if and when delivered, designed to help mature suppliers unlock financing , and they get paid out as tons are removed and sequestered . Named deals I can document: | Supplier | Value | Tons | Delivery window | |---|---|---|---| | Charm Industrial (first offtake, 2023) | $53M | 112,000 | 2024–2030 | | Vaulted Deep (2024) | $58.3M | — | — | | Stockholm Exergi (2024) | $48.6M | — | 2028–2030 | | 280 Earth (2024) | $40M | — | — | | CO280 (Dec 2024) | $48.0M | 224,500 | 2028–2030 | | CREW Carbon (Dec 2024) | $32.1M | 71,878 | 2025–2030 | | Terradot (2024) | $27M | 90,000 | — | | CarbonRun (2024) | $25.4M | 55,442 | — | | Planetary (Aug 2025, ocean alkalinity) | $31.3M | — | deliveries starting 2026 | | Hafslund Celsio (waste-to-energy, Norway) | $31.6M | 100,000 | 2029–2030 | | Phlair | $30.6M | 47,000 | 2027–2030 | | NULIFE GreenTech (Dec 2025) | US$44.2M | 122,000 | 2026–2030 | Others exist without public dollar figures in the sources I could reach (Lithos, CarbonCapture, Arbor, Graphyte, Eion, Reverion, Deep Sky). Treat this table as ~$470M of the $698M — it is not a complete ledger. ## 4. Prepurchases (small, upfront, high-risk) Prepurchases are the venture-style tail of the portfolio: early-stage suppliers get paid upfront for tons not yet delivered . Documented rounds: - **Fall 2023 (3rd round): $7M** — Stripe, Shopify and H&M spent $7 million across 12 companies: Airhive, Alkali Earth, Banyu Carbon, Carbon Atlantis, CarbonBlue, CarbonRun, EDAC Labs, Holocene, Mati, Planetary, Spiritus and Vaulted . - **Sept 2024 (4th round): $4.5M** — from 9 companies (Alithic, Alt Carbon, Anvil, Capture6, Exterra, Flux, NULIFE, Planeteers, Silica) for buyers Stripe, Shopify, Alphabet, H&M and Match via Watershed , covering close to 8,000 tons . - **July 2025 (5th round): $1.75M** — to Karbonetiq, Limenet and pHathom, on behalf of Google, Stripe, Shopify and others . Cumulatively, prepurchases are a rounding error on the dollar total but dominate the deal count: as of the 2024 annual letter, Frontier was first-ever customer for 26 of 33 prepurchase companies and first offtaker for 9 of its 11 offtakes . Rough order of magnitude across all five rounds: **~$20–30M**, i.e. under 5% of contracted dollars. I could not retrieve the 2022 and spring-2023 round totals, so that range is an estimate. ## 5. What has actually been *delivered and paid* This is the number almost never quoted, and it is tiny. Seven portfolio companies delivered ~23,000 tons in 2025, roughly 2x the prior year, and broke ground on 1.4m tons of new annual capacity; Frontier forecasts 50,000+ tons this year . For 2024, Vaulted delivered 12,000 tons — the most of any carbon removal company by a significant margin — while Charm injected 2,500 tons of bio-oil . Cumulative deliveries are therefore on the order of **40,000–60,000 tons against 1.8M tons contracted (~2–3%)**. At the portfolio's ~$388/ton blended price, that implies roughly **$15–25M of cash actually disbursed on offtakes** — my estimate, not a Frontier disclosure, and imprecise because early tons priced well above average and prepurchase cash is paid upfront regardless. ## 6. Caveats that should temper the headline - **Prepurchase dollars carry zero recourse.** Frontier states that a small portion of demand is contracted as prepurchases, and buyers "understand that there is no recourse if the supplier is unable to deliver tons." - **Offtake terms are deliberately lenient**, so contracted dollars can evaporate without penalty: under Frontier's template, suppliers aren't punished for under-delivering as long as they hit a minimum share of annual volume; if they miss even that, the buyer can terminate but the supplier faces no damages . Frontier also does not require security or collateral, because it seeks very limited punitive damages and its counterparties are often small startups . - **Frontier itself concedes delivery risk**: because the industry is early-stage, "it's likely that some suppliers will be delayed or fail to deliver," which is why it withholds registry and protocol detail until tons are verified. - **The clock got longer.** The original $1B was a 2022–2030 program; the $1.8B total now runs to 2040. - **Frontier is no longer the market's anchor buyer.** For scale, Microsoft had contracted roughly 45 million tonnes of CDR by January 2026, growing to nearly 50 million tonnes before an April 2026 procurement pause — roughly 25x Frontier's contracted volume. ## Bottom line - **Pledged: $1.8B** through 2040 (of which $915M was added only in June 2026 and is essentially unspent). - **Contracted: ~$698M** across 53 projects and >1.8M tons — about 39% of the pledge. - **Offtakes: the overwhelming majority** of that ~$698M ($166M in 2023, $279M in 2024, ~$170M in the year to June 2026). - **Prepurchases: roughly $20–30M**, spread across ~35–40 early-stage companies. - **Actually delivered and paid: on the order of $15–25M**, against ~40–60k tons. **Biggest gaps in the public record:** Frontier does not publish a cumulative cash-disbursed figure, a full deal-by-deal price list for every offtake, per-buyer allocations, or a reconciliation of the $1.94B in announced pledges against the stated $1.8B total. Anyone modelling Frontier's real-world capital deployment should treat the $1.8B as a demand signal and the ~$698M as the contracted ceiling, not as money in suppliers' hands.
Judge’s note — All five gold checklist items are substantively covered: the $1B/nine-year 2022 launch intent (with an explicit flag on the $925M vs $1B discrepancy), the ~$100M 2023 member expansion with named new members, the $915M June 2026 Growth AMC, five distinct prepurchase cohort rounds spanning $1.75M–$11M, and an 18-deal offtake table clearly separated from the AMC pledge layer. The answer's central contribution is exactly the distinction the question invites — pledged (~$1.8B) vs contracted (~$716M) vs delivered (~10% of tonnes) — and it avoids the common double-counting error while noting the arithmetic tension between $1B + $100M + $915M and Frontier's own $1.8B headline. Beyond-corpus findings include the Trellis reconciliation (~$700M / 1.8 Mt / 53 projects, matching the bottom-up sum), 2025 delivery volumes (~23 kt) and 2026 forecast (>50 kt), Charm and Vaulted per-supplier delivery progress, the 1.4 Mt of capacity that broke ground, the Graphyte prepurchase absent from the corpus, Anthropic/Salesforce membership, CarbonCapture Inc.'s disappearance from the portfolio page, and the $2M R&D grant/Innovation program track. Citations are named and dated (Frontier's own pages, ESG Today, Trellis, DataCenterDynamics, Carbon Herald) though not URL-level for every figure, and derived $/t values are explicitly labeled as computed rather than quoted. Coverage honesty is exemplary: it flags the prepurchase total as a floor, states no post-mid-August-2026 information was found, admits the Exergi price is uncertain, and points to CDR.fyi for cross-checking. The delivered-dollars range ($50–70M) is a soft estimate but is labeled as such. Structure and decision-usefulness are high — pathway and vintage breakdowns, the DAC price premium observation, and the overlapping-public-subsidy caveat all add analytical value beyond mere tallying. Minor deductions only for unverifiable per-deal precision and reliance on derived rather than sourced per-tonne figures.
I'll start by checking what the corpus covers, then pull Frontier's full record. Let me compute the totals carefully (the corpus has some aggregate records that duplicate individual deals) and check the web for anything after the corpus freshness date. Now let me check the web for anything after the corpus freshness date (Aug 2026) and for prepurchase rounds the corpus may have missed. ## Short answer Two very different numbers, and they get conflated constantly: | Measure | Amount | What it means | |---|---|---| | **Pledged (AMC capital)** | **~$1.8 billion** | Money buyers have said they will spend, mostly not yet contracted | | **Actually contracted (signed offtakes)** | **~$689 million** for ~1.85 Mt | Legally binding purchase agreements with named suppliers | | **Contracted (prepurchases, cohort rounds)** | **~$27 million** | Small, no-recourse, early-stage cheques | | **Delivered and paid against** | **~150–200 kt of CO₂, order of $50–70M** | The only tonnes that exist | So of $1.8bn pledged, roughly **$716M is under contract** and the overwhelming majority of that is a promise to pay *if and when* tonnes are delivered — largely between 2027 and 2030. --- ## 1. The pledges (AMC layer) **Original AMC (2022):** Frontier launched as an advance market commitment modelled on the pneumococcal vaccine AMC, with Stripe, Alphabet, Shopify, Meta and McKinsey Sustainability planning to commit **$1 billion of permanent carbon removal over nine years, through 2030** (frontierclimate.com/writing/launch, in the corpus). Note a discrepancy worth flagging: ESG Today reports the launch figure as **$925 million**, with the round number reached later. The corpus separately records a **$100M expansion in April 2023** when Autodesk, H&M Group, JPMorgan Chase and Workday joined (frontierclimate.com/writing/new-members) — if you add that to a $1B launch you double-count; if you add it to $925M you land at ~$1.025bn. **Growth AMC (17 June 2026): $915 million.** Frontier's own page describes ~10–15 focused bets via 8–10 year offtakes to carry projects to FID at commercial scale, contracting as far out as 2040. Press coverage (DataCenterDynamics, Yahoo Finance, ESG Today, 17–18 June 2026) confirms this brings **total commitments to $1.8 billion**, and that **Anthropic** joined as the first AI-company member alongside Salesforce. **Critical caveat:** these are commitments to *purchase*, not cash deployed, not escrowed, and not paid on signature. They are also fungible across pathways — the Growth AMC names no supplier, tonnage or price. --- ## 2. The offtakes — the real money This is where the substance is. Eighteen distinct Frontier-facilitated offtakes in the corpus, all with disclosed values (Frontier is the only large buyer that publishes a price on every deal): | Announced | Supplier | Pathway | Tonnes | $M | $/t | Delivery | |---|---|---|---|---|---|---| | 2023-05-18 | Charm Industrial | bio-oil injection | 112,000 | 53.0 | 473 | 2024–30 | | 2023-11-16 | Heirloom | DAC | 26,900 | 26.6 | 989 | by 2030 | | 2023-11-16 | CarbonCapture Inc. | DAC | 45,500 | 20.0 | 440 | by 2028 | | 2023-12-07 | Lithos Carbon | ERW | 154,240 | 57.1 | 370 | 2024–28 | | 2024-05-01 | Vaulted Deep | slurry injection | 152,480 | 58.3 | 382 | 2024–27 | | 2024-06-18 | Stockholm Exergi | BECCS | 179,998 | 48.6 | (~270) | 2028–30 | | 2024-07-11 | 280 Earth | DAC | 61,571 | 40.0 | 650 | 2024–30 | | 2024-09-23 | CarbonRun | River alkalinity | 55,442 | 25.4 | 458 | 2025–29 | | 2024-12-12 | Terradot | ERW | 90,000 | 27.0 | 300 | 2025–29 | | 2024-12-17 | CO280 | BECCS (pulp mill) | 224,500 | 48.0 | 214 | 2028–30 | | 2024-12-17 | CREW Carbon | Wastewater alkalinity | 71,878 | 32.1 | 447 | 2025–30 | | 2025-02-27 | Phlair | DAC (electrochem) | 47,000 | 30.6 | 651 | 2027–30 | | 2025-03-25 | Eion | ERW | 78,707 | 33.0 | 419 | 2027–30 | | 2025-04-01 | Hafslund Celsio | BECCS (WtE) | 100,000 | 31.6 | 316 | 2029–30 | | 2025-07-08 | Arbor | BECCS | 116,000 | 41.0 | 353 | 2028–30 | | 2025-08-26 | Planetary | OAE | 115,211 | 31.3 | 272 | 2026–30 | | 2025-11-25 | Reverion | BECCS (SOFC) | 96,000 | 41.0 | 427 | 2027–30 | | 2025-12-18 | NULIFE GreenTech | bio-oil injection | 122,000 | 44.2 | 362 | 2026–30 | | **Total** | **18 deals** | | **1,849,427** | **688.8** | **372 avg** | | *(Per-tonne prices are derived from value ÷ tonnes, not quoted by Frontier. The Exergi figure is bracketed because the corpus declines to derive it — tonnage was pending the Swedish state reverse auction at announcement.)* **This reconciles with Frontier's own accounting.** Trellis (17 June 2026) reported that since 2022 members have signed contracts worth almost $700 million for 1.8 million tons across 53 projects — within a rounding error of the $688.8M / 1.85 Mt I get by summing the corpus deal-by-deal. ### By pathway | Pathway | $M | Tonnes | $/t | Share of spend | |---|---|---|---|---| | BECCS / industrial BiCRS | 210.2 | 716,498 | 293 | 31% | | Biomass/bio-oil injection | 155.5 | 386,480 | 402 | 23% | | DAC | 117.2 | 180,971 | **648** | 17% | | Enhanced rock weathering | 117.1 | 322,947 | 363 | 17% | | Ocean/river alkalinity | 88.8 | 242,531 | 366 | 13% | DAC takes 17% of the money for under 10% of the tonnes. That price gap ($648/t vs $293/t for BECCS) is the single clearest signal in Frontier's portfolio. ### By year - **2023:** 4 deals, $156.7M, 338,640 t — $463/t - **2024:** 7 deals, $279.4M, 835,869 t — $334/t - **2025:** 7 deals, $252.7M, 674,918 t — $374/t - **2026:** no offtake in the corpus through the 4 Aug 2026 ingest cutoff — consistent with Frontier pausing new offtakes while raising the Growth AMC, but I cannot rule out an unrecorded deal. --- ## 3. The prepurchases — small by design Cohort rounds, no-recourse, meant to buy option value on early-stage suppliers. Frontier's own disclosures page states buyers who prepurchase understand there is no recourse if the supplier fails to deliver. | Round | $M | Recipients | |---|---|---| | 2022-06 (spring) | 2.40 | AspiraDAC, Calcite-Origen, Lithos, RepAir, Travertine, Living Carbon | | 2022-12 (fall) | 11.00 | Arbor, Captura, Arca, Carbon To Stone, Cella, CREW, InPlanet (7,011 t) | | 2023-09 (fall) | 7.00 | Airhive, Alkali Earth, Banyu, Phlair, CarbonBlue, CarbonRun, EDAC, Holocene, Mati, Planetary, Spiritus, Vaulted (12 companies) | | 2024-09 (fall) | 4.50 | Alithic, Alt Carbon, Anvil, Capture6, Exterra, Flux, NULIFE, Planeteers, Silica | | 2025-07 (summer) | 1.75 | Karbonetiq, Limenet, pHathom (2,982 t) | | **Corpus total** | **26.65** | | | 2025-11 Graphyte *(not in corpus)* | 0.25 | Graphyte "Carbon Casting," Arkansas (graphyte.com, Carbon Herald, 20 Nov 2025) | **Gap flag:** the corpus holds five cohort rounds; Frontier has run more than that, and its writing page also references **$2M in R&D grants** and a 2026 Innovation program targeting OAE, surficial mineralization and open-system MRV. Treat ~$27M as a floor for the prepurchase track, not a complete count. It is under 4% of contracted spend either way — the offtakes are, as Frontier itself says, the bulk of its spend. **The prepurchase-to-offtake pipeline works, though:** Vaulted, CarbonRun, Planetary, Phlair, Arbor, CREW and NULIFE all took a small prepurchase first and later won an eight-figure offtake. That's the AMC design functioning as intended. --- ## 4. What has actually been delivered — the number nobody leads with This is the honest bottom line. Against ~1.85 Mt contracted: - **~23,000 tonnes** delivered by seven portfolio companies in **2025**; **>50,000 tonnes forecast for 2026**, more than doubling for a second consecutive year (Frontier Growth AMC page; ESG Today, 17 June 2026). - **Charm Industrial:** ~100,600 t delivered of 112,003 contracted (Frontier portfolio page, Aug 2026) — the only offtake near completion. - **Vaulted Deep:** ~52,100 t delivered of 154,146 contracted. - Frontier also reports seven companies **broke ground on 1.4 Mt of new annual capacity** in 2025. So cumulative deliveries are plausibly around **150–200 kt** — roughly **8–11% of contracted tonnes** — with the bulk of both tonnes and cash sitting in 2028–2030 delivery windows for facilities that in several cases have not reached FID. --- ## 5. Caveats and where I'd push back on any headline number 1. **"$1.8bn committed" ≠ $1.8bn spent.** Roughly **$716M is contracted**; the remainder is a purchase intent. The Growth AMC's $915M has, as of the latest records, no supplier attached. 2. **Contracted ≠ paid.** Offtakes are delivery-contingent. Frontier's contract template pays for CRUs "if and when delivered." 3. **Counterparty failure is real, not hypothetical.** CarbonCapture Inc. ($20M / 45,500 t, 2023) no longer appears on Frontier's portfolio page as of Aug 2026 after Project Bison was paused in late 2024; delivery status is unconfirmed. That is a live example of why signed ≠ delivered. 4. **Contract prices decline over time.** Charm's price falls at least 37% and up to 75% across 2024–2030; CarbonCapture's fell at least 46%. Derived $/t figures are contract averages, not spot prices. 5. **Overlapping public money.** Stockholm Exergi reached FID in March 2025 on the back of Sweden's SEK 36bn BECCS reverse auction (~SEK 20bn over 15 years), and Hafslund Celsio sits inside Norway's Longship programme. Frontier's cheque is not the marginal dollar in either case. 6. **Coverage limits.** The corpus deal freshness is 4 Aug 2026 (ingest 15 Aug 2026); the Graphyte prepurchase was absent from it, so smaller items are demonstrably under-captured. Anything announced after mid-August 2026 is invisible to both the corpus and my searches. CDR.fyi is the fuller transaction scoreboard for cross-checking. **The one-line version:** Frontier has pledged $1.8bn, contracted about $716M of it for ~1.85 Mt at an average of ~$372/tonne, and taken delivery of somewhere near 10% of those tonnes — with the money and the risk both concentrated in 2028–2030.
Judge’s note — All five gold items are substantively present. The answer's central move — separating the AMC pledge (pledged, uncontracted), signed offtakes (contracted, $688.8M / 1.85 Mt across 18 named deals with derived $/t), and small prepurchases/R&D grants — is exactly the breakdown the question asks for, and it is executed with a deal-level table, by-year and by-pathway subtotals, and an explicit de-duplication note for the CO280+CREW and CarbonCapture+Heirloom aggregate announcements. AMC layers are all dated and sourced: April 2022 launch ($925M, five founding members), April 2023 ~$100M expansion crossing $1B, and the June 2026 $915M Growth AMC to $1.8B with Anthropic/Salesforce, plus the horizon shift from 2030 to 2040. Prepurchases are handled honestly: the corpus holds none (threshold artefact), the fall 2024 cycle is verified at $4.5M/9 companies with suppliers named, other cycles are flagged as unverified, and a $20–30M cumulative estimate is labelled as an estimate. Beyond-corpus contributions are strong and varied: Growth AMC details, fall 2024 prepurchase cohort, Frontier's own delivery reporting (~23,000 t in 2025, 1.4 Mt/yr under construction, >50,000 t forecast for 2026), the 2024 annual-letter reconciliation ($279M vs $166M, exposing a ~$9M 2023 gap), the ESG Dive $169.8M/463,700 t since June 2025 figure implying an unrecorded H1-2026 offtake, and the $2M Leilac/SaltX lime grant. The commitment-vs-delivery-vs-disbursement distinction is made sharply (contract ceilings, declining price curves, pay-on-delivery), which is the single most important caveat for an expert reader. Weaknesses are minor: $925M + ~$100M + $915M sums to ~$1.94B rather than the stated $1.8B and the rounding is not reconciled; the prepurchase range ($1.75M–$11M per round) is approximated rather than enumerated; and no item is dated after 2026-08-14, so the post-cutoff probe is satisfied only via progress-report reconciliation rather than genuinely newer announcements. Uncorroborated signals (CarbonCapture delisting, Innovation program) are correctly labelled.
I'll start by checking what the Carbon Curve corpus covers, then pull Frontier's full record. Now let me compute the totals precisely and check for prepurchase records, then search the web for what the corpus says it can't cover. The corpus holds Frontier's offtakes but explicitly no prepurchase records. Let me compute the offtake arithmetic and then search the web for the AMC total, prepurchase cycles, and anything since the corpus cutoff. Now the web searches the boundary field calls for — prepurchases, the AMC's headline total, and anything after the 2026-08 cutoff. ## Short answer Three different numbers get called "Frontier's commitment," and they are not the same money: | Layer | Amount | What it actually is | |---|---|---| | **The AMC pledge** | **$1.8B** total (April 2022 launch $925M + ~$100M new members 2023 + **$915M** "Growth AMC" June 2026) | A commitment by buyer companies to *spend* on removal by 2040 — not contracted to any supplier | | **Signed offtakes (contracted)** | **$688.8M across 18 agreements / 1,849,427 t** (Carbon Curve records, through 2025-12-18) | Binding pay-for-delivery contracts with named suppliers | | **Prepurchases + R&D grants** | Single-digit $M per cycle (2024 cycle: **$4.5M across 9 companies**) | Small upfront cheques to early-stage suppliers; **not held in the Carbon Curve corpus at all** | So of a $1.8B pledge, roughly **$0.69B is contracted to actual suppliers** in the records I can verify, and only a small fraction of *that* has been delivered and paid. --- ## 1. The offtakes — the enumerable part (Carbon Curve, all `review_status: approved`) Frontier is the one buyer in this corpus that discloses a value on every deal, so this is unusually complete. Note the database holds 20 Frontier deal records, but two are aggregate wrappers (the CO280+CREW $80M announcement and the CarbonCapture+Heirloom $47M announcement) that duplicate their components — I de-duplicated to **18 unique offtakes** before summing. | Date | Supplier | $M | Tonnes | $/t (derived) | Pathway | Where | |---|---|---|---|---|---|---| | 2023-05-18 | Charm Industrial | 53.0 | 112,000 | 473 | bio-oil injection | US | | 2023-11-16 | CarbonCapture Inc. | 20.0 | 45,500 | 440 | DAC | US | | 2023-11-16 | Heirloom | 26.6 | 26,900 | 989 | DAC | US | | 2023-12-07 | Lithos Carbon | 57.1 | 154,240 | 370 | ERW | US | | 2024-05-01 | Vaulted Deep | 58.3 | 152,480 | 382 | slurry injection | US | | 2024-06-18 | Stockholm Exergi | 48.6 | 179,998 | n/a* | BECCS | SE | | 2024-07-11 | 280 Earth | 40.0 | 61,571 | 650 | DAC | US | | 2024-09-23 | CarbonRun | 25.4 | 55,442 | 458 | river alkalinity | CA | | 2024-12-12 | Terradot | 27.0 | 90,000 | 300 | ERW | BR | | 2024-12-17 | CO280 | 48.0 | 224,500 | 214 | BECCS (pulp mill) | US | | 2024-12-17 | CREW Carbon | 32.1 | 71,878 | 447 | wastewater alkalinity | US | | 2025-02-27 | Phlair | 30.6 | 47,000 | 651 | electrochemical DAC | CA | | 2025-03-25 | Eion | 33.0 | 78,707 | 419 | ERW | US | | 2025-04-01 | Hafslund Celsio | 31.6 | 100,000 | 316 | waste-to-energy BECCS | NO | | 2025-07-08 | Arbor | 41.0 | 116,000 | 353 | BECCS | US | | 2025-08-26 | Planetary | 31.3 | 115,211 | 272 | OAE | CA | | 2025-11-25 | Reverion | 41.0 | 96,000 | 427 | biogas fuel-cell BECCS | DE | | 2025-12-18 | NULIFE GreenTech | 44.2 | 122,000 | 362 | bio-oil injection | CA | | **Total** | **18 suppliers** | **688.8** | **1,849,427** | **372 blended** | | | \*The Exergi record notes the announcement left tonnage pending Sweden's BECCS reverse auction; the tonnage was taken from Frontier's portfolio page, so the corpus deliberately derives no per-tonne price. **By year:** 2023 $156.7M / 338,640 t ($463/t) → 2024 $279.4M / 835,869 t ($334/t) → 2025 $252.7M / 674,918 t ($374/t). **By pathway:** - BECCS — 5 deals, $210.2M, 716,498 t, $293/t (31% of dollars) - Biomass injection (bio-oil/slurry) — 3 deals, $155.5M, 386,480 t, $402/t - DAC — 4 deals, $117.2M, 180,971 t, **$648/t** (17% of dollars for 10% of tonnes) - ERW — 3 deals, $117.1M, 322,947 t, $363/t - OAE — 2 deals, $63.4M, 187,089 t, $339/t - River alkalinity — 1 deal, $25.4M, 55,442 t, $458/t Every per-tonne figure above is **derived by division** from a disclosed total, not quoted by Frontier. The database's own price-disclosure note is worth repeating: 21 of 59 deals corpus-wide carry a value, covering just 4.5% of contracted tonnes, and Frontier is essentially the sole discloser. **A useful internal cross-check:** a Carbon Curve *signal* (uncorroborated, from Frontier's own 2026-08-06 BiCRS post) claims Frontier members have committed "over $360M in offtake agreements with eight BiCRS companies." My BECCS + biomass-injection subtotal is $365.7M across exactly eight suppliers. The records and the claim line up. --- ## 2. The AMC pledge — what web search adds - Frontier launched April 2022 with **$925M** in commitments from Stripe, Alphabet, Shopify, Meta and McKinsey Sustainability (frontierclimate.com; ESG Today, 17 June 2026). - April 2023: Autodesk, H&M Group, JPMorgan Chase and Workday joined, committing "a combined $100M ... over the next eight years," taking the AMC past $1B (frontierclimate.com/writing/new-members, 12 April 2023). - **17 June 2026: an additional $915M "Growth AMC," taking the total to $1.8B**, with Anthropic joining as the first AI company alongside Salesforce (frontierclimate.com/writing/growth-amc; ESG Today; Utility Dive, 22 June 2026). Frontier's homepage now describes itself as buying "$1.8B of permanent carbon removal by 2040" — note the horizon moved from 2030 to 2040. Two things to be precise about. First, the Growth AMC is framed by Frontier and by press as *investment in scaling companies*, not simply more offtake — Utility Dive and ESG Dive describe it as a financing pledge. Whether all $915M converts into tonne purchases is unresolved. Second, the original $1B was "by 2030"; four years in, roughly $689M of it is contracted, which is on pace but leaves a meaningful tail. Frontier's delivery numbers, from its own June 2026 post: seven portfolio companies delivered **~23,000 tonnes in 2025** (about 2× the prior year), broke ground on **1.4 Mt/yr of new annual capacity**, and forecast **>50,000 tonnes in 2026**. Against 1.85 Mt contracted, that is the gap between committed and delivered — the honest headline is that Frontier has *contracted* ~$689M and *received* a low tens of thousands of tonnes. --- ## 3. Prepurchases and R&D grants — the corpus's blind spot `search_records` for Frontier prepurchases returns **zero deals**. That is a threshold artefact, not absence of activity: prepurchases are typically $250k–$1M cheques, far below the corpus inclusion bar. What I could verify by search: - **Fall 2024 cycle: $4.5M across 9 companies** — Alithic, Alt Carbon, Anvil, Capture6, Exterra Carbon Solutions, Flux, NULIFE, Planeteers, Silica, on behalf of Stripe, Shopify, Alphabet, H&M Group and others (frontierclimate.com/writing/fall-2024-prepurchases; Trellis, 19 Sept 2024). This was the fourth prepurchase cohort. - Frontier's own application page references prepurchases announced in September 2023 and September 2024 and a 2025 purchasing cycle; the portfolio page carries "Prepurchase – 2025" entries (one as small as **333 tonnes**). - Two Carbon Curve **signals** (uncorroborated, Frontier's own writing): a **$2M R&D grant** to Leilac and SaltX for zero-carbon lime (22 Jan 2026), and a 2026 restructuring of the prepurchase track into an "Innovation program" that explicitly funds R&D rather than only buying tonnes, with flexible cheque sizes. Treat both as source claims, not verified records. **Honest gap:** I could not verify the dollar totals of the 2022, 2023 and 2025 prepurchase cycles in this session (search budget exhausted). If each ran in the $2.5–7M range, cumulative prepurchases plausibly sit in the **$20–30M** area — i.e. ~3–4% of Frontier's committed dollars, but the source of most of its pipeline. Vaulted Deep went from prepurchase to a $58.3M offtake inside a year; NULIFE from a 2024 prepurchase to a $44.2M offtake. Don't judge that layer by its size. --- ## 4. Where the two sources disagree 1. **2023 offtake total.** Frontier's own 2024 annual letter states 2024 offtakes totalled **$279M, up from $166M in 2023**. The corpus sums 2024 at **$279.4M** — a near-exact match — but 2023 at only **$156.7M**. Even using Frontier's rounded $47M headline for CarbonCapture+Heirloom you reach $157.1M. **~$9M of 2023 commitment is unaccounted for** in the deal records. Possible explanations: prepurchases folded into the letter's total, an undisclosed or unannounced agreement, or a definitional difference. Unresolved. 2. **A missing 2026 offtake.** ESG Dive (22 June 2026) reports Frontier "has contracted to remove 463,700 tons of carbon dioxide for $169.8 million since June 2025." The corpus deals announced after June 2025 (Arbor, Planetary, Reverion, NULIFE) total **449,211 t / $157.5M** — leaving roughly **14,500 t and $12.3M**, at ~$850/t, unaccounted for. That strongly implies **at least one Frontier offtake announced in H1 2026 that the corpus does not yet hold** (its latest Frontier record is 2025-12-18). The high implied price points to DAC or a small novel pathway. I could not identify it before exhausting search calls. 3. **Delivery figures.** The corpus notes Frontier's portfolio page listing ~100.6k tonnes delivered against Charm's 112,003 contracted, and ~52,100 against Vaulted's 154,146, as of 2026-08. Those sit awkwardly against Frontier's statement of ~23,000 tonnes delivered across seven companies in 2025. The two may be measuring different things (cumulative vs annual, or contracted-and-scheduled vs verified-and-retired), but I would not quote either without checking the portfolio page directly. 4. **One deal may be dead.** A corpus signal flags that **CarbonCapture Inc. no longer appears on Frontier's portfolio page** as of 2026-08, and Project Bison (Wyoming) was reported paused in late 2024. The $20M / 45,500 t offtake's status is unconfirmed. If it has been unwound, the $688.8M figure is overstated by that amount. Uncorroborated — Frontier has made no recorded public statement. --- ## What neither source covers - **Cash actually disbursed.** Offtakes are pay-on-delivery, often with milestone structures and declining price curves (Charm's contract price falls at least 37% and up to 75% between 2024 and 2030; CarbonCapture's by at least 46%). "$688.8M committed" is a contract ceiling paid over 2024–2030, not money out the door. Nobody publishes the disbursement schedule. - **How the $915M Growth AMC splits** between tonne purchases, equity/project finance and other instruments. - **Per-buyer attribution.** Frontier deals are signed on behalf of a rotating list of members and Watershed-channel buyers; no deal record discloses who paid what share. - **Prepurchase cycle totals for 2022, 2023 and 2025**, and any 2026 Innovation-program awards beyond the $2M lime grants. - **Anything after 2026-08-04** (corpus deal cutoff) or 2026-08-15 (last ingest). Frontier typically announces offtakes in clusters late in the year, so assume additional deals exist that neither layer here captures. CDR.fyi is the fuller transaction scoreboard for cross-checking.
Judge’s note — The answer is an unusually well-built offtake ledger: 18 deals, dated, priced, per-tonne, with named Frontier writing-page URLs, pathway and geography rollups, a de-duplication caveat, and an internal cross-check against Frontier's own $360M BiCRS aggregate. Structurally it is excellent and its epistemic hygiene is exemplary — it says plainly what it cannot see and refuses to launder outside knowledge as sourced fact. But against the question actually asked — total committed across AMCs, prepurchases and offtakes — it under-delivers on three of five checklist items. The original $1B/nine-year AMC (2022) is not surfaced; worse, the answer offers '~$925M/2022–2030' as the 'commonly quoted' figure, which is inaccurate. The ~$100M 2023 AMC expansion with new members is absent entirely. Prepurchases are acknowledged as a category but quantified only as vague '~$500k-scale cheques across ~30+ suppliers,' with none of the actual round totals ($1.75M–$11M each) — a gesture, not a finding. The $915M Growth AMC is surfaced with correct figure and June 2026 date, though only as an unverified one-line media summary rather than a confirmed commitment. The offtake layer is comprehensively covered with the right named suppliers. Beyond-corpus: the answer is candidly corpus-bound and produces no post-2026-08-14 findings. It does deliver some spend-vs-commitment reconciliation — Charm ~100.6k of 112k tonnes delivered, Vaulted ~52.1k of 154k, plus CDR.fyi's ~3% market delivery rate and the CarbonCapture/Project Bison status flag — which is the most valuable non-ledger content here. Roughly two distinct such findings. Net: a rigorous, honest, decision-useful partial answer whose headline figure ($688.8M contracted) is a defensible floor but leaves the AMC-vs-prepurchase-vs-offtake breakdown the user requested largely unreconciled, with one affirmatively wrong AMC number.
I'll start by checking corpus coverage and pulling Frontier's entity record. The entity record covers offtakes through Dec 2025. Let me check for 2026 activity, prepurchases, and the growth fund referenced in the media layer. Let me chase the prepurchase layer and the $915M growth fund specifically. Let me check the funding layer and the media item on the growth fund. ## Headline number **Across the 18 distinct offtake agreements Frontier has publicly announced and that this corpus holds, the total disclosed commitment is $688.8 million for 1,849,427 tonnes of durable CDR — a volume-weighted average of ~$372/tonne.** Every one of those deals carries a disclosed dollar value, which makes Frontier almost unique as a buyer (the corpus notes that only 21 of 59 tracked deals disclose value at all, covering 4.5% of tracked tonnes, and that disclosure "is concentrated in one buyer"). One methodological warning before the breakdown: the corpus stores two announcements twice — once as a combined record and once split by supplier (CarbonCapture+Heirloom = $46.6M; CO280+CREW = $80M). I have used the split records only. Naively summing every Frontier "deal" record returns ~$815M, which is wrong by $126.6M. ## Year by year | Year | Deals | Committed | Tonnes | $/t | |---|---|---|---|---| | 2023 | 4 | $156.7M | 338,640 | $463 | | 2024 | 7 | $279.4M | 835,869 | $334* | | 2025 | 7 | $252.7M | 674,918 | $374 | | **Total** | **18** | **$688.8M** | **1,849,427** | **$372** | \*Stockholm Exergi ($48.6M / 179,998 t) has no derivable per-tonne price — Frontier's June 2024 announcement left tonnage pending the Swedish state reverse auction, and the tonnage shown comes from Frontier's portfolio page as of Aug 2026. Excluding it, 2024 averages $352/t. ## Every deal **2023 — $156.7M** - Charm Industrial, 18 May 2023: $53.0M / 112,000 t / $473 (bio-oil injection; Frontier's first offtake; contract price declines 37–75% over 2024–2030) — [frontierclimate.com/writing/first-offtake](https://frontierclimate.com/writing/first-offtake) - CarbonCapture Inc., 16 Nov 2023: $20.0M / 45,500 t / $440 (DAC, Project Bison, Wyoming) — [carboncapture-heirloom](https://frontierclimate.com/writing/carboncapture-heirloom) - Heirloom, 16 Nov 2023: $26.6M / 26,900 t / $989 (limestone-looping DAC) — same source - Lithos Carbon, 7 Dec 2023: $57.1M / 154,240 t / $370 (first ERW offtake globally) — [writing/lithos](https://frontierclimate.com/writing/lithos) **2024 — $279.4M** - Vaulted Deep, 1 May: $58.3M / 152,480 t / $382 — [writing/vaulted](https://frontierclimate.com/writing/vaulted) - Stockholm Exergi, 18 Jun: $48.6M / 179,998 t — [writing/exergi](https://frontierclimate.com/writing/exergi) - 280 Earth, 11 Jul: $40.0M / 61,571 t / $650 — [writing/280-earth](https://frontierclimate.com/writing/280-earth) - CarbonRun, 23 Sep: $25.4M / 55,442 t / $458 (world's first river alkalinity enhancement offtake) — [writing/carbonrun](https://frontierclimate.com/writing/carbonrun) - Terradot, 12 Dec: $27.0M / 90,000 t / $300 — [writing/terradot](https://frontierclimate.com/writing/terradot) - CO280, 17 Dec: $48.0M / 224,500 t / $214 (largest Frontier offtake by tonnage at announcement) — [co280-crew](https://frontierclimate.com/writing/co280-crew) - CREW Carbon, 17 Dec: $32.1M / 71,878 t / $447 — same source **2025 — $252.7M** - Phlair, 27 Feb: $30.6M / 47,000 t / $651 (first electrochemical DAC offtake) — [writing/phlair](https://frontierclimate.com/writing/phlair) - Eion, 25 Mar: $33.0M / 78,707 t / $419 — [writing/eion](https://frontierclimate.com/writing/eion) - Hafslund Celsio, 1 Apr: $31.6M / 100,000 t / $316 — [writing/hafslundcelsio](https://frontierclimate.com/writing/hafslundcelsio) - Arbor, 8 Jul: $41.0M / 116,000 t / $353 — [writing/arbor](https://frontierclimate.com/writing/arbor) - Planetary, 26 Aug: $31.3M / 115,211 t / $272 (largest OAE offtake to date at announcement) — [writing/planetary](https://frontierclimate.com/writing/planetary) - Reverion, 25 Nov: $41.0M / 96,000 t / $427 — [writing/reverion](https://frontierclimate.com/writing/reverion) - NULIFE GreenTech, 18 Dec: $44.2M / 122,000 t / $362 — [writing/nulife](https://frontierclimate.com/writing/nulife) ## By pathway | Pathway | Committed | Tonnes | $/t | Share of $ | |---|---|---|---|---| | BECCS | $210.2M | 716,498 | ~$293 | 31% | | Biomass injection (bio-oil/slurry) | $155.5M | 386,480 | $402 | 23% | | DAC | $117.2M | 180,971 | $648 | 17% | | Enhanced weathering (ag) | $117.1M | 322,947 | $363 | 17% | | OAE (incl. wastewater) | $63.4M | 187,089 | $339 | 9% | | River alkalinity enhancement | $25.4M | 55,442 | $458 | 4% | DAC is the clearest structural story: 17% of the money for 10% of the tonnes, at roughly 2.2× the portfolio average price. **A useful cross-check:** Frontier's own 6 Aug 2026 post claims members "have committed over $360M in offtake agreements with eight BiCRS companies" ([deploying-bicrs-responsibly](https://frontierclimate.com/writing/deploying-bicrs-responsibly), held in the corpus as an uncorroborated signal). Summing BECCS + biomass-injection here gives $365.7M across exactly eight suppliers (Charm, Vaulted, NULIFE, Reverion, Hafslund, CO280, Exergi, Arbor). The deal-level records reconcile with Frontier's own aggregate. ## By geography US ~$335.6M (10 deals) · Canada ~$131.5M (Phlair/Alberta, Planetary/Halifax, CarbonRun/Nova Scotia, NULIFE/Saskatoon) · Europe $121.2M (Exergi SE, Hafslund NO, Reverion DE) · Brazil $27.0M (Terradot). Minor conflict to flag: the NULIFE record describes it as "Frontier's third Canadian offtake," but four deals carry a CA jurisdiction tag — Phlair's Munich HQ may explain the discrepancy. ## What this number is *not* — the important gaps **1. Prepurchases are entirely absent from the corpus.** A direct search for Frontier's prepurchase program returns zero records. Frontier's early-stage program (roughly $500k-scale cheques across ~30+ suppliers per cycle since 2022) is a real committed sum that is simply not captured here. The $688.8M therefore understates total Frontier commitments, and the offtake/prepurchase split you asked for cannot be produced from this corpus. A signal dated 2026 notes Frontier is "evolving its early-stage 'prepurchase program'" into a 2026 Innovation program that funds R&D directly with "more flexible check sizes" ([innovation-program-2026](https://frontierclimate.com/writing/innovation-program-2026)) — so the category is changing shape as well as being unrecorded. **2. The AMC ceiling versus deployed capital.** Frontier is described in its entity record as an "advance market commitment... founded by Stripe, Alphabet, Shopify, Meta and McKinsey." The corpus does not hold a record of the headline AMC size or its expiry, so I cannot cite the commonly quoted ~$925M/2022–2030 figure from this evidence base — treat it as outside-knowledge context, not a sourced fact here. What *is* sourced is that $688.8M of contracted offtakes has been announced. **3. A separate $915M "growth fund" appears in the media layer but nowhere in the deal layer.** Noah Deich's 17 Jun 2026 piece is summarised as "Frontier's $915M carbon removal growth fund read against the history of solar PV" ([ndeich.substack.com](https://ndeich.substack.com/p/long-live-cdr)). The corpus holds only that one-line summary — no size confirmation, no structure, no deals attached. If this is a new vehicle rather than a restatement of the original AMC, total Frontier-mobilised capital is far above $689M. This is the single biggest open question in the numbers and needs primary-source verification. **4. Non-offtake commitments.** A Jan 2026 signal records $2M in R&D grants to Leilac and SaltX for zero-carbon lime, on behalf of Stripe, Shopify and Google ([zero-carbon-lime-grants](https://frontierclimate.com/writing/zero-carbon-lime-grants)). The Quebec Surficial Mineralization Hub at Thetford Mines (Apr 2026, with Carbon Removal Canada) commits 10,000 tonnes of serpentinite tailings and pilot space, with no dollar figure disclosed ([surficial-mineralization](https://frontierclimate.com/writing/surficial-mineralization)). Both are uncorroborated source claims. **5. No 2026 offtakes are recorded at all.** The last Frontier deal held is NULIFE, 18 Dec 2025, against a corpus deal-freshness date of 4 Aug 2026 (ingest 15 Aug). Frontier was demonstrably active in 2026 on grants, hubs and program design. Whether it signed no new offtakes for ~8 months or the corpus simply hasn't caught them, I can't tell from here — but a genuine pause would be significant, given CDR.fyi's observation that the market is heavily concentrated in Microsoft and Frontier. **6. Committed ≠ delivered ≠ paid.** Roughly 100,600 of Charm's 112,003 contracted tonnes and only ~52,100 of Vaulted's 154,146 had been delivered as of Aug 2026 (Frontier portfolio page). CDR.fyi's June 2026 analysis puts the market-wide delivery rate at ~3%, rising to ~13% once Microsoft and Frontier are excluded — i.e. Frontier's book is heavily forward-weighted, with delivery windows running to 2030. Separately, the $20M / 45,500 t CarbonCapture Inc. offtake is flagged: the company no longer appears on Frontier's portfolio page as of Aug 2026 and Project Bison was reported paused in late 2024, with delivery status unconfirmed and no public Frontier statement recorded. Treat $688.8M as *contracted*, not *spent*. **Bottom line:** $688.8M across 18 announced offtakes and 1.85 Mt is the defensible, fully-sourced floor. The true all-in figure — including prepurchases, the 2026 innovation/R&D lines, and whatever the $915M growth fund actually is — is higher, and the corpus cannot currently quantify the difference.
Asked, verbatim: “What is the current policy landscape for durable carbon removal in the EU?”
Judge’s note — Very strong on the two central pillars: the CRCF architecture is described at instrument level (Reg. 2024/3012, Implementing Reg. 2025/2358, Delegated Reg. 2026/285 covering DACCS/BioCCS/biochar, pending carbon-farming and bio-based product acts, ERW/OAE only 'under consideration', registry consolidation), and the 2026 ETS review is covered with real design detail (central buyer/Removals Authority, no direct offsetting, 250 Mt allowances auctioned 2031-2040, ~48 Mt/yr by 2040, CRCF-certified DACCS/BioCCS eligibility only, third-country storage from 2031, payment on delivery, supersession of the Buyers' Club) plus a genuinely useful caveat that 250 Mt is allowances not tonnes and that 1 EUA \u2260 1 CDR unit. It does not, however, name Article 9c or systematically work through the CDR-only vs CDR-and-CCS scope distinction across instruments \u2014 CCS-adjacent files (NZIA injection capacity, CO2 transport/storage package, Industrial Carbon Management Strategy) are only listed as unverified, so the scope-differentiation dimension is essentially absent. On the beyond-corpus probe it scores nothing substantive: no consultation deadlines, no post-mid-August-2026 developments, and member-state programmes (Sweden BECCS auction, Danish NECCS tenders, SDE++) are named only in a list of things it could not verify \u2014 honest, but zero content. Citations are consistently attributed to named, checkable sources (Carbon Gap, Linklaters, ICAP, CATF, HSF Kramer, AlliedOffsets) with regulation numbers, though several specific instrument numbers/dates (2026/285, 2026/667, COM(2026) 616) carry hallucination risk and are asserted with more confidence than the hedging elsewhere would suggest. Coverage honesty is exemplary \u2014 an explicit, specific list of what was not verified and why, plus identification of the pre-2031 demand gap and storage capacity as the likely binding constraint. Structure and decision-usefulness are high for a practitioner: clear bottom line, table for the ETS design, timeline/politics, and a 'what to watch' list.
I'll research the current EU policy landscape for durable carbon removal across several tracks. ## Bottom line (as of mid-August 2026) The EU has moved from *rule-setting* to *implementation and demand creation* on durable CDR in the past 12 months. Two things now define the landscape: (1) the **CRCF certification architecture is live** for permanent removals, and (2) the Commission's **17 July 2026 ETS review proposal** would create the first large-scale public demand channel — but it is a proposal, not law, and political agreement is not expected before Q1 2027. --- ## 1. Certification: the CRCF is now operational for durable methods - **Framework law:** Regulation (EU) 2024/3012 (the Carbon Removals and Carbon Farming Regulation) established the first EU-wide **voluntary** certification framework, covering permanent removals, carbon farming and carbon storage in products (European Commission; European Accreditation). - **Governance layer:** **Implementing Regulation (EU) 2025/2358**, published in the Official Journal in **November 2025**, set the rules on certification schemes, certification bodies and audits. Certification bodies must be accredited to **EN ISO/IEC 17065** by a national accreditation body under Regulation (EC) 765/2008, or recognised as competent by a national competent authority (European Accreditation summary). - **The key milestone: Delegated Regulation (EU) 2026/285**, adopted **3 February 2026**, sets the certification methodologies for three permanent activities — **DACCS, BioCCS (biogenic emissions capture with storage), and biochar** (European Commission, "Certification methodologies"; Carbon Management Europe, 3 Feb 2026). It defines what counts as a tonne, how permanence must be ensured, and how leakage/reversal liability is handled. Practical effect: DAC, BECCS and biochar projects can now apply for EU certification, giving corporate buyers a government-backed quality benchmark for the first time (Senken analysis, June 2026). - **Registries:** certification schemes must run interoperable public registries; these are to be replaced by a **single Union-wide registry within four years** (Carbon Gap CRCF tracker). - **Still pending:** delegated acts for **carbon farming** and **carbon storage in bio-based construction products** were flagged for adoption during 2026; **enhanced rock weathering and ocean alkalinity enhancement** remain only "under consideration" — the Commission held a dedicated workshop on them on **25 September 2025** but has not committed to methodologies. Note that biochar's inclusion as "permanent" was contested during drafting, and Carbon Management Europe describes the outcome as a compromise pairing science-based accounting with "reviewable flexibility" for early deployment. ## 2. Demand: the 17 July 2026 ETS review is the main event Article 30 of the ETS Directive obliged the Commission to report on how negative emissions could be accounted for in emissions trading. It discharged that duty through the **ETS revision proposal COM(2026) 616 of 17 July 2026**, presented alongside an Electrification Action Plan and the €100bn Industrial Decarbonisation Bank (Carbon Gap; CATF, 17 July 2026; Linklaters). Design, as proposed: | Element | Proposal | |---|---| | Mechanism | **No direct offsetting.** The Commission acts as **central buyer** via a new "Removals Authority"; emitters cannot buy removal credits for compliance (Carbon Gap; Enerdata; Linklaters) | | Volume/finance | **250 million additional allowances** issued and auctioned **2031–2040**, plus **~10 Mt contingency** to bridge the gap between removal costs and the EUA price (Carbon Gap; Carbon Herald) | | Ramp | Purchases ramping to roughly **48 Mt/yr by 2040**, cumulative ~250 Mt units by 2040 (Enerdata; Carbon Herald) | | Eligibility | Only **CRCF-certified DACCS and BioCCS** initially — chosen because reversals can be monitored — subject to ETS MRV and liability rules (Linklaters; Carbon Gap) | | Geography | Anticipates storage in **third countries from 2031**, if that country's ETS is linked to the EU ETS (HSF Kramer) | | Procurement style | Competitive allocation where feasible, with payment on delivery; this **supersedes the voluntary EU Buyers' Club** as the primary near-term demand instrument (AlliedOffsets) | **Important caveat on the headline number:** the 250 Mt figure is a quantity of *allowances to be auctioned*, not a guaranteed tonnage of removals. As Carbon Gap and AlliedOffsets both stress, actual volumes purchased depend on the EUA clearing price, removal prices and programme design. Because the purchased removals sit outside the cap, the linear reduction factor does not apply to them, and one EUA ≠ one CDR unit — a price-convergence problem the proposal does not fully resolve. ## 3. The surrounding architecture - **2040 target:** the amended European Climate Law (Regulation (EU) **2026/667**, per AlliedOffsets) locks in the 2040 goal and permits **international credits up to a ceiling described as 5% of 1990 net emissions from 2036**, an interpretation that has generated wide market speculation. Removals integration into the ETS was already agreed in principle in the 2040 framework (Carbon Brief). - **Safeguard clause:** the Commission must report to Parliament and Council by **January 2033** on the international credits market (integrity, verification, supply/demand). If high-quality, cost-effective credits are unavailable, the **LRF reverts to 2.7% from 2036**, and set-aside allowances go to the Industrial Decarbonisation Bank. A further report on removals is due by **end-2034** (ICAP). - **Interacting files:** CBAM began phasing in at the start of 2026, with free allocation phased out 2026–38 — though the Commission has proposed reinstating 15% of the allocation otherwise removed from 2028 (Carbon Brief). A separate fallback-benchmarks proposal of the same date adds roughly €6bn in free allocation for 2026–30 (HSF Kramer; ICAP). Innovation Fund revenue remains the main current public funding source for technology-based CDR (Carbon Gap). ## 4. Timeline and politics Council and Parliament are **targeting political agreement by Q1 2027**, with the substantive fight starting from **September 2026** (Linklaters). Expect the removals provisions to be contested on three axes: whether direct compliance use should replace central purchasing; whether biochar and, later, ERW/OAE get in; and whether the removals cap increase is a legitimate flexibility or a loosening of ambition. CATF's reaction was explicitly cautionary — it warned the package "loosens the system in anticipation of removals and international credits that still have to materialise." ## 5. Gaps and uncertainty — flagged honestly - **Pre-2031 demand is thin.** The purchasing programme starts in 2031. Between now and then, durable CDR in the EU depends on the voluntary market, the Innovation Fund/Industrial Decarbonisation Bank, and national schemes — the sequencing risk for projects reaching FID in 2027–29 is real and unresolved. - **I could not verify several adjacent items in this session** (my search budget was exhausted): the current status of national procurement programmes (e.g. Sweden's BECCS reverse auction, Denmark's CCUS/NECCS tenders, the Dutch SDE++), the state of the CO₂ transport-and-storage regulatory package flowing from the 2024 Industrial Carbon Management Strategy, NZIA implementation of the 50 Mt/yr injection-capacity objective, the LULUCF review, and how removals are treated under ESRS/green-claims rules. Treat anything you read from me on those as unconfirmed until checked. - **Whether the carbon-farming and bio-based-products delegated acts have actually been adopted** since spring 2026 is not confirmed by the sources I retrieved — they were "expected in 2026." - **Storage capacity, not certification, may be the binding constraint** on whether 48 Mt/yr by 2040 is deliverable; none of the sources I found quantified the permitted EU injection capacity required to meet it. **What to watch next:** the Parliament ENVI rapporteur's draft report on COM(2026) 616 (autumn 2026); Council general approach; the promised follow-up legislation on the international-credits purchasing facility "later in 2026" (ICAP); and CRCF scheme recognitions and the first certified units under Delegated Regulation 2026/285.
Judge’s note — All three gold items are substantively covered. CRCF is handled in depth (Regulation (EU) 2024/3012, adoption/OJ/entry-into-force chronology, three unit types, Delegated Regulation (EU) 2026/285 covering DACCS/BioCCS/biochar, Implementing Regulation (EU) 2025/2358 on schemes and audits, pending carbon-storage-in-products act, Union Registry due 2028, ERW/marine gaps). The 17 July 2026 ETS review with the new Article 9c central procurement mechanism is described with concrete parameters (250 Mt allowances plus ~10 Mt reserve, 2031–2040 auctioning, Removals Authority, competitive tenders, payment on delivery, ~48 Mt/yr by 2040, rejection of direct offsetting, Article 14(1a) netting carve-out, Q1 2027 political-agreement target). Instrument scope distinctions are present at the instrument level rather than as an explicit taxonomy: CRCF-certified-but-ETS-ineligible biochar, geological-storage-only eligibility, Sweden's BECCS-only reverse auction versus Denmark's CCS Fund awarding a cement-capture (non-removal) project, Innovation Fund as a mixed carbon-management instrument, and NZIA injection obligations flagged as CCS infrastructure outside the CDR files — enough to count, though it is inferred rather than framed as a cdr-only vs cdr-and-ccs distinction. Beyond-corpus substantive findings: several member-state and implementation-level details that go past a bare EU-instrument listing (Sweden SEK 36bn 2026–2046 auction plus the Stockholm Exergi SEK 20bn award and reported ~15% CCS budget trim; Denmark's DKK 28.7bn fund and Aalborg Portland up to DKK 16.5bn; Germany's €111m 2026 line including €11.5m procurement and €320m through 2033; Innovation Fund €2.7bn/54 projects March 2026; Implementing Regulation 2025/2358 and the 2028 Registry deadline; Buyers' Club go-live and December 2026 purchase target; City of Stockholm 50 kt/yr 15-year contract). However, the answer explicitly fails the post-2026-08-14 part of the probe — it admits a search-tool limit blocked verification of anything after mid-August 2026, and no consultation deadlines are given. Citations are exceptionally traceable: named regulations with numbers and dates, plus attributed secondary sources (Taylor Wessing, A&O Shearman, ICAP, Linklaters, AlliedOffsets, CDR.fyi, E3G, Carbon Gap, DG CLIMA/ISCC). Coverage honesty is outstanding: it distinguishes verified from unverified items (Innovation Fund BECCS Stockholm award amount, Commission COM number, German LT-NET strategy), flags month-precision records, states the corpus cutoff and the 14-jurisdiction limit, and warns that absence of FR/NL/ES instruments is a coverage gap not inaction. Structure is clean and decision-useful, with a sharp bottom line (certification-first, demand-later) and well-chosen tensions (250 Mt is allowance volume not a purchase commitment; ~€25bn funding gap; certification/eligibility split). Marked down slightly from top marks for the unmet post-cutoff verification and absence of consultation-deadline detail.
I'll start by checking what the corpus covers, then dig into EU-specific policy. Let me pull recent changes and search the web for anything past the corpus freshness dates. ## Bottom line The EU now has the most complete durable‑CDR policy stack of any jurisdiction, but it is still **certification‑first and demand‑later**: the certification layer is law, the compliance‑demand layer is a July 2026 legislative proposal that would not buy a tonne before 2031, and the near‑term money is coming from grants and national schemes rather than from the ETS. --- ## 1. Certification layer — in force **CRCF Regulation (EU) 2024/3012** is the umbrella. Per the Carbon Curve policy corpus, it was adopted by the European Parliament at first reading on 10 April 2024, formally adopted by Council on 19 November 2024, published in the Official Journal on 6 December 2024 and entered into force on 26 December 2024. It is pathway‑neutral and covers three unit types: permanent removals, carbon farming, and carbon storage in products. The pathway rules arrive as delegated acts. **Delegated Regulation (EU) 2026/285** — the substantive methodologies — was adopted in early February 2026 (the corpus records the adoption event as February 2026 at month precision; Taylor Wessing's 13 Feb 2026 note dates the act to 3 February 2026). The Commission's own certification‑methodologies page confirms it sets methodologies for three permanent‑removal activities: DACCS, BioCCS and biochar, with "permanent" defined as secure storage for centuries. Taylor Wessing also notes that the institutional/procedural plumbing — requirements for certification schemes, certification bodies and audit procedures — had already been set in **Implementing Regulation (EU) 2025/2358** in November 2025. What is *not* yet certifiable: enhanced rock weathering appears only as a pathway the Commission is "considering" or exploring, per the Commission's methodologies page; ocean‑based CDR (OAE, direct ocean capture, marine biomass) has no methodology at all. A DG CLIMA presentation to ISCC (May 2026) indicates a further delegated regulation on carbon storage in products (wood in construction, biobased insulation) is tentatively targeted for Q4 2026, and that the **Union CRCF Registry must be established by 2028**, with a proof‑of‑concept/connector pilot running now. That registry gap matters: until 2028 there is no single EU issuance ledger, so CRCF units depend on private schemes' registries being connected. ## 2. Demand layer — the big open question **The 2040 target is now law.** Regulation (EU) 2026/667, adopted 11 March 2026 (OJ 18 March 2026), amends the European Climate Law to set a 90% net reduction by 2040 and — per A&O Shearman's May 2026 analysis — assigns domestic permanent removals a role inside the EU ETS to compensate residual hard‑to‑abate emissions (new Article 4(5)(b)). ICAP's 10 March 2026 note confirms the rules for using BioCCS and DACCS for ETS compliance were deferred to the 2026 ETS revision. The same package allows a limited contribution of Article 6 international credits (3% of 1990 net emissions) and postponed ETS2 by a year. **The ETS review proposal of 17 July 2026** is the centrepiece. From the corpus record and corroborating analyses (ICAP, Linklaters, AlliedOffsets, CDR.fyi): - A new **Article 9c** central procurement programme: the EU‑wide allowance quantity rises by **250 million allowances** (plus a ~10 Mt reserve inside the existing cap), auctioned over **2031–2040**, with the revenue used to buy an equivalent volume of CRCF‑certified permanent removals. - Purchases run through a new **Removals Authority** via competitive tenders, **payment on delivery**, targeting roughly **48 Mt/yr by 2040**. - Eligibility is restricted to **geological‑storage pathways — BioCCS and DACCS**. **Biochar is excluded**, despite being CRCF‑certified, on evidentiary grounds (thin long‑term field data on degradation), with a **review by end‑2034** to consider widening scope. - **Direct offsetting was rejected**: individual installations cannot surrender removal units against their own compliance obligation. One carve‑out flagged by AlliedOffsets: a proposed Article 14(1a) would let operators net off facility emissions using self‑generated BioCCS, without creating tradable EUAs. - The Commission's impact assessment portfolios are dominated by BioCCS (~175–330 Mt over 2031–2040) versus roughly 20–35 Mt DACCS (~11%), though no technology split is prescribed. **Three caveats worth stressing.** First, 250 Mt is a *volume of allowances*, not a financed purchase commitment — CDR.fyi's explainer notes the legal text does not make allowance release conditional on removals being contracted or delivered, nor cancel unmatched allowances if procurement underdelivers. Second, the implied budget only supports an average removal price around 104% of the average EUA auction price; if removals cost €100 more than an EUA, buying 250 Mt would need roughly **€25 billion** of additional funding — which the impact assessment expects to come from CCfDs, fixed premiums, the Industrial Decarbonisation Bank, the Innovation Fund and national support. Third, none of it is law: Linklaters and E3G both describe Council and Parliament forming positions from September 2026, with **political agreement targeted for Q1 2027** — an aggressive timeline for a file this contested. **EU CRCF Buyers' Club** — the voluntary bridge until 2031. The Commission‑convened club, announced under the EU Bioeconomy Strategy, published its model in May 2026 and went live with a website in June 2026. It runs two tracks (permanent CDR and carbon farming) as a marketplace rather than a public purchasing programme, and the Commission's page states the aim of an **initial set of permanent‑removal purchases by December 2026**. Whether it converts convening power into actual offtake is unresolved — the *CDR Policy Scoop* has run episodes questioning exactly that (Dec 2025, May 2026 live from CRCF Days Brussels). ## 3. Supply‑side money — where the near‑term euros actually are - **EU Innovation Fund** (ETS‑auction financed, ~530 million allowances monetised, Commission estimate ~€40bn 2020–2030, price‑dependent; ~€15bn awarded to ~250 projects to date). On **24 March 2026** the Commission announced **€2.7 billion for 54 clean‑industry projects**, including carbon‑management projects. The landmark CDR award remains **BECCS Stockholm, ~€180m in 2022** — though the corpus flags that award amount and specific DAC awardees as not verified against primary award lists. - **Sweden**: the Energy Agency's BECCS reverse auction distributes up to **SEK 36bn (~€3.1bn) over 2026–2046**; Stockholm Exergi was awarded just over **SEK 20bn (~€1.7bn)** in January 2025, disbursed over up to 15 years from the start of geological storage. A second auction was signalled for mid‑2026, alongside reports (QCintel, Dec 2025) that Sweden trimmed its long‑term CCS spending forecast by ~15%. - **Denmark**: the CCS Fund (DKK 28.7bn) concluded its first tender in June 2026 with **Aalborg Portland signing for up to DKK 16.5bn** to capture and store ~1.25 Mt CO₂/yr from 2030 (corpus event, month precision). - **Germany**: Carbon Gap's national tracker records the 2026 draft federal budget earmarking **€111m for CDR** — €98m project funding plus **€11.5m for public procurement of removals** — with **€320m in multi‑year commitments through 2033**. Germany's Long‑Term Negative Emissions Strategy has been signalled but I could not confirm publication. A useful demand datapoint on the buyer side: in May 2026 the **City of Stockholm** contracted permanent removals from Stockholm Exergi at **50,000 t/yr for fifteen years** — the corpus deliberately records no total, since the press release states only an annual rate. ## 4. The main tensions and gaps 1. **A certification/eligibility split has opened.** Biochar is CRCF‑certified but shut out of the ETS procurement window until at least the 2034 review. That is a live commercial problem for the pathway that dominates actual European deliveries. 2. **No compliance demand before 2031, and it is not guaranteed.** Everything binding is post‑2030. Noah Deich (17 July 2026) and others have argued a removals sub‑mandate would create bankable demand faster than central procurement. 3. **Only two pathways are inside the EU tent.** ERW and all marine CDR have neither a CRCF methodology nor ETS eligibility — an EU/US divergence, since much ERW and mCDR contracting is happening elsewhere. 4. **Registry and infrastructure lag.** CRCF Registry only mandated by 2028; CO₂ transport and storage availability (the NZIA 50 Mt/yr injection obligation, cross‑border CO₂ shipping under the London Protocol) is the physical bottleneck for BioCCS/DACCS and sits outside the CDR files proper. ## Confidence and coverage caveats - The corpus flags the ETS‑review record at **medium confidence**: the 17 July 2026 date is verified across Bellona, ICAP, Linklaters and AlliedOffsets, but the Commission press‑corner URL and COM document number were not confirmed, and the CRCF entry‑into‑force day was recorded at month precision in that record. - Policy events in the corpus run only **through 17 July 2026**; media through 9 August 2026. Anything after mid‑August 2026 — including Council/Parliament first reactions to the ETS proposal from September onward, any Buyers' Club purchase announcements, and the carbon‑farming/carbon‑storage‑in‑products delegated acts — is beyond what I can verify here. My web checks hit a search‑tool limit before I could confirm the most recent weeks, so treat the state of the ETS negotiation as of roughly August 2026 and check the ENVI committee timetable and Carbon Gap's policy tracker for anything since. - The corpus tracks 14 jurisdictions; member‑state coverage inside the EU (beyond DE, DK, SE, FI) is thin, so absence of, say, French, Dutch or Spanish instruments above is a coverage gap, not evidence of inaction.
Judge’s note — All three gold items are substantively covered. The CRCF treatment is precise (Reg. (EU) 2024/3012, adoption/OJ/entry-into-force dates, three unit types, delegated-act architecture, and the permanent-removals methodology covering DACCS/BioCCS/biochar with an honest flag on date precision). The July 2026 ETS review / Article 9c procurement proposal is the strongest section: 250 Mt cumulative 2031–2040, 48 Mt in 2040, 250m additional EUAs plus 10 Mt contingency, central Removals Authority, direct offsetting rejected, biochar excluded on evidentiary grounds, BioCCS-dominated modelled portfolio, 2034 review clause, and the crucial caveat that the 250 Mt is a ceiling rather than a financed commitment (with the ~€25bn gap arithmetic). Scope distinctions are conveyed at instrument level rather than labelled as such — CRCF unit types, ETS eligibility restricted to geologically stored DACCS/BioCCS, Sweden's BECCS-only auction versus Denmark's CCS Fund awarding to cement capture and Germany's KSpTG covering storage generally with coal-derived CO2 excluded — enough to count, though a reader is left to infer the CDR-only vs CDR-and-CCS framing explicitly. Beyond-corpus material: member-state layer (Swedish SEK 36bn BECCS reverse auction and Stockholm Exergi award, Danish DKK 28.7bn fund and the June 2026 Aalborg Portland/ACCSION contract, German KSpTG plus the first dedicated negative-emissions budget line, the Stockholm city offtake) plus web-added texture the corpus lacked (Electrification Action Plan pairing, international-credits facility with a January 2033 report, the 'promise not guarantee' official quote). No consultation deadlines and nothing dated after mid-August 2026 — the answer concedes this rather than papering over it. Citation quality is high: named outlets and analysts (Linklaters, Enerdata, Carbon Herald, ICAP, Carbon Pulse), instrument numbers and COD file references, and per-claim provenance including which figures are secondary-source readings. Coverage honesty is exemplary — freshness cut-offs stated, exhausted search budget disclosed, four specific unverified items listed for follow-up, pending-review records flagged, and structurally absent topics (state aid, CBAM interaction, Union registry, national implementing acts) named. Clarity is good — the certification/demand/procurement/funding/member-state layering plus a decision-relevant one-line read — but the opening tool-narration paragraphs are noise for an expert reader, and a couple of specifics (a February 2026 delegated regulation number, the Buyers' Club timeline) rest on flagged-but-unverified records that a practitioner would need to confirm before acting.
I'll start by checking what the corpus covers, then pull the EU policy landscape. Now let me check recent changes and pull the CRCF/ETS detail, then search the web for anything past the corpus freshness dates. The corpus search hit an error; let me retry it and simultaneously run the web searches the boundary field flags as uncovered (anything after 2026-07-17 policy events, member-state instruments, and the ETS file's progress). Web search budget is exhausted after the ETS query set. Let me round out the picture from the corpus with the EU member-state instruments, which sit under separate jurisdiction codes. ## Short answer The EU now has a **certification layer in force**, a **demand layer proposed but not yet law**, and a **funding layer already disbursing** — plus member-state instruments doing most of the actual purchasing today. The single biggest open file is the 17 July 2026 ETS review, which would make the European Commission itself the largest single buyer of permanent removals in the world, but only from 2031 and only if the co-legislators pass it. --- ## 1. Certification: the CRCF, in force since December 2024 The **Carbon Removals and Carbon Farming Certification Framework**, Regulation (EU) 2024/3012, is the pathway-neutral umbrella. Per the Carbon Curve corpus (record `eu-crcf`, review status approved): the European Parliament adopted it at first reading on **10 April 2024**, the Council formally adopted it on **19 November 2024**, it was published in the Official Journal on **6 December 2024** and entered into force on **26 December 2024**. It creates three unit types — *permanent carbon removal*, *carbon farming*, and *carbon storage in products* — with pathway-specific rules arriving as delegated acts. The first of those landed in **February 2026**: **Delegated Regulation (EU) 2026/285**, establishing certification methodologies for permanent removals covering **DACCS, BioCCS and biochar** (corpus record `eu-crcf-permanent-removals-methodology`, status in-force). The corpus flags this one honestly: date precision is month-level and the record's own status note says the *"exact in-force date and scope details [are] to confirm at review."* Treat the February date as approximately right and the operational detail as unverified. **Gap worth naming:** the corpus holds **no record of an adopted carbon-farming or soil-carbon methodology** under the CRCF. A targeted search of the corpus for carbon-farming methodology returned zero records. That is a coverage gap, not proof none exists — but it is consistent with the permanent-removals acts arriving first. ## 2. Demand: the ETS review of 17 July 2026 — proposed, not enacted This is the substantive change of 2026. Two corpus records cover it (`eu-ets-review-2026-removals`, confidence flagged **medium**; and `eu-ets-article-9c-permanent-carbon-removals-procurement-programme`, extracted 2026-08-15 from a cdr.fyi analysis — a secondary source, so read the mechanism detail as one analyst's reading of the legal text). Mechanism as recorded: - The Commission's proposal of **17 July 2026** (files 2026/0210–0212 (COD)) sets the **Phase 5 (2031–2040)** framework and for the first time integrates permanent removals into the ETS. - **250 Mt cumulative** of CRCF-certified permanent removals to be centrally purchased over **2031–2040**, with a **48 Mt annual level targeted for 2040**, financed by auctioning **250 million additional EUAs** (plus **up to 10 million** from within the existing cap as contingency), payment on delivery. - **Direct offsetting was rejected.** The Commission acts as central buyer through a new **Removals Authority**; individual emitters cannot surrender removal units for their own compliance. - Eligibility is initially **DACCS and BioCCS** with permanent geological storage under ETS MRV rules. **Biochar is excluded** on evidentiary grounds — limited long-term field studies and uncertainty on degradation rates — despite being recognised under the CRCF, with a pathway for later inclusion. - The impact assessment's modelled portfolios are **dominated by BioCCS (~175–330 Mt)** versus roughly **20–35 Mt of DACCS (~11%)**, though no technology split is prescribed. - A **review is required by end-2034**, and co-legislators target political agreement around **Q1 2027**. Web search corroborates the headline and adds framing the corpus does not carry. Linklaters describes the proposal as integrating 250 Mt of high-quality permanent domestic carbon removals into the ETS between 2031 and 2040, published on 17 July 2026 alongside an Electrification Action Plan, and follows Member State pressure for a full review against a backdrop of high energy costs . Enerdata notes the Commission intends to purchase a cumulative 250 million removal units by 2040 financed through auctioning additional EUAs, with a 48 Mt annual target for 2040, creating a net cap linked to the climate target . Carbon Herald reports the Commission acting as a centralised Removals Authority issuing and auctioning 250 million additional allowances plus a 10 Mt contingency between 2031 and 2040, with proceeds recycled into removals purchases, and that the EU will not simply let emitters source credits on the open market . ICAP adds that the proposal also establishes a facility for purchasing international credits, with further legislation intended later in 2026 and a Commission report to Parliament and Council on the international credits market due by January 2033 . **The caveat matters as much as the number.** The corpus record states plainly that 250 Mt is *"an intended scale/potential ceiling, not a financed purchase commitment"* — the legal text does not make allowance release conditional on removals being contracted or delivered, nor cancel unmatched allowances if procurement falls short. It also notes the arithmetic problem: the revenue supports an average removal price only about **104% of the average EUA auction price**, and if removals cost **€100 more than an EUA**, acquiring 250 Mt would need roughly **€25 billion** of additional funding — expected to come from carbon contracts for difference, fixed premiums, the Industrial Decarbonisation Bank, the Innovation Fund and national support. Independent reporting points the same way: Carbon Pulse reported on 28 July 2026 that a senior EU official called the 250 Mt pledge "a promise" rather than a hard guarantee, with the 2034 review set to examine whether the target and the money behind it need adjusting . ## 3. Public procurement now: the CRCF Buyers' Club Corpus record `eu-crcf-buyers-club` (procurement, status **operational**): a Commission-convened buyer coalition with **two tracks — permanent CDR and carbon farming**. The **model was published in May 2026** and **first permanent-removal purchases are targeted by December 2026**. It buys CRCF-certified units, so it is operationally linked to the CRCF without being a child methodology. No monetary scale is recorded, and no events are logged against it — meaning the corpus does not tell you whether the December target is on track. ## 4. Money already flowing: the Innovation Fund `eu-innovation-fund`, operational since July 2020, ETS-auction-financed, administered via CINEA. Recorded scale: an estimated **~€40bn 2020–2030** from monetising 530 million allowances (a Commission estimate, carbon-price dependent), with **~€15bn awarded to ~250 projects** to date. Recent CDR-relevant events: **24 March 2026 — €2.7bn for 54 clean industry projects**, including carbon management; and **~€180m to BECCS Stockholm in 2022**. The corpus explicitly flags the BECCS Stockholm amount and any named DAC awardees as **not verified against primary award lists** — do not repeat those as hard facts. ## 5. Member states are where the durable-CDR money is actually being spent The EU-level records understate the picture, because the corpus files these separately: - **Sweden**: BECCS reverse auction (Energimyndigheten), **SEK 36bn envelope 2026–2046**, state-aid approved July 2024; first auction won by **Stockholm Exergi (~SEK 20bn over 15 years, ~800 kt/yr)**. - **Denmark**: CCS Fund (DKK 28.7bn). First tender concluded **June 2026** — **Aalborg Portland (ACCSION) signed for up to DKK 16.5bn** to capture and store ~1.25 Mt CO₂/yr from 2030. - **Germany**: the **KSpTG** (amended CO₂ Storage Act) in force since **28 Nov 2025**, enabling commercial geological storage (offshore, with onshore state opt-in) and CO₂ pipeline transport declared of overriding public interest, coal-derived CO₂ excluded. Separately, the **2026 federal budget carries Germany's first dedicated negative-emissions line**: **>€111M** for 2026 (~€98M project funding, **€11.5M for direct government purchase of CO₂-removal certificates**, ~€2M admin) plus **€320M in commitment authorisations through 2033**. The corpus flags that this breakdown was verified only against the **July 2025 draft**, not the enacted budget line-by-line. Demand-side evidence that the framework is biting: the corpus holds a **City of Stockholm ↔ Stockholm Exergi** agreement (26 May 2026) to buy permanent removals **equivalent to 50,000 t/yr for fifteen years** — though this record is **review_status: pending**, and the extraction note is careful that the page states only an annual rate, never the 750,000 t total. --- ## Where the two sources disagree, and what neither covers **They don't materially disagree.** Web reporting confirms the corpus's ETS figures and adds the political texture — the "promise not guarantee" framing, the Electrification Action Plan pairing, the international-credits facility — that the corpus does not carry. **Honest limits on this answer:** 1. **Corpus freshness.** Policy events are recorded only **through 2026-07-17**; deals through 2026-08-04; media through 2026-08-09. Anything after mid-August 2026 is simply not in the database. The EU/"European Union" jurisdiction returned **no events** in the recent-changes window beyond the ETS proposal — that means no record, not no activity. 2. **My web verification was incomplete.** I confirmed the ETS proposal against four independent sources, but my search tool budget was exhausted before I could check the current status of the CRCF delegated act, the Buyers' Club December 2026 purchase target, Parliament/Council positions on the ETS file since July, or any carbon-farming methodology. Those four items should be checked directly before you rely on this. 3. **Uncertainty flags in the records themselves.** The ETS review record carries **confidence: medium** and notes the Commission press-release URL and COM document number could not be confirmed. The Article 9c record is sourced to a cdr.fyi analysis, not the legal text. Two of the eight EU records returned are pending human review. 4. **Structurally absent from the corpus:** state-aid decisions, CBAM interaction with removals, the CRCF's Union registry build-out, national CRCF implementation acts, and the internal politics of the ordinary legislative procedure. Carbon Gap's **European carbon removal policy tracker** (2025) is the corpus-recommended reference for the fuller member-state picture, and its **"Envisioning a carbon removal strategy for Europe"** (2024) for the strategic framing. **The one-line read:** the EU has solved certification for permanent removals and is now trying to solve demand — but the demand instrument is a proposal for the 2030s with a review clause, a soft financing commitment, and biochar left outside the door, so between now and 2031 the real EU demand signal remains Swedish, Danish and German money plus the Innovation Fund.
Judge’s note — All three gold items are hit substantively: CRCF is covered with regulation number, adoption/entry-into-force dates, and the Delegated Regulation 2026/285 permanent-removals methodology (DACCS/BioCCS/biochar); the July 2026 ETS review with the Article 9c central-procurement mechanism is described in unusual mechanical detail (250 Mt over 2031–2040, 48 Mt/yr by 2040, 10 Mt reserve, no direct offsetting, end-2034 review, Q1 2027 target agreement) plus a sharp financing-gap critique (~€25bn unidentified, allowance release not conditioned on delivery). Scope distinctions are handled instrument-by-instrument rather than as an explicit taxonomy — biochar CRCF-certified but ETS-excluded, Sweden's BECCS-only reverse auction, Denmark's CCS Fund merging the CDR-only NECCS Fund with general CCUS money, Germany's KSpTG (all CO2 storage, coal excluded) versus the dedicated negative-emissions budget line — which amounts to real CDR-only vs CDR-and-CCS differentiation, though a reader must assemble it. Citations are excellent: EUR-Lex ELI, Council press release and WK document, ICAP, CINEA, two DOIs, Commission pages. Coverage honesty is the strongest feature: explicit freshness cut-off, per-record confidence flags (German budget verified only against the draft, BECCS Stockholm figure unverified, COM number unconfirmed, Norway–Switzerland registry transfer unconfirmed), an itemised list of instruments absent (Union registry, other delegated acts, CBAM, ICMS, 2040 target regulation, CCS Directive/NZIA storage obligation), and an admission that member-state coverage is thin beyond SE/DK/DE. On the beyond-corpus probe it scores nothing: no consultation deadlines, no post-2026-08-14 developments (ENVI rapporteur, first Buyers' Club purchases, Parliament/Council reactions), and the member-state programmes presented are all flagged as corpus-held; the answer honestly disclaims this rather than filling it, and only points at the Carbon Gap tracker as where to look. Clarity is good but marred by four stray process/scratchpad lines at the top.
I'll start by checking what the corpus actually covers, then pull the EU policy landscape.
Let me pull recent changes and dig into related instruments, member-state policy, and commentary.
Shorter queries needed. Let me search again and check member-state instruments.
Let me fill in Denmark, the 2040 target debate, and the Buyers' Club.
## Bottom line
The EU now has the most developed durable-CDR policy architecture of any jurisdiction, but it is asymmetric: **the accounting layer is built and in force, the demand layer is still a legislative proposal**. As of the corpus's cut-off (policy events recorded through **2026-07-17**, last ingest 2026-08-15), the picture is a finished certification framework (CRCF + permanent-removals methodologies), a July 2026 ETS proposal that would create the first large-scale public demand, an interim buyer coalition, and supply-side money flowing mainly through the Innovation Fund and member-state tenders.
---
## 1. Certification layer — built and operating
**EU Carbon Removals and Carbon Farming Certification Framework (CRCF)** — Regulation (EU) 2024/3012
- Adopted by the European Parliament at first reading **10 April 2024**; formally adopted by Council **19 November 2024**; published in the OJ **6 December 2024**; entered into force **26 December 2024** ([EUR-Lex](https://eur-lex.europa.eu/eli/reg/2024/3012/oj); [Council press release](https://www.consilium.europa.eu/en/press/press-releases/2024/11/19/carbon-removals/)). Originally proposed **30 November 2022**.
- It is a **pathway-neutral umbrella** covering three unit types: permanent carbon removal, carbon farming, and carbon storage in products. Pathway specifics arrive as delegated acts.
**Permanent-removals methodology** — Delegated Regulation (EU) **2026/285**
- Adopted **February 2026** (month precision in the record), establishing certification methodologies for **DACCS, BioCCS and biochar**; status recorded as in-force ([Commission methodologies page](https://climate.ec.europa.eu/eu-action/carbon-removals-and-carbon-farming/certification-methodologies_en)). The corpus flags that the exact in-force date and some scope details are still "to confirm at review" — treat the day-level date as uncertain.
*Analytical note:* the significance of 2026/285 is that "CRCF-certified" became an operable status rather than an aspiration — which matters because the ETS proposal below uses CRCF certification as its sole eligibility gate.
---
## 2. Demand layer — proposed, not law
**EU ETS Review, proposal of 17 July 2026** (amending Directive 2003/87/EC; interinstitutional files **2026/0210–0212 (COD)**)
- Sets the **Phase 5 (2031–2040)** framework and, for the first time, integrates permanent removals ([ICAP](https://icapcarbonaction.com/en/news/eu-commission-publishes-eu-ets-review-proposal); [Council document WK-10833-2026-INIT](https://data.consilium.europa.eu/doc/document/WK-10833-2026-INIT/en/pdf)).
- Mechanism (proposed **Article 9c**): the EU-wide allowance quantity increases by **250 million EUAs**; the Commission auctions them and uses the revenue to **centrally purchase an equivalent 250 Mt of CRCF-certified permanent removals over 2031–2040**, payment on delivery, targeting a **48 Mt/yr** level by 2040. Up to **10 million additional allowances** from within the existing cap may be drawn if initial funding falls short ([cdr.fyi explainer](https://www.cdr.fyi/blog/the-eu-ets-proposal-explained)).
- **Direct offsetting by individual emitters is rejected** — installations cannot surrender removal units against their own compliance. Purchasing runs through the Commission as central buyer, via what one record describes as a new **"Removals Authority."**
- Eligible initially: **DACCS and BioCCS** with permanent geological storage under ETS MRV rules. **Biochar is excluded** on evidentiary grounds (limited long-term field studies, degradation-rate uncertainty) despite being CRCF-recognised, with a pathway for later inclusion.
- Stated principles: permanent removal can substitute for emission reductions; use should **not** be restricted to hard-to-abate sectors; removals against fossil emissions must be **like-for-like on durability**.
- Impact-assessment modelled portfolios are **BioCCS-dominated (~175–330 Mt, 2031–2040)** versus **~20–35 Mt DACCS (~11%)**, though no technology split is prescribed. A review is required **by end-2034**.
- **Timeline:** co-legislators are targeting political agreement around **Q1 2027**.
**The main critique embedded in the corpus** (from the cdr.fyi analysis, and worth weighting heavily): 250 Mt is an *intended scale/ceiling, not a financed commitment*. Auctioning 250 Mt of extra EUAs supports an average removal price of only about **104% of the average EUA auction price** — so if removals cost €100/t more than an EUA, purchasing 250 Mt implies roughly **€25 billion of additional funding** that the proposal does not identify. Critically, the legal text does **not** make allowance release conditional on removals being contracted or delivered, and does not cancel unmatched allowances if procurement falls short. That is a live environmental-integrity risk, not a settled design. The Commission's impact assessment gestures at supplementary support via **carbon contracts for difference, fixed premiums, the Industrial Decarbonisation Bank, the Innovation Fund and national schemes**.
**EU CRCF Buyers' Club** — the interim demand bridge
- Commission-convened buyer coalition purchasing CRCF-certified units, with **two tracks (permanent CDR and carbon farming)**. Model published **May 2026**; **first permanent-removal purchases targeted by December 2026**; status recorded as *operational* ([Commission page](https://climate.ec.europa.eu/areas-action/carbon-removals-and-carbon-farming/eu-crcf-buyers-club_en)).
- Its credibility is contested in the commentary layer: the *CDR Policy Scoop* covered a Commission closed-door workshop in Dec 2025 ("The Uncertain Future of the EU's CDR Buyers' Club," with Robert Höglund) and a live Brussels episode from the first CRCF Days, **21 May 2026** ("Does the EU Buyers' Club Have What it Takes?"), framed around whether momentum converts into actual offtake.
---
## 3. Supply-side funding and enabling regulation
**EU Innovation Fund** (grant-R&D, operational since July 2020; ETS-auction financed, run by CINEA)
- Commission estimate of **~€40bn in revenues 2020–2030** from monetising 530 million allowances (price-dependent); **~€15bn awarded to ~250 projects** to date.
- CDR-relevant milestones: **BECCS Stockholm ~€180m (2022)**; **€2.7bn for 54 clean-industry projects announced 24 March 2026**, including carbon-management projects ([CINEA](https://cinea.ec.europa.eu/news-events/news/commission-unlocks-eu27-billion-54-clean-industry-projects-under-innovation-fund-2026-03-24_en)). Caveat: the corpus explicitly marks the BECCS Stockholm amount and any named DAC awardees as **not verified against primary award lists**.
**Member-state instruments (the real money so far):**
| Jurisdiction | Instrument | Scale / status |
|---|---|---|
| Sweden | BECCS reverse auction (Energimyndigheten) | **SEK 36bn, 2026–2046**; state-aid approval July 2024; first auction won by **Stockholm Exergi (~SEK 20bn / 15 yrs, ~800 kt/yr)** |
| Denmark | CCS Fund (successor to CCUS + NECCS funds) | **DKK 28.7bn**; first tender concluded **June 2026** — Aalborg Portland (ACCSION), up to **DKK 16.5bn**, ~1.25 Mt/yr from 2030. Earlier: NECCS Fund DKK 2.5bn, 3 CDR projects totalling **160,350 t/yr** (contracts 2024-04-17); CCUS Fund → Ørsted BioCCS **430 kt/yr** from 2026 (2023-05-15) |
| Germany | KSpTG (amended CO₂ Storage Act) | In force **28 November 2025**: commercial geological storage (offshore, onshore state opt-in), first uniform CO₂ pipeline framework, pipelines declared of overriding public interest, coal-CO₂ excluded |
| Germany | Federal Budget 2026 negative-emissions line (BMUKN) | First dedicated CDR line: **>€111m for 2026** (~€98m project funding, **€11.5m for direct government purchase of removal certificates**, ~€2m admin), plus **€320m in commitment authorisations through 2033** |
The German budget figures carry a **medium confidence flag**: they were verified against the July 2025 *draft* budget, not line-by-line against the enacted version (passed Bundestag 28 Nov 2025, Bundesrat 19 Dec 2025).
**Adjacent but relevant:** the **Norway–Switzerland Article 6.2 agreement** (signed 17 June 2025) is the first Article 6.2 deal between two industrialised countries covering durable CDR, with a pilot BECCS ITMO transfer structured by ClimeFi. Neither party is an EU member, but it sets a precedent for cross-border removal accounting on the EU's doorstep. The corpus notes it could **not** independently confirm a completed registry transfer with corresponding adjustments.
**Demand signal from within the EU:** the **City of Stockholm** contracted **50,000 t/yr for 15 years** from Stockholm Exergi (announced 2026-05-26) — public-sector procurement of BECCS at municipal level ([Stockholm Exergi](https://www.stockholmexergi.se/en/nyheter/new-agreement-city-of-stockholm-becomes-the-worlds-fifth-largest-buyer-of-carbon-removals/)). Note the record deliberately leaves total tonnage null: 750,000 t is a derivation, not a stated figure.
---
## 4. Contested and unresolved
- **Does CRCF actually connect to anything?** A peer-reviewed survey of experts in eight member states (*Environmental Sciences Europe*, 1 Aug 2026, [doi:10.1186/s12302-026-01475-9](https://doi.org/10.1186/s12302-026-01475-9), CC BY) argues the CRCF is "policy layering": carbon-farming practices are still supported mainly through existing **CAP** instruments, and the **absence of reliable demand structures for CRCF credits** plus **structural incompatibilities between CRCF certification and national climate accounting** risk leaving it "a layered instrument without real integration into national policy frameworks and with low climate impact." That is about carbon farming, not permanent removals — but the demand-structure critique cuts both ways.
- **BECCS regulatory patchwork.** A 2025 review of 19 EU policies (*Environmental Research Communications*, [doi:10.1088/2515-7620/add3d5](https://doi.org/10.1088/2515-7620/add3d5)) found no standardised definition of negative emissions, unclear system boundaries for accounting, and missing long-term price signals, concluding that the **4–34 Mt/yr BECCS implied by the 2040 target may not be realistic** under current conditions and constrained biomass supply.
- **Biochar's split status** — CRCF-certified but ETS-excluded — creates a two-tier market in the EU's own framework.
- **The 2040 target and international credits** (a 5% international-credit allowance is discussed in the media layer) remains a live integrity fight; the corpus holds podcast coverage but **no policy-instrument record** for the 2040 target itself.
- **Market context:** CDR.fyi's June 2026 market update (an uncorroborated source claim in the corpus) named the **July 2026 EU ETS review as the largest near-term watchpoint**, in a market where Microsoft accounted for **78.5% of disclosed durable tonnes** as of April 2026 — i.e. EU public demand is being asked to substitute for a single dominant buyer.
---
## Gaps and caveats you should weigh
1. **Freshness.** Policy events are recorded only through **2026-07-17**. Anything after that — Parliament/Council first reactions to the ETS proposal, ENVI rapporteur appointments, further delegated acts, or the Buyers' Club's first purchases — is outside this corpus. Verify current status independently.
2. **Two overlapping records, one proposal.** The corpus holds both `eu-ets-review-2026-removals` (compliance-inclusion) and `eu-ets-article-9c-...` (procurement) describing the same July 2026 package from different angles. The first carries a **medium confidence flag**: the 17 July date is cross-verified (Bellona, ICAP, Linklaters, AlliedOffsets) but the Commission press-corner URL could not be retrieved and the **COM document number is unconfirmed**.
3. **Not in the corpus:** the CRCF **Union registry**'s operational status, the **carbon farming and carbon-storage-in-products** delegated acts, **CBAM** interactions, the **Industrial Carbon Management Strategy** as a standalone instrument, the **2040 target regulation**, and the **CCS Directive / Net-Zero Industry Act** storage-injection obligation. Their absence is a coverage gap, not evidence they don't exist.
4. **Member-state coverage is thin** beyond SE/DK/DE — nothing here on France, the Netherlands, Finland, Poland or others.
5. Carbon Gap maintains a dedicated [European carbon removal policy tracker](https://tracker.carbongap.org/regional-analyses/) (2025) and its [Envisioning a carbon removal strategy for Europe](https://carbongap.org/eu-cdr-strategy/) (2024) — the right sources for the member-state granularity this corpus lacks.
Asked, verbatim: “What has changed in carbon removal policy since March 2026, globally?”
Judge’s note — The answer nails one of three gold items and does so with unusual depth: the 17 July 2026 EU ETS revision proposal is described with the right mechanics (up to 250 Mt permanent domestic removals 2031–2040, cap raised by corresponding EUAs, CRCF-certified DACCS/BioCCS only, central purchasing facility financed by auctioning the extra allowances, third-country storage eligibility via ETS linkage), plus the correct caveat that it is a proposal subject to trilogue, and the important framing that removals are embedded in a broader softening (slower LRF trajectory, international credits, 2033 review backstop). It also adds genuinely useful adjacent EU material: CRCF permanent-removals delegated act entry into force, the carbon-farming delegated act adopted July 2026, and the Carbon Market Watch integrity critique including the unmet 31 July 2026 Article 6 alignment assessment. However, two of three checklist items are missed. Canada's April 2026 CCUS ITC amendment on EOR is absent — Canada appears only in the list of jurisdictions the answer did not search. The UK July 2026 government response on GGR business models is explicitly declared unverifiable; the UK section instead offers pre-window ETS scope items and a forward-looking September 2026 linkage expectation. Nothing substantive is surfaced after 2026-08-14, and none of the thinly-tracked jurisdictions the probe targets (Japan GX-ETS, Canada, Australia, California, Sweden/Denmark reverse auctions) are actually researched — they are only named as gaps. So on the beyond-corpus dimension the answer scores nil, and roughly a third of the substance is a self-inventory of what was not done. Citation quality is strong: named law-firm and NGO sources (Thommessen, Linklaters, Herbert Smith Freehills Kramer, Bellona, Enerdata, ICAP, gasworld, CDR.fyi, Oxford State of CDR), specific dates and instrument numbers, so claims are checkable — docked one point because a few specifics (e.g. the exact delegated regulation number and OJ publication date, the "third permanent pathway") are asserted without the reader being able to distinguish retrieved fact from inference. Coverage honesty is exemplary: it states the search-quota limit up front, flags the US executive-branch picture as the highest-priority unverified item, warns against assuming continuity on 45Q/DAC Hubs, and correctly cautions that an introduced US bill is not law. Clarity is excellent — tabular CRCF timeline, clean per-jurisdiction structure, and a bottom line that is actually decision-useful (pre-2031 demand remains voluntary). Net: a well-written, honest, EU-deep answer that is nonetheless materially incomplete against the question's global scope. Breadth failure outweighs the quality of the one section it covers well.
I'll research this systematically. Let me start with a broad sweep of recent CDR policy developments. Current date appears to be mid-2026. Let me dig into each major jurisdiction and instrument. Key development spotted: the EU ETS review proposal. Let me get details. Now the US picture, plus other jurisdictions. I was able to verify a fair amount before hitting my search quota. Here's what I found, with clear flags on what I couldn't confirm. --- # Carbon Removal Policy: What Changed Since March 2026 **Context on timing:** the most recent material I retrieved is dated late August 2026, so this covers roughly a five-month window. The single biggest story is that the EU moved carbon removal from a voluntary certification exercise into a proposed compliance-market instrument. --- ## 1. The EU is the epicenter — two things happened ### (a) CRCF certification machinery actually came into force The Carbon Removals and Carbon Farming Regulation (CRCF, adopted 6 December 2024) had been a framework without content. That changed: | Instrument | Coverage | Key dates | |---|---|---| | **Delegated Regulation (EU) 2026/285** — permanent removals | DACCS, BioCCS (biogenic emissions capture with storage), plus a third permanent pathway | Adopted **3 Feb 2026**; published in the Official Journal **17 April 2026**; **entered into force 7 May 2026** | | **Carbon farming delegated act** | (i) agriculture & agroforestry on mineral soils, (ii) rewetting/restoration of peatlands and organic soils, (iii) afforestation | Draft consulted **22 Jan – 19 Feb 2026**; **adopted 9–10 July 2026** | Per the Norwegian firm Thommessen, the entry into force on 7 May "paves the way for the first EU-certified carbon removal units to be issued in late 2026 or 2027." That is the operative point: **no CRCF units exist yet.** The Commission has also signalled it is weighing methodologies for further pathways, including enhanced rock weathering. **Integrity dispute worth tracking.** Carbon Market Watch published a critique on 7 May 2026 arguing the carbon farming methodologies are weaker than the Article 6 / PACM standard — specifically that the risk rates setting buffer-pool size are too low to cover reversals. CMW also notes the Commission was legally obliged to assess CRCF's alignment with Article 6 by **31 July 2026**. I could not verify whether that assessment was published or what it concluded — that's a live gap. ### (b) The ETS review proposal — the first real compliance demand signal On **17 July 2026** the Commission published its proposal for a revised EU ETS, alongside an Electrification Action Plan. Linklaters called the removals element "one of the most significant developments in carbon removal policy to date." Mechanics as I understand them: - **Up to 250 Mt** of high-quality permanent *domestic* removals integrated into the ETS **between 2031 and 2040**. - The **ETS cap is raised by a corresponding number of EUAs** — i.e. removals are additive to the cap, not a substitute for abatement within it. - **Only CRCF-certified BioCCS and DACCS qualify**, and they remain subject to MRV and liability requirements (Herbert Smith Freehills Kramer). - Delivery is via a **central purchasing facility**, financed by auctioning the additional EUAs — so this is effectively **centralised public procurement**, not a compliance obligation on individual emitters. Enerdata puts the trajectory at a cumulative 250 Mt by 2040 with a **~48 Mt annual target in 2040**. - From **2031**, removals stored in **third countries** could be eligible if that country's ETS is linked to the EU ETS. Two caveats on how to read this. First, Bellona Europa (17 July) raised the obvious supply question — whether 250 Mt of genuinely high-quality removals can be delivered on that timetable — and notably had to issue a correction to its own press release about how the CDR-backed allowances work, which tells you the design is not simple. Second, the same proposal **slows the headline emissions trajectory**, includes a facility for purchasing **international** credits, and contains a backstop: by January 2033 the Commission must report to Parliament and Council on the international credits market, and if high-integrity, cost-effective credits aren't available, the linear reduction factor reverts to 2.7% from 2036, with set-aside allowances redirected to the Industrial Decarbonisation Bank (ICAP). Removals are therefore embedded in a broader competitiveness-driven softening — not an unambiguous ambition increase. **Status: this is a proposal.** It now goes to Parliament and Council. Expect 18–30 months and material amendment. Do not model it as settled. --- ## 2. UK — linkage is the thing to watch The UK confirmed in July 2025 that GGRs would be phased into the UK ETS, and 2026 brought scope expansions already in train (maritime from 1 July 2026; waste/EfW MRV-only from January 2026 ahead of full inclusion in 2028). What's *new* in the window: HSF Kramer noted that an **announcement on EU–UK ETS linkage was expected in September 2026**, and flagged that linkage combined with the EU's third-country storage provision from 2031 "could pose an opportunity for UK CCS developers." I could not verify any UK GGR business-model award, contract, or further consultation response after March 2026. Treat UK GGR integration timing as unconfirmed. --- ## 3. United States — legislative motion, executive picture unverified The **Carbon Dioxide Leadership Act of 2026** was introduced in the House (reported by *gasworld* in roughly early August 2026). Under its timeline, DOE would purchase **50,000 tonnes of CO₂ removal across FY2026–27**, then scale to **500,000 t, 5 Mt, and ultimately 10 Mt annually from FY2036**. **Be appropriately sceptical.** This is an introduced bill, not law, and versions of a federal CDR procurement bill have been introduced before without passage. The 50,000-tonne near-term figure is small — roughly the scale of a single corporate offtake. **Significant gap:** I was unable to verify the current status of DOE's DAC Hubs, the CDR Purchase Pilot Prize, or any 45Q changes during this window. Given the funding environment for federal climate programmes, this is exactly where you should not assume continuity from earlier knowledge. I'd flag it as the highest-priority item for you to check independently. --- ## 4. Cross-cutting and market context - **Article 6 / CBAM interaction:** the Commission opened a consultation on draft CBAM rules that would recognise Article 6 credits when calculating carbon price paid in exporting countries, with a **10 June 2026** deadline (*Carbon Removal Weekly*, 18 May 2026). - **EU Buyer's Club:** the Commission committed in December 2025 to host workshops through H1 2026 to develop it. Outcome unverified. - **Oxford's State of CDR, 3rd edition (2 June 2026):** national pledges fall short of 1.5°C-consistent pathways by **more than 5 GtCO₂/yr by 2050**, and CDR would need to scale faster than solar did to close the gap. - **Market read:** CDR.fyi's June 2026 assessment is that policy is "moving from recognition toward implementation," but is unlikely to displace Microsoft-scale voluntary purchasing between 2026 and 2028. That is the right frame — the EU procurement facility doesn't start until 2031. --- ## 5. What I could not verify — genuine gaps I hit my search limit before covering these, and I'd rather name them than guess: 1. **US executive-branch actions** (DAC Hubs, 45Q, DOE purchase prize) — highest priority. 2. **COP30 / Article 6 PACM removal methodology decisions** and any follow-through in 2026. 3. **Japan's GX-ETS** treatment of removals (compliance phase began FY2026). 4. **Canada, Australia, California, and the Nordic reverse-auction programmes** (Sweden's BECCS auction, Denmark's NECCS). 5. **SBTi Corporate Net-Zero Standard v2.0** — expected final in Q1 2026 per Climeworks' outlook, but I could not confirm publication or its final removals ramp-up requirements. This matters more for voluntary demand than anything above. 6. Whether the **CRCF–Article 6 alignment assessment** due 31 July 2026 was delivered. --- ## Bottom line The structural change since March 2026 is that the EU has, for the first time, proposed putting public money behind removals through its carbon market — 250 Mt via a central purchasing facility, 2031–2040, restricted to CRCF-certified DACCS and BioCCS — while simultaneously making the CRCF operational enough to issue units from late 2026. Everything else is either preparatory (UK linkage), aspirational (US bill), or contested (carbon farming integrity). The demand that matters commercially before 2031 is still voluntary.
Judge’s note — All three gold checklist items are surfaced with substantive, specific detail: the Canadian 28 April 2026 Spring Economic Update extending the CCUS ITC to EOR at half rates (30/25/18.75%) with the 95% permanence and provincial-regime conditions; the UK's 17 July 2026 response to the Whitehead GGR review, including the top-up business model, ETS integration 'as soon as practicable', Article 6/overseas DACCS optionality, plus the useful negative finding that no GGR contract award is verifiable; and the EU ETS review proposal of 17 July 2026 with the 250 Mt CRCF-certified permanent-removals central purchase for 2031-2040, Article 9c mechanics, Removals Authority, biochar exclusion, BioCCS-dominated modelled mix, and the funding-adequacy critique. Additional context (Japan GX-ETS Phase 2, US DAC Hubs restoration, California SB 643 death, Denmark CCS Fund tender, Australia CFI amendment bill, Swiss KlV) is well-calibrated and dated. The major shortfall is the beyond-corpus probe: the answer openly states it could not execute the verification searches, so it contributes zero findings after 2026-08-14 and nothing on the named silent jurisdictions (Norway, Sweden, Switzerland post-February, Germany, Kenya, Article 6/PACM, ICVCM methodology decisions) or untracked ones (China, India, Brazil, Gulf). It correctly identifies these as gaps rather than absences, which is honest but leaves the question's global scope only partially answered. Citation quality is high — named instruments, interinstitutional file numbers, delegated regulation numbers, specific dates, and explicit sourcing distinctions (primary text vs. CDR.fyi commentary vs. trade press) — though some items rest on secondary reporting and the ISO 14060/SBTi material is flagged as uncorroborated. Coverage honesty is exemplary: freshness boundary stated up front, unverified claims labelled, a dedicated gaps section, and a warning not to read absence as inaction. Clarity and decision-usefulness are strong, with a synthesis ('compliance-market design advancing while cash disbursement lags') that an expert would find actionable, plus a useful correction on the California SB 643 timeline. Overall a rigorous, well-caveated answer that fully covers the corpus-derived gold but fails the extension requirement.
I'll start by checking what the corpus covers, then pull recent changes. Now let me pull the structured policy landscape and search the web for the silent jurisdictions and anything after the corpus freshness date. ## Important caveat before the substance Two limits shape this answer: 1. **Corpus freshness.** The Carbon Curve corpus holds policy events **only through 17 July 2026** (deals through 4 August 2026, last ingest 15 August 2026). Anything after mid-July 2026 is invisible to it. 2. **I could not run the verification searches.** The tool's own boundary field instructs me to web-search the silent jurisdictions (Norway, Sweden, Switzerland, Germany, Kenya, EU-level, international mechanisms) and the post-freshness window. My web search quota was exhausted after the first call returned no usable results, so **nothing below has been checked against sources newer than the corpus**. Treat this as a floor, not a census — particularly for the last several weeks and for jurisdictions outside the 14 tracked (no China, India, Brazil, Gulf states, or subnational programmes beyond California). --- ## The headline change: the EU moved removals into compliance-market design The single biggest development in the window is the **European Commission's EU ETS review proposal, published 17 July 2026** (interinstitutional files 2026/0210–0212 (COD)), which for the first time integrates permanent carbon removals into the ETS. The corpus records the mechanics as follows: - **Volume and vehicle:** a cumulative **250 Mt of CRCF-certified permanent removals** to be purchased over **2031–2040**, targeting an annual level of ~48 Mt by 2040. - **Architecture (proposed Article 9c):** the EU-wide allowance quantity increases by **250 million EUAs**; the Commission auctions them and uses the revenue to centrally purchase removal units through a new **Removals Authority**. **Direct offsetting by individual emitters was rejected** — installations cannot surrender removals against their own obligations. - **Eligibility:** initially **DACCS and BioCCS** with permanent geological storage under ETS MRV rules. **Biochar is excluded** on evidentiary grounds (limited long-term field studies, degradation-rate uncertainty), despite being recognised under the CRCF, with a pathway for later inclusion. - **Modelled mix:** the impact assessment is dominated by BioCCS (~175–330 Mt 2031–2040) versus roughly 20–35 Mt DACCS (~11%), though no technology split is prescribed. - **Timeline and fragility:** co-legislators are targeting political agreement around **Q1 2027**; a review is required by end-2034. Critically, the corpus flags (via CDR.fyi's analysis) that the legal text does **not** make allowance release conditional on removals being contracted or delivered, nor cancel unmatched allowances if procurement falls short — and that the funding implied supports an average removal price only ~104% of the average EUA auction price. If removals cost €100 more than an EUA, 250 Mt would need roughly **€25bn** of additional funding. Supporting EU moves in the same window: the **Innovation Fund awarded €2.7bn to 54 clean-industry projects on 24 March 2026**, including carbon-management projects. Just before the window, the Commission adopted **Delegated Regulation (EU) 2026/285** (February 2026) establishing CRCF certification methodologies for DACCS, BioCCS and biochar — the legal basis the ETS proposal now leans on. The **CRCF Buyers' Club** published its model in **May 2026**, with first permanent-removal purchases targeted by **December 2026**. --- ## United Kingdom: reaffirmation, not acceleration On **17 July 2026** — the same day as the EU proposal — the government published its **response to the Independent Review of Greenhouse Gas Removals** (Dr Alan Whitehead). It reaffirms the **GGR Business Model** as a top-up payment to an agreed price for credits sold, plus coverage of CCUS transport and storage network costs; commits to integrating GGRs into the **UK ETS "as soon as practicable"**; routes first geological-storage removals through CCUS cluster sequencing; reserves the right to use **Article 6** cooperation and to consider whether high-integrity **overseas DACCS** should complement domestic deployment — while planning to meet carbon budgets through 2037 with domestic action alone. Context the corpus adds, with an explicit confidence flag: **no signed GGR contract or contract award could be verified as of August 2026**, and the gov.uk business-model page showed no updates after 27 August 2025. The underlying UK ETS integration decision (July 2025 response) still targets legislating by **end-2028** and operation by **end-2029**, with a **200-year minimum storage** requirement and UK-only removals initially. So the July 2026 response is directionally positive but did not shorten the timeline or produce money on the table. --- ## Japan: the first mandatory compliance system of the window went live The **GX-ETS mandatory phase (Phase 2) commenced 1 April 2026** under the Amended GX Promotion Act, covering entities emitting ≥100,000 tCO₂/year — roughly 300–400 entities, about 60% of national emissions. J-Credits and JCM credits are usable for compliance up to a **10% cap**, with a price corridor of **JPY 4,300 ceiling / JPY 1,700 floor** (FY2027 values). The government-run allowance trading market under the GX Acceleration Agency is scheduled to open **autumn 2027**. Uncertainty flagged in the corpus: whether **international** voluntary-market removal credits (DACCS/BECCS/blue carbon beyond J-Credit and JCM) remain eligible in the mandatory phase could not be confirmed — approved for the voluntary phase in April 2024, but reported under reconsideration as of July 2025 with no final decision verified. --- ## North America: mixed signals **United States** - **DAC Hubs restored (April 2026).** After a federal court ruled the October 2025 mass grant terminations unconstitutional (suit led by the City of St. Paul), DOE's preserved-awards list sent to the House Appropriations Committee **retained funding for both flagship hubs** — Project Cypress (Louisiana) and the South Texas DAC Hub — roughly $1.2bn combined. But at least **10 of the original 21 awards** were terminated in October 2025 and are not restored: the programme is intact at the top and descoped underneath. The corpus notes no DOE primary press release confirming restoration was located; it rests on trade press reporting. - **California SB 643 died (2 March 2026).** The veto was **sustained** — the Legislature did not override. Note the sequence: Newsom vetoed on **13 October 2025** citing budget constraints (not opposition to CDR), pointing to SB 840's $85M climate innovation allocation; March 2026 is only the formal termination. A state purchase programme of ~$50M is therefore off the table for this session. - **45Q** is unchanged since the OBBBA amendment of 4 July 2025 ($180/t DAC storage or utilisation; $85/t point-source). **DOE's CDR Purchase Pilot Prize** survives with a $45M FY26 appropriation (up to 10 winners, up to $3M each). **Canada** - **28 April 2026 Spring Economic Update** proposes making **enhanced oil recovery an eligible storage use under the CCUS ITC**, at half the standard rates: 30% for DAC capture equipment, 25% other capture, 18.75% transport/storage/use (versus 60/50/37.5% for dedicated geological storage). EOR projects must permanently store ≥95% of CO₂ intended for EOR in a province with an adequate storage regime (Alberta, BC, Saskatchewan expected to qualify). Applies from 28 April 2026, subject to legislation; rates halve again 2036–2040. This is a widening of the tent that CDR advocates will read ambivalently. - Federal **CDR procurement** (≥CAD $10M by 2030, RFI issued Feb 2025) remained **pre-award** as of August 2026. --- ## Denmark, Australia, Switzerland - **Denmark, June 2026:** the **CCS Fund's first tender concluded**, with **Aalborg Portland (ACCSION)** signing for up to **DKK 16.5bn** to capture and store ~**1.25 Mt CO₂/yr from 2030**. This is the largest single subsidy award in the window, but it is **cement-sector fossil CCS, not CDR** — relevant to removals mainly through shared transport-and-storage infrastructure. Whether the remaining ~DKK 12bn goes to a second tender was unverified. - **Australia:** exposure draft of the **Carbon Credits and Other Legislation Amendment (Integrity and Transparency) Bill 2026** released **30 April 2026**, consultation closed **22 May 2026** — the first amendments to the CFI Act in over a decade (new Carbon Abatement Integrity Committee replacing ERAC, expanded native title consent, new ministerial integrity powers). **No evidence it had been introduced to Parliament as of August 2026.** Do not confuse it with the separate private member's bill debated 22 June 2026 on the NSW native forest management method. - **Switzerland:** the KlV amendment consultation (federal administration net-zero 2040, residual emissions balanced via negative-emission technologies) **closed 12 February 2026** — just before this window — and **no Federal Council adoption could be verified** as of 11 August 2026. First roadmaps are planned for the 2028–2032 legislature, to be adopted in 2027. --- ## Standards and registries (adjacent to policy, and moving faster) - **Puro Standard General Rules v4.4** approved **7 May 2026** (Puro was confirmed **ICVCM CCP-eligible at programme level on 11 December 2025**; individual methodologies not yet CCP-assessed). - **CORSIA**: eight programmes are approved for the 2024–2026 phase including **Isometric** (announced 26 November 2025) — the first durable-CDR-only registry eligible, opening a compliance channel. Actual availability of units with corresponding adjustments remains constrained. - Two corporate net-zero standards moved in the window, but the corpus holds these only as **uncorroborated signals** from CDR.fyi commentary (18 June 2026): the **draft ISO 14060** net-zero standard reportedly requires companies to begin buying CDR **within five years** of setting a target and scale to full counterbalancing (≥100-year durability), whereas **SBTi Corporate Net Zero Standard V2** imposes **no CDR purchase requirement until 2035**, then 1% scaling to 100% of residual emissions. If accurate, ISO would be the first major corporate standard treating CDR procurement as part of the transition plan from the outset. **This needs primary-source verification.** --- ## What this adds up to The window's pattern is **compliance-market design advancing while cash disbursement lags**. Japan's GX-ETS went mandatory; the EU proposed a 250 Mt central-purchase channel; the UK reaffirmed ETS integration — but the EU programme does not start until 2031, UK operation is targeted for end-2029 with no contract signed, Canada's procurement is still pre-award, and California's purchase programme is dead. CDR.fyi's June 2026 assessment (an uncorroborated signal in the corpus) frames this bluntly: policy is progressing but **not broad or fast enough to replace Microsoft-scale demand near-term**, with Microsoft accounting for **78.5% of all disclosed durable CDR tonnes purchased as of April 2026**. ## Known gaps you should not read as "nothing happened" - **No corpus records in this window for:** Norway, Sweden, Germany, Switzerland (post-February), Kenya, or international mechanisms (Article 6 / PACM / ICVCM methodology decisions). - **Untracked entirely:** China, India, Brazil, Gulf states, most subnational programmes, COP-cycle outcomes. - **Post-17 July 2026 policy events are absent**, including any Council/Parliament reaction to the EU ETS proposal, and I was unable to check them. If you want, I can re-run the targeted web verification on the highest-value unknowns — EU ETS proposal reaction since July, ISO 14060 status, UK GGR contract award, and the Article 6/PACM removals track — when search capacity is available.
Judge’s note — All three gold checklist items are surfaced with substantive instrument-level detail, not gestures: the Canadian Spring Economic Update EOR eligibility at half rates (30/25/18.75% vs 60/50/37.5%, ≥95% permanent storage, designated provinces, 28 April 2026 effect) with an explicit read on direction of travel; the UK 17 July 2026 response to the Whitehead review (top-up payment model, T&S cost coverage, UK ETS integration 'as soon as practicable', Article 6 / overseas DACCS optionality, plus the 200-year storage and legislate-by-2028 context and the fact no signed GGR contract was found); and the EU ETS review proposal in depth (CRCF-only DACCS/BioCCS, central-buyer Removals Authority, 250 Mt 2031–2040, ~48 Mt/yr by 2040, biochar exclusion rationale, Article 9c, the crucial point that allowance release is not conditional on delivery, ~€25bn implied cost, Q1 2027 political agreement target). Beyond-corpus additions are real but thin and almost all clustered on the same EU file: the international-credits purchasing facility with the January 2033 reporting obligation and LRF reversion to 2.7% from 2036 (ICAP), the ~€6bn benchmarks/free-allocation proposal (HSF Kramer), and the forward-looking Carbon Gap Policy Levers Playbook launch on 16 September — the only genuinely post-2026-08-14 item. It does not close any of the silent jurisdictions (Switzerland, Germany, Norway, Sweden, Kenya, international mechanisms/PACM) or non-tracked ones (China, India, Brazil), and says so plainly rather than bluffing. Citation quality is high: named publishers, regulation and file numbers ((EU) 2026/285, 2026/0210–0212 (COD)), dated events, and explicit medium-confidence flags with the reason (trade-press-only sourcing for DOE hub restoration, month-precision on Denmark). Honesty is exemplary — a dedicated gaps section separating structural corpus limits from unclosed search gaps, plus a firewall between verified records and CDR.fyi 'signals'. Structure is decision-useful (chronological table plus detail on the material items), marred only by a few lines of leftover process narration at the top. Deductions: limited genuinely new-jurisdiction or post-window reporting, and the SB 643 / Japan GX-ETS / Denmark items, while useful, are corpus recitation rather than new ground.
I'll start with the coverage baseline, then pull the policy deltas. The corpus gives me the tracked policy deltas. Now let me pull the instrument-level detail and run the suggested searches for jurisdictions that returned nothing. The corpus flags seven tracked jurisdictions with no events in this window, and policy events only recorded through 17 July 2026. Let me search for what it can't hold. ## What changed in CDR policy since 1 March 2026 **Method note first:** this draws on the Carbon Curve database (tracked policy events recorded through **17 July 2026**; 33 instruments across 14 jurisdictions), plus web searching to cover what the corpus structurally cannot hold. I was able to run only three web searches before hitting the tool's usage limit, so several gaps below remain genuinely unclosed — I flag them explicitly at the end rather than papering over them. --- ## 1. The biggest change: the EU proposed the first compliance-market channel for durable removals **Database record (approved, 2026-08-11 / 2026-08-15):** On **17 July 2026** the Commission published the EU ETS review proposal (interinstitutional files 2026/0210–0212 (COD)), which for the first time integrates permanent carbon removals into the ETS. The design is *indirect*: only CRCF-certified removals (Regulation (EU) 2024/3012), initially **DACCS and BioCCS**, purchased by the Commission acting as central buyer via a new **Removals Authority** — direct offsetting by emitters was rejected. **250 Mt cumulative over 2031–2040**, targeting ~48 Mt/yr by 2040, financed by auctioning additional EUAs, paid on delivery. Co-legislators target political agreement around **Q1 2027**. The corpus records a second, more granular instrument record for the same mechanism (**proposed Article 9c**), which adds: emitters cannot use removals for their own compliance; **biochar is excluded** on evidentiary grounds (limited long-term field studies, degradation-rate uncertainty) despite CRCF recognition, with a pathway for later inclusion; the Commission's impact assessment models a portfolio dominated by BioCCS (~175–330 Mt) versus ~20–35 Mt DACCS (~11%); review required by end-2034. Funding is 250 million additional EUAs plus up to 10 million from within the existing cap. Critically, that record notes the legal text **does not make allowance release conditional on removal contracting or delivery**, nor cancel unmatched allowances if procurement falls short — so 250 Mt is an intended ceiling, not a financed commitment. If purchased removals cost €100 more than an EUA, acquiring 250 Mt would need roughly **€25 billion** of additional funding. **What web search added.** Linklaters (sustainablefutures.linklaters.com, published ~3 weeks ago) calls this <cite index="0">"one of the most significant developments in carbon removal policy to date"</cite> and frames it as following intense Member State pressure for a review against high energy costs. ICAP confirms the cap increase mechanism and adds a detail the corpus does not hold: the proposal also establishes **a facility for purchasing international credits**, with further legislation expected later in 2026 and a Commission report to Parliament and Council due **by January 2033** — and if high-integrity international credits prove unavailable, the linear reduction factor reverts to 2.7% from 2036. Herbert Smith Freehills Kramer (published ~3 days before this response) confirms the 250 Mt figure and notes a separate benchmarks proposal worth roughly **€6 billion in additional free allocation 2026–2030**. Bellona Europa's 17 July response raises the supply-side concern directly — whether enough high-quality removals can be delivered in the timeframe. Carbon Gap's tracker characterises it as a public authority purchasing removals financed by allowance auctioning rather than direct compliance use. **Related, just before the window:** the corpus records Delegated Regulation **(EU) 2026/285**, adopted **February 2026**, establishing CRCF certification methodologies for DACCS, BioCCS and biochar — the legal foundation the ETS proposal depends on. **Also inside the window but recorded as an instrument attribute rather than a dated event:** the **EU CRCF Buyers' Club** published its model in **May 2026**, with first permanent-removal purchases targeted by **December 2026**. That is a near-term milestone worth watching and the corpus holds no event timeline for it. --- ## 2. Other database-recorded policy events, chronologically | Date | Jurisdiction | Event | |---|---|---| | 2026-03-02 | US (California) | **SB 643 veto sustained** — Legislature did not override, terminating the CDR Purchase Program bill | | 2026-03-24 | EU | **Innovation Fund: €2.7bn to 54 clean industry projects**, including carbon management | | 2026-04-01 | Japan | **GX-ETS mandatory phase (Phase 2) in force** | | 2026-04-01 | US | **DOE Regional DAC Hubs funding restored** | | 2026-04-28 | Canada | **CCUS ITC amendment proposed: EOR made eligible** | | 2026-04-30 | Australia | ACCU **Integrity and Transparency Bill exposure draft** released | | 2026-05-07 | Finland (registry) | **Puro Standard General Rules v4.4** approved | | 2026-05-22 | Australia | ACCU consultation **closed** | | 2026-06-01 | Denmark | **CCS Fund first tender concluded** | | 2026-07-17 | UK | **Government response to the GGR Independent Review** | **Detail on the material ones:** - **California SB 643** — worth stating precisely because the dates are easy to garble: Newsom vetoed the bill on **13 October 2025** citing budget constraints (pointing to SB 840's $85M climate innovation allocation as an alternative), and **2 March 2026** is the date the veto was formally sustained. The bill would have created a state CDR purchase program reported at up to ~$50M. It is now terminal. That removes what would have been the first US state CDR procurement programme. - **Japan GX-ETS** — mandatory phase began **1 April 2026** under the Amended GX Promotion Act, covering entities emitting ≥100,000 tCO₂/yr (roughly 300–400 entities, ~60% of national emissions). J-Credits and JCM credits usable up to a **10% cap**; price corridor ceiling JPY 4,300 / floor JPY 1,700 (FY2027 values). A government-run allowance market under the GX Acceleration Agency is scheduled for **autumn 2027**. Important caveat carried in the record: whether *international* voluntary removal credits (DACCS/BECCS/blue carbon beyond J-Credit and JCM) are eligible in the mandatory phase **could not be confirmed** — approved for the voluntary phase in April 2024 but reported as under reconsideration for possible revocation as of July 2025. - **US DOE DAC Hubs** — after a federal court ruled the October 2025 mass grant terminations unconstitutional (suit led by the City of St. Paul), DOE's **April 2026** preserved-awards list sent to House Appropriations retained both flagship hubs (Project Cypress, Louisiana; South Texas DAC Hub), roughly **$1.2bn combined**. At least 10 of the 21 original awards had been terminated in October 2025, so the programme is descoped but its two large hubs survive. **Confidence flag: medium** — no DOE OCED primary press release confirming restoration was located; this is sourced from trade press (Decarbonfuse, Carbon Herald, Latitude Media) reporting on the list sent to appropriators, and the date is month-precision only. - **Canada CCUS ITC** — the Spring Economic Update (**28 April 2026**) proposes making **enhanced oil recovery an eligible storage use** at half the standard rates: 30% for DAC capture equipment, 25% other capture, 18.75% transport/storage/use, versus 60%/50%/37.5% for dedicated geological storage. EOR projects must permanently store ≥95% of CO₂ intended for EOR, in a province designated as having an adequate storage regime (Alberta, BC, Saskatchewan expected to qualify). Applies from 28 April 2026, subject to legislation. This is the most consequential *direction-of-travel* change in the window: it widens a CDR-adjacent tax credit toward oil production. - **Denmark CCS Fund** — first tender concluded **June 2026**: Aalborg Portland (ACCSION) signed for up to **DKK 16.5bn** to capture and store ~1.25 Mt CO₂/yr from 2030, out of a DKK 28.7bn envelope. **Confidence: medium** — award day is month-precision, ens.dk had not updated at fetch time, and whether the remaining ~DKK 12bn goes to a second tender is unverified. Note this is a fossil-cement award, not CDR; the fund is open to both. - **UK GGR** — the **17 July 2026** government response to the Whitehead Independent Review reaffirms the GGR Business Model (top-up payment to an agreed price for credits sold, plus coverage of CCUS transport and storage network costs), aims to integrate GGRs into the UK ETS **"as soon as practicable"**, and routes first geological-storage removals through CCUS cluster sequencing. It reserves the right to use Article 6 cooperation and will consider whether high-integrity **overseas DACCS** should complement domestic deployment — while planning to meet carbon budgets through 2037 with domestic action alone. Context from the existing instrument records: the UK ETS decision (July 2025) sets a **200-year minimum storage** requirement, UK-based removals only initially, gross cap maintained, legislating by end-2028 and operational by end-2029. As of the corpus's August 2026 check, **no signed GGR contract had been found** and the gov.uk business model page showed no updates after 27 August 2025. - **Australia ACCU** — first amendments to the Carbon Farming Initiative Act in over a decade (new Carbon Abatement Integrity Committee replacing ERAC, expanded native title consent, new ministerial integrity powers). As of August 2026 there was **no evidence the bill had been introduced to Parliament**. Caution flagged in the record: a separately named private member's bill on the NSW Improved Native Forest Management method is a different bill and is often conflated with this one. --- ## 3. Where database and web sources diverge or need reconciling - On the EU ETS, they **agree on the substance** but the web sources carry material the corpus lacks: the international-credits purchasing facility, the January 2033 reporting obligation, the LRF reversion clause, and the €6bn free-allocation benchmarks proposal. Treat the corpus record as the structural skeleton and the law-firm/ICAP coverage as the mechanism detail. - The corpus itself flags a **conflict on the EU ETS publication URL**: date 2026-07-17 verified across Bellona, ICAP, Linklaters and AlliedOffsets, but the ec.europa.eu press-corner release could not be retrieved and the COM document number is unconfirmed. --- ## 4. Market-side signals touching policy — uncorroborated, not records The corpus holds these as **signals**, deliberately kept apart from verified records. Do not treat as fact: - A **CDR.fyi analysis (12 June 2026)** argues policy is progressing (CRCF methodologies adopted, EU ETS assessment, UK ETS integration and CfD-style models, German CDR funding, the Switzerland–Norway Article 6.2 pathway, Japan's mandatory GX-ETS, Canadian procurement, limited CORSIA eligibility, restored US DAC Hubs funding) but **is not broad or fast enough to replace Microsoft-scale demand near-term**, naming the July 2026 EU ETS review as the largest watchpoint. It claims Microsoft accounted for **78.5%** of all disclosed durable CDR tonnes as of April 2026. - A **CDR.fyi post (18 June 2026)** contrasts the ISO 14060 draft net-zero standard (CDR purchases required within five years of setting a target) with **SBTi Corporate Net Zero Standard V2** (no CDR purchase requirement until 2035, then 1% scaling to 100% of residual emissions). Both are private standards, not policy, but they shape compliance-adjacent demand. --- ## 5. What neither source covers — the honest gaps **Structural corpus limits:** - Policy events recorded only **through 17 July 2026**. Anything after that date is absent by construction. - **14 jurisdictions tracked.** Seven of them — Switzerland, Germany, the EU-as-filed, international mechanisms, Kenya, Norway, Sweden — returned **zero events** for this window. That is no record, not no activity. Switzerland's KlV amendment consultation closed 12 February 2026 with **no adoption announcement found as of 11 August 2026**, and a separate post-2030 climate policy consultation including a CCS/NET framework act was expected by Q2 2026 — whether it opened is unverified. - **Subnational and non-OECD coverage is thin.** China, India, Brazil, Indonesia, the Gulf states and the African Union have no policy instruments in the corpus at all, despite India and Brazil appearing repeatedly on the *deal* side in this same window. **Gaps I could not close because the web search budget ran out:** - **Article 6.4 / PACM** removals standards and any 2026 Supervisory Body decisions — potentially the single most globally consequential CDR governance track, and entirely absent here. - **Germany**: whether the €111M+ 2026 negative-emissions budget line (including €11.5M for direct government purchase of removal certificates) has produced a procurement mechanism. The corpus's own figures were verified only against the July 2025 *draft* budget, not the enacted one. - **US federal** developments after April 2026 — 45Q rulemaking, DOE CDR Purchase Pilot Prize awards (24 Phase 2 semifinalists, up to 10 winners at up to $3M each, $45M FY26 appropriation, no award event recorded), Class VI primacy applications. - **Canada's federal CDR procurement** (≥CAD $10M by 2030, RFI February 2025) remains **pre-award** as of August 2026 with no recorded event — worth checking directly. - **China, India, Brazil, COP31 preparation** — no coverage from either source. One forward-looking item surfaced incidentally: Carbon Gap is launching a **Carbon Removal Policy Levers Playbook in Brussels on 16 September**, convening international, EU, national and local governments — a useful marker for where European policy attention is heading next.
Judge’s note — All three gold checklist items are covered substantively and with unusual specificity: the Canadian Spring Economic Update EOR extension to the CCUS ITC (28 Apr 2026, half rates 30/25/18.75%, \u226595% permanent storage condition), the UK DESNZ response to the Whitehead GGR review (17 July 2026, top-up payment model, UK ETS integration, 200-year durability, Article 6 optionality), and the EU ETS review proposal (17 July 2026, new Art. 9c, 250 Mt central procurement 2031\u20132040, CRCF-certified DACCS/BioCCS only, biochar excluded), including a sharp critique that the 250 Mt is a ceiling rather than a financed commitment. It adds well-chosen secondary material (Innovation Fund March 2026, CRCF Buyers' Club, DAC Hubs restoration via litigation, California SB 643 death, Denmark CCS Fund tender with the correct caveat that it is point-source not CDR, Japan GX-ETS Phase 2, Australian ACCU integrity bill, Puro v4.4). Where it falls short is exactly on the beyond-corpus probe. The answer explicitly bounds itself at the corpus cut-off (events through 17 July 2026) and surfaces no verified policy development after 2026-08-14; the thin jurisdictions are correctly *named* (Switzerland, Germany, Norway, Sweden, Kenya, plus China/India/Brazil/Indonesia/Gulf) and the Switzerland KlV / Article 6-PACM / ISO 14060 / Switzerland\u2013Norway 6.2 threads are flagged as unverified or record-less \u2014 but none are resolved with independent evidence. So the global picture stops short of the frontier the question asks about, and several potentially checkable leads are left dangling. Citation quality is good but mixed: named institutional sources (ICAP, CINEA, DESNZ, Council doc WK-10833-2026-INIT, CDR.fyi) and explicit confidence downgrades where only trade press supports a claim (DAC Hubs restoration) or where a primary URL/COM number could not be verified; no fabricated precision detected. Coverage honesty is exemplary \u2014 a dedicated section separating gaps from inaction, freshness limits stated up front, uncorroborated signals labelled as such. Clarity and decision-usefulness are high, with a synthesis that correctly isolates the design convergence on compliance-market integration and the back-loading of EU money.
I'll start by checking what the corpus actually covers, then pull recent changes. Let me drill into the major instruments and check the jurisdictions that returned no events. ## Bottom line Between 1 March 2026 and the corpus's policy cut-off (**events recorded through 17 July 2026**; last ingest 15 Aug 2026), the single most consequential development is the **European Commission's 17 July 2026 EU ETS review proposal**, which for the first time puts permanent CDR inside the EU's compliance carbon market — via central government purchasing rather than direct offsetting. Around it, the picture is one of **demand-side instruments consolidating in Europe, US federal support partially rescued by litigation rather than legislation, and sub-federal US procurement dying**. --- ## 1. European Union — the big one **EU ETS Review proposal, 17 July 2026** (interinstitutional files 2026/0210-0212 (COD), amending Directive 2003/87/EC) — status: *proposed*, now in the ordinary legislative procedure. - Creates a **new Article 9c**: the EU-wide allowance quantity rises by **250 million EUAs**, which the Commission auctions and uses the revenue to **centrally purchase an equivalent 250 Mt of permanent removals over 2031–2040**, targeting a ~48 Mt/yr level by 2040. A new **Removals Authority** would act as central buyer; payment on delivery. - **Direct offsetting by individual emitters was rejected.** Eligibility is restricted to **CRCF-certified (Regulation (EU) 2024/3012) DACCS and BioCCS** with permanent storage under ETS MRV rules. **Biochar is excluded** on evidentiary grounds (limited long-term field data, degradation-rate uncertainty), despite CRCF recognition — with a pathway for later inclusion. - The Commission's impact assessment models portfolios **dominated by BioCCS (~175–330 Mt, 2031–2040) versus ~20–35 Mt DACCS (~11%)**, though no technology split is prescribed. Review required by end-2034. - **Important caveat carried in the corpus:** 250 Mt is an *intended scale/ceiling, not a financed purchase commitment*. The legal text does **not** make allowance release conditional on removals being contracted or delivered, nor cancel unmatched allowances if procurement falls short. Auctioning 250 Mt of EUAs only funds removals at ~104% of the average EUA auction price; if removals cost €100/t more than an EUA, the gap is roughly **€25 billion** (analysis: CDR.fyi, "The EU ETS proposal explained"). Co-legislators target political agreement around **Q1 2027**. - Sources: ICAP (icapcarbonaction.com), Council doc WK-10833-2026-INIT, CDR.fyi. *Corpus flags "medium" confidence: the exact ec.europa.eu press release URL and COM document number could not be verified.* **Other EU changes in the window:** - **Innovation Fund, 24 March 2026** — €2.7 billion awarded to 54 clean-industry projects, including carbon-management (CCS/CCU/DAC-relevant) projects (CINEA). The corpus flags that specific DAC awardees within the ~10 carbon-management projects are **unconfirmed**. - **EU CRCF Buyers' Club** — Commission-convened buyer coalition; **purchasing model published May 2026**, two tracks (permanent CDR / carbon farming), with **first permanent-removal purchases targeted by December 2026**. This is a real change inside the window but is recorded as an instrument attribute, not a dated event. - Context immediately preceding the window: **Delegated Regulation (EU) 2026/285** (adopted Feb 2026) established CRCF certification methodologies for **DACCS, BioCCS and biochar** — the certification layer the ETS proposal now leans on. --- ## 2. United Kingdom **17 July 2026 — Government response to the Independent Review of Greenhouse Gas Removals** (Dr Alan Whitehead), published by DESNZ. It: - Reaffirms the **GGR Business Model**: a **top-up payment to an agreed price** for credits sold, plus coverage of CCUS transport-and-storage network costs. - Commits to integrate GGRs into the **UK ETS "as soon as practicable"** as the long-term demand signal (the July 2025 decision set: ex-post allowance award, **minimum 200-year storage**, UK-only removals initially, gross cap maintained, legislate by end-2028, operational by end-2029). - First geological-storage removals to come via **CCUS cluster sequencing**. - Notably, **reserves the right to use Article 6 cooperation** and will consider whether high-integrity **overseas DACCS** should complement domestic deployment — while planning to meet carbon budgets **through 2037 with domestic action alone**. **What has *not* happened:** as of the August 2026 corpus check, **no signed GGR contract or award** could be found (HyNet Track-1 expansion: Ince BECCS/Evero as Priority, Climeworks' Silver Birch DACCS as Standby, both listed Aug 2025), and no government response to the April 2025 GGR/Power BECCS revenue-support regulations consultation. Both instruments remain at "proposed". --- ## 3. United States — restoration by court order, and a state programme dies - **DOE Regional DAC Hubs restored, ~April 2026.** After a federal court ruled the October 2025 mass termination of 200+ clean-energy grants unconstitutional (suit led by the City of St. Paul), DOE's preserved-awards list sent to the House Appropriations Committee **retained funding for both flagship hubs** — Project Cypress (LA) and South Texas DAC Hub — roughly **$1.2B combined** of the $3.5B BIL appropriation. At least 10 of the original 21 awards were terminated in Oct 2025 and remain gone; the programme is **descoped but with its two anchors intact**. *Confidence: medium — no DOE/OCED primary press release confirming restoration was located; sourced from trade press (Decarbonfuse, Carbon Herald, Latitude Media), month precision only.* - **California SB 643 terminated, 2 March 2026.** The legislature declined to override Newsom's 13 October 2025 veto, so the veto was formally sustained and the bill died. It would have created a state CDR purchase programme (~$50M reported); the veto message cited budget constraints and pointed to SB 840's $85M climate innovation allocation instead. Note this bill had passed 37–0 in the Senate and 78–1 in the Assembly — the failure was fiscal, not ideological. - **Unchanged in the window:** 45Q ($180/t DAC geological storage or utilization; $85/t point-source; begin-construction deadline 1 Jan 2033) — last amended by OBBBA on 4 July 2025. Class VI primacy stands at six states after Texas (effective 15 Dec 2025), with **Louisiana's Class VI moratorium (EO JML 25-119, Oct 2025) still recorded as in place**. DOE's CDR Purchase Pilot Prize is recorded as preserved with a **$45M FY26 appropriation** (24 Phase 2 semifinalists, up to 10 winners at up to $3M each) — but no dated event since March 2026, so its award status is a gap. --- ## 4. Canada **28 April 2026 — Spring Economic Update proposes extending the CCUS ITC to enhanced oil recovery**, at half the standard rates: **30% for DAC capture equipment, 25% for other capture, 18.75% for transport/storage/use**, versus 60%/50%/37.5% for dedicated geological storage. EOR projects must permanently store **≥95%** of the CO₂ intended for EOR, in a province with a designated adequate storage regime (Alberta, BC, Saskatchewan expected to qualify). Applies from 28 April 2026, subject to legislation; rates halve again for 2036–2040. This is a genuinely two-sided change: it broadens the subsidy base but deliberately prices EOR storage below dedicated storage — and it is the first material dilution of the "dedicated geological storage only" principle in the ITC. **Canada federal CDR procurement** (≥CAD $10M by 2030, announced Oct 2024, RFI Feb 2025) is recorded as **still pre-award as of August 2026** — no change in the window. --- ## 5. Nordics - **Denmark, June 2026:** the **DKK 28.7bn CCS Fund** concluded its first tender — **Aalborg Portland (ACCSION) contracted for up to DKK 16.5bn** to capture and store ~1.25 Mt CO₂/yr from 2030. Read carefully: this is **cement point-source CCS, not CDR**; the fund is pathway-agnostic. The corpus cannot verify whether the remaining ~DKK 12bn goes to a second tender, or that tender's status. Source: Carbon Herald; ens.dk overview page not yet updated at fetch time. - **Sweden:** no policy event recorded in the window. The SEK 36bn BECCS reverse auction (Stockholm Exergi, ~SEK 20bn / ~800 kt/yr) is unchanged. The related commercial signal — the **City of Stockholm's 26 May 2026 purchase of 50,000 t/yr for 15 years** from Stockholm Exergi, making it "the world's fifth-largest buyer of carbon removals" — is a municipal procurement, not a tracked instrument. --- ## 6. Asia-Pacific - **Japan, 1 April 2026:** the **GX-ETS mandatory phase (Phase 2) entered into force**, covering entities emitting **≥100,000 tCO₂/yr** (based on FY2023–2025 average emissions). The government-run allowance trading market, operated by the GX Acceleration Agency, is scheduled to **open in autumn 2027** (recorded deadline: 1 Oct 2027, month precision). Source: ICAP. *The corpus does not record whether or how removals are eligible under GX-ETS — a material gap for anyone assessing this as a CDR demand signal.* - **Australia:** DCCEEW released the **exposure draft of the Carbon Credits and Other Legislation Amendment (Integrity and Transparency) Bill 2026** on **30 April 2026**; consultation **closed 22 May 2026**; government is considering feedback before parliamentary introduction. This is ACCU-scheme integrity reform, mostly relevant to land-sector/conventional-durability credits. --- ## 7. Standards, registries and corporate accounting - **Puro.earth General Rules v4.4 approved, 7 May 2026** — the latest version identified as of Aug 2026. - Two **uncorroborated but important signals** (source claims, not verified records): - **ISO 14060 draft net-zero standard** vs **SBTi Corporate Net Zero Standard V2**: per CDR.fyi (18 June 2026), ISO would require companies to **begin buying CDR within five years of setting a target** and scale to full counterbalancing, with **≥100-year durability** for residual emissions; SBTi V2 imposes **no CDR purchase requirement until 2035**, then 1% scaling to 100% of residuals by the net-zero date. If accurate, ISO would be the first major corporate standard treating CDR procurement as part of the transition plan from the outset. **Treat as unverified.** - CDR.fyi's June 2026 market update names the **EU ETS review as the largest near-term policy watchpoint**, and lists a **Switzerland–Norway Article 6.2 durable CDR transfer pathway** — which the corpus holds **no policy record for**. --- ## What I could *not* establish — be careful here These are real gaps, not evidence of inaction: 1. **Article 6 / PACM / CORSIA / ICVCM.** Searches for international mechanisms and CORSIA/ICVCM removal eligibility returned **zero records**. Given COP-cycle activity and the UK's explicit Article 6 language in its July 2026 response, this is a coverage hole, not a quiet period. 2. **Switzerland.** The KlV amendment consultation (federal administration net-zero 2040, residual emissions balanced via negative-emissions technologies, "Modell ecCo") **closed 12 Feb 2026** — just before your window — and **no Federal Council adoption could be verified as of 11 Aug 2026**. A separate consultation on post-2030 climate policy including a CCS/NET framework act was expected by Q2 2026 per BAFU documents; whether it opened is **unverified**. 3. **Germany, Norway, Sweden, Kenya** returned **no events** in this window despite being tracked. Germany's two instruments (KSpTG in force 28 Nov 2025; the **first-ever dedicated federal negative-emissions budget line — >€111M for 2026, including €11.5M for direct government purchase of CO₂-removal certificates, plus €320M in commitment authorisations through 2033**) both pre-date March 2026, and the corpus notes those figures were verified only against the **July 2025 draft**, not the enacted budget. 4. **Coverage limits:** 33 policy instruments across 14 jurisdictions. **Subnational and non-OECD coverage is explicitly thin** — nothing here on China, India, Brazil, Indonesia or the Gulf, despite deal activity in India, Brazil, Indonesia and Uganda in the same period. 5. **Freshness:** policy events stop at **17 July 2026**. Anything from late July or August 2026 is outside this dataset entirely. --- ## Reading the trend Three things are worth separating. First, **Europe is building the demand plumbing** (CRCF methodologies → CRCF Buyers' Club → ETS Article 9c) but the money is back-loaded to the 2030s and, on the corpus's own reading of the legal text, **not actually guaranteed** — the allowances get released whether or not removals are bought. Second, **the US is in a holding pattern where the courts, not Congress or the states, are doing the preserving** — DAC Hubs survived by litigation while California's purchase programme died by fiscal veto. Third, **the compliance-market question has now converged on one design choice**: EU (central buyer, no direct offsetting), UK (ex-post allowances inside a maintained gross cap, 200-year floor), Japan (mandatory ETS live, removal eligibility unclear). Whoever gets this right first sets the template — and on current schedules, that's the UK in 2029 and the EU in 2031.
Asked, verbatim: “How strong is the field evidence for ocean alkalinity enhancement? How much of the literature is actual in-water measurement vs modeling?”
Judge’s note — The answer is a well-organized, expert-sounding synthesis with a strong field-trial inventory (Planetary/PML Cornwall 2022 and Halifax, Vesta Duck NC olivine berm, an electrochemical outfall trial with a negative detectability result, CSIRO Tasmania NaOH release with ~370 µatm pCO2 drop, LOC-NESS Gulf of Maine EPA-permitted trial with Feb 2026 preliminary ecological results, Mendes intertidal, Paul mesocosm) — roughly eight distinct named, dated, sourced findings that go well beyond a generic answer. Citations are mostly named authors/journals/years and thus checkable, and the answer is admirably honest about tiers of evidence (press briefing vs peer review) and about the central verification gap (no directly measured air–sea flux; all credited tonnage model-derived). However, it fails on the quantitative core of the question. It never surfaces the corpus scale (~308 in-scope OAE papers), never gives the field-trial count of 16 or the ~6% share (it hedges with 'on the order of a dozen' and 'low tens'), and never provides the lab (~89) / modeling (~80) / review (~52) breakdown, substituting an explicitly acknowledged guess of '10:1 or more.' The honesty about not having a bibliometric count is creditable but does not substitute for the classification data the question asks for. Direction is right; the numbers — the actual answer to 'how much' — are absent.
I'll research this systematically — looking at the published field trial record, the modeling-vs-measurement balance in the literature, and the known verification gaps. ## Bottom line The field evidence for ocean alkalinity enhancement (OAE) is **real but thin, young, and structurally incomplete**. As of mid-2026 there are on the order of a dozen published or presented open-water field experiments worldwide, essentially all of which demonstrate the *first* link in the causal chain (you can add alkalinity, raise pH, depress seawater pCO₂, and track the plume). **No field trial has yet directly measured the atmospheric CO₂ uptake it claims** — that step is, without exception, supplied by models. The tonnage sold and credited by OAE companies is therefore model-derived, not measured. --- ## The actual field record | Trial | Date / place | Material | What was shown | |---|---|---|---| | Planetary Technologies / Plymouth Marine Laboratory | Sept 2022, St Ives Bay, Cornwall UK | Mg(OH)₂ via wastewater outfall | First peer-reviewed OAE field result — Kitidis et al., *Communications Earth & Environment* (2024), on magnesium hydroxide reducing aqueous CO₂ in ocean-discharged wastewater. PML provided independent monitoring after lab tests and modelling. | | Planetary / Dalhousie | Sept 2023–ongoing, Halifax Harbour | Mg(OH)₂ | Longest-running operational site. *Knowable Magazine* (Mar 2026) reports over 3,600 tonnes CO₂ removed since Sept 2023 — a **credited/accounted** figure, not a directly observed flux. | | Vesta | Spring 2024, Duck, North Carolina | ~8,200 t olivine sand (D50 = 0.17 mm), berm 640 m offshore at 7.6 m depth | First coastal enhanced weathering field trial; 4 months of pre-deployment baseline plus monthly monitoring (AGU 2024, preliminary geochemistry). | | Electrochemical OAE trial | Nov 2024 & Feb 2025 | Electrochemically derived aqueous alkalinity | *Frontiers in Environmental Engineering* (Jul 2025). Notably a **negative detectability result**: even after a >6-fold increase in discharge volume, elevated pH was detectable only just outside the outfall pipe; biological impact data could not be collected, and near-field modelling (PLUMES2.0) was needed to fill the gap. | | CSIRO / Australian trial | 2025, coastal Tasmania | Continuous aqueous NaOH | Wynn-Edwards et al., *Scientific Reports* (Dec 2025) — first Australian OAE field trial; plume tracked, and pCO₂ fell by up to 370 µatm at the release point. | | LOC-NESS (WHOI) | Aug 2025, Wilkinson Basin, Gulf of Maine | ~16,000+ gallons purified NaOH over a 6-hour dispersal, plus Rhodamine WT dye | First **EPA-permitted** OAE trial in US federal waters. EPA approved in April 2025 after a year-long review, two comment periods totalling 75 days and 250+ comments; NOAA Fisheries and USFWS consultations found minimal potential environmental impact. The patch was tracked for four days using shipboard sampling, gliders, a long-range AUV, drifters and remote sensing. | | Intertidal alkalinity enhancement | 2025 | — | Mendes et al., *JGR Biogeosciences* (Feb 2025). | Plus mesocosm work, which sits between lab and field — e.g. a one-month open-ocean mesocosm study in a low-nutrient ecosystem (Paul et al., *Biogeosciences* 2024) found biogeochemical functioning largely stable but flagged that long-term carbon-storage capability "may be limited at high alkalinity concentration." **LOC-NESS preliminary findings** (announced at a press briefing at the Ocean Sciences Meeting in Glasgow, Feb 2026): seawater chemistry returned to baseline within expected timeframes, and no measurable impact was detected on bacteria, phytoplankton, zooplankton, fish larvae or lobster larvae. Treat these as **preliminary and not yet peer-reviewed** — a conference press release is the weakest tier of evidence in this table. --- ## What the field evidence does and does not establish Think of OAE as a three-link chain. The field record covers them very unevenly: 1. **Alkalinity delivery and carbonate-system response — well supported.** Multiple independent trials, different materials (NaOH, Mg(OH)₂, olivine, electrochemical), different geographies, consistently show the expected pH rise and pCO₂ drop. This link is no longer seriously in doubt. 2. **Air–sea CO₂ equilibration — not directly measured anywhere.** Equilibration takes months to years while a dispersed patch becomes undetectable within days. A 2026 *Frontiers in Climate* paper notes that field trials confirm pCO₂ reductions and their propagation, but that mixing, dilution and advection make the signal hard to detect reliably beyond a limited spatial extent. The 2023 OAE Guide field-experiments chapter (Cyronak et al.) makes the same point bluntly — dilution of the alkalinity and ultimately the CO₂ signal makes MRV harder. 3. **Durability and ecological safety at scale — largely untested.** Null ecological results so far come from small, short, dilute perturbations. That is weak evidence about repeated or industrial-scale addition, and near-zero evidence about chronic exposure or benthic effects. The MRV chapter of the OAE Guide (Ho et al., 2023) is explicit that observations should quantify alkalinity delivery and monitor for secondary precipitation and biotic calcification, that subduction is a real risk for nearshore trials, that tracers such as SF₆ are crucial, and that **coupling explicit OAE modelling to field trials is a necessary milestone for robust MRV**. That is an admission, from the field's own reference text, that measurement alone cannot close the accounting. --- ## The modeling-vs-measurement split **Direct answer, with an explicit caveat: I could not locate a peer-reviewed bibliometric study that quantifies this split, so treat any ratio as an estimate rather than a citable number.** What is citable: - The 2021 US National Academies report on ocean CDR states that the vast majority of results on OAE chemical and biological impacts come from **modeling studies**, and calls for urgent empirical data from lab and mesocosm work and small field trials. That is the field's own baseline assessment. - A 2024 *Earth's Future* global Earth-system-model study (Palmiéri & Yool) notes that modeling studies have typically investigated OAE at basin or global scale — and, tellingly, that in their simulation the added alkalinity is indistinguishable from background alkalinity. That is precisely the detection problem that dooms direct field verification. My own reading of the corpus, offered as a judgment rather than a count: the published literature is **overwhelmingly model- and lab-dominated — plausibly on the order of 10:1 or more in favour of modeling, lab and desk studies over open-water field measurement**, with open-water field experiments numbering in the low tens versus hundreds of modeling, laboratory, techno-economic and review papers. The direction of the imbalance is not in dispute; the exact ratio is. Note also that the balance has shifted meaningfully since the NASEM assessment — 2024–2026 produced the first genuine cluster of peer-reviewed field results, so a 2021-vintage characterisation now understates the empirical base. Two further distortions worth naming: - **Publication asymmetry.** Modeling papers are cheap and fast; field trials cost millions, require years of permitting, and produce one paper. The literature count systematically understates the effort going into fieldwork. - **Grey literature.** Much of the most operationally relevant data sits in company reports, registry documentation and conference abstracts (e.g. the Duck, NC geochemistry is an AGU 2024 abstract), not in indexed journals. --- ## The specific gaps I would flag to anyone relying on this evidence 1. **No measured air–sea flux attribution.** Every tonne credited to date rests on a model of equilibration efficiency. This is the single largest verification gap in the sector. 2. **Secondary precipitation / "runaway" carbonate loss.** If alkalinity is added too fast or too concentrated, CaCO₃ precipitates and consumes the alkalinity you paid for. Mesocosm work suggests storage capability may degrade at high alkalinity concentrations. Field data on the practical dosing ceiling remain sparse. 3. **Olivine dissolution kinetics are poorly constrained in situ.** Laboratory work published in *Geochimica et Cosmochimica Acta* (2023) found continuously tumbled olivine dissolved 8–19× faster than under stagnant conditions — an order-of-magnitude parameter uncertainty that propagates straight into cost and tonnage estimates. Duck, NC is the first real-world test. 4. **Detectability limits are worse than expected.** The electrochemical trial's finding that a 6× volume increase still produced only near-outfall detectability is an important, under-discussed result. 5. **Ecological null results are low-power.** "No measurable impact" over four days on a diluted patch is not evidence of safety at deployment scale. Trace-metal release (nickel, chromium from olivine) over multi-year timescales is essentially unmeasured in the field. 6. **Independence.** Several key datasets are company-generated or company-funded, even where monitoring was contracted out to independent labs like PML. The LOC-NESS trial is unusual in being academically led and regulator-permitted. --- ## What would actually move the needle Peer-reviewed publication of the full LOC-NESS dataset (the current results are a press briefing); a multi-year mass balance on the Duck, NC olivine berm showing how much of the 8,200 tonnes actually dissolved and where the alkalinity went; and — the hard one — a sustained-addition experiment with a deliberate tracer (SF₆ or similar) large and long enough to observe air–sea equilibration rather than infer it. Until that third thing exists, OAE carbon accounting remains a modelled quantity with field-validated inputs, not a measured one.
Judge’s note — All three gold checklist items are surfaced precisely: 308 classified OAE papers described as the complete OpenAlex result set as of 2026-08-14; 16 field-trial papers given as 5.2% of 308 / 6.4% of classified; and the lab (89), modeling (80), review (52) distribution laid out in a table. The answer goes well beyond the checklist with a critical reclassification of the 16 'field trials' into mesocosms, true open-water alkalinity releases (~3–4 papers, ~1% of the literature), and non-alkalinity/observational studies — exactly the kind of analysis an expert practitioner needs. Specific in-water results are reported with DOIs (Tasmania pCO2 drop of 22–77 µatm downstream; Sequim Bay pH signal confined to ~2.5 m; Martha's Vineyard tracer study achieving 10% of potential CDR; ~1,050-day 95% equilibration estimate from Bergen), plus lab efficiency risks (Hartmann, Moras) and the He & Tyka equilibration modeling. Beyond-corpus: the LOC-NESS Gulf of Maine August 2025 EPA-permitted trial is surfaced with roughly three distinct substantive findings (trial existence/permit status and Feb 2026 Ocean Sciences presentation; 65,000 L NaOH over 1 km² with rhodamine and multi-platform monitoring; preliminary 2–10 t CO2 removed over four days, ~50 t estimated, no significant adverse effects), correctly flagged as preliminary and not peer-review-confirmed. Citation quality is very high (DOIs, journal names, dates), with secondary web sources named rather than laundered. Coverage honesty is exemplary: it flags 19% unclassified papers, that only 10 of 16 field-trial records were retrieved (so the open-water count is a floor), LLM-assigned study types from abstracts, mesocosm/open-water conflation as its own reclassification, exclusion of preprints/non-English, and the exhausted search budget preventing verification of delivered tonnages and LOC-NESS publication status. The offtake/policy section usefully frames the evidence-to-commitment gap and correctly notes the Ebb deal is mostly optional volume and CREW is wastewater, not open ocean. Minor deductions only: one thinly-sourced secondary citation for the LOC-NESS tonnage, and a small internal arithmetic slip (lab+modeling+review = 221 of 250 is 88%, but the '68%' and 'two-thirds' framings are stated loosely against different denominators).
I'll start by checking what the corpus covers, then pull the full evidence profile for ocean alkalinity enhancement.
Now let me pull the field-trial list in full and check MRV, then search the web for developments since the corpus freshness date.
## Bottom line
OAE has crossed from "theory and tanks" into "actual water" only in the last ~2 years, and the crossing is thin. In the Carbon Curve corpus — which holds a *complete* OpenAlex result set for the OAE pathway query as of **2026-08-14** — there are **308 classified OAE papers**, of which **16 (5.2%) are tagged field-trial** and **80 (26%) are modeling**. That's a **5:1 modeling-to-field ratio**. And when you read the 16 field-trial papers, most are not open-ocean alkalinity releases at all.
---
## 1. The raw split
| Study type | Papers | % of 308 | % of the 250 classified |
|---|---|---|---|
| Lab (bench + tank) | 89 | 28.9% | 35.6% |
| **Modeling** | **80** | **26.0%** | **32.0%** |
| Review | 52 | 16.9% | 20.8% |
| **Field-trial** | **16** | **5.2%** | **6.4%** |
| Survey (social science) | 8 | 2.6% | 3.2% |
| LCA | 5 | 1.6% | 2.0% |
| Unclassified (no rights-cleared abstract) | 58 | 18.8% | — |
Source: `get_evidence_profile('oae')`, Carbon Curve corpus. Coverage caveat that matters: study types were assigned by an LLM from title + abstract only (v1, 2026-08-12), 19% couldn't be classified, and preprints and non-English work are excluded. Treat these as ±a few points, not exact.
Two structural reads:
- **Desk-and-bench work is two-thirds of the classified literature** (lab + modeling + review = 221 of 250, 88%; lab + modeling alone = 68%).
- **Empirical work is heavily biased toward the wrong scale.** Lab studies outnumber field studies 5.6:1, and the lab studies are mostly beaker-scale carbonate chemistry and single-species/community response experiments.
---
## 2. "Field trial" is doing a lot of work in that 5.2%
Of the 16, the corpus surfaced 10. Sorting them by what was actually done:
**Enclosed mesocosms (in the sea, but not *of* the sea):**
- Helgoland, North Sea — six alkalinity levels, ΔTA up to 1,250 µmol/kg, 39 days (Biogeosciences, 2026-07-08, doi:10.5194/bg-23-4691-2026)
- Bergen fjord, Norway — five levels, ΔTA 0–600 µmol/kg, 47 days (Biogeosciences, 2026-01-08, doi:10.5194/bg-23-137-2026)
- Norwegian fjord, 53-day *Oikopleura dioica* study (PLoS ONE, 2026-03-30, doi:10.1371/journal.pone.0344503)
**Deliberate open-water alkalinity release with in-water chemistry — the real thing:**
- **Tasmania, Australia** — first Australian OAE field trial, continuous aqueous NaOH addition at a coastal site (Scientific Reports, 2025-12-12, doi:10.1038/s41598-025-31606-w)
- **Sequim Bay, WA** — Ebb Carbon electrochemical alkalinity discharged via a wastewater outfall (Frontiers in Environmental Engineering, 2025-09-18, doi:10.3389/fenve.2025.1641277)
- **Autonomous lab-on-a-chip TA analyzer deployed during an Mg(OH)₂-slurry trial in 2023** (Communications Engineering, 2026-04-22, doi:10.1038/s44172-026-00665-w)
**Not alkalinity at all:**
- Martha's Vineyard **rhodamine dye tracer** release, Sept 2023 — dye used as a *proxy* for alkalinity; no TA was added (Biogeosciences, 2025-10-10, doi:10.5194/bg-22-5511-2025)
- South China Sea 26-year time series of **natural** TA increase (+0.56 µmol/kg/yr) — observational, not an intervention (GRL, 2025-09-02, doi:10.1029/2025gl117872)
- Oyster-farming field mesocosms (PNAS, doi:10.1073/pnas.2504004122) and Narragansett Bay bivalve-shell dissolution (L&O Letters, doi:10.1002/lol2.70033) — adjacent, coastal-alkalinity-cycle work
So the count of **peer-reviewed papers reporting a deliberate alkalinity addition to open coastal/ocean water with in-water measurement is roughly 3–4 — about 1% of the OAE literature.** No commercial deployment has published a closed, peer-reviewed, measurement-based CDR quantification.
---
## 3. What the in-water work actually found — and why it's sobering
The field results are largely *methodological* successes and *quantification* warnings:
- **Tasmania:** at the point of release, pCO₂ fell by up to 370 µatm with TA up ~545 µmol/kg — but downstream maximum pCO₂ decreases were only 22–77 µatm, i.e. signal strengths of **<1–5%**, with the plume dispersing "within meters" of the addition point (doi:10.1038/s41598-025-31606-w).
- **Sequim Bay:** the elevated-pH region was confined to within **~2.5 m of the discharge pipe**; surrounding temperature, salinity, turbidity and oxygen showed no measurable change. Explicitly framed as proof-of-concept for a conservative release (doi:10.3389/fenve.2025.1641277).
- **Martha's Vineyard tracer study:** over 36 hours, the calculated CDR signal amounted to just **10% of the total potential CDR** — a 4 µatm fCO₂ rise, 0.004 pH units, 1.8 µmol/kg DIC. Detectability thresholds were TA >10 µmol/kg, pH >0.01, fCO₂ >10 µatm (doi:10.5194/bg-22-5511-2025).
- **In-situ TA analyzer:** first autonomous lab-on-chip TA deployment during a slurry trial recorded a ~40 µmol/kg TA increase after roughly 210 tonnes of alkaline addition, measured 60 m from the discharge (doi:10.1038/s44172-026-00665-w).
- **The equilibration problem, measured:** the Bergen mesocosm estimated that reaching 95% air–sea equilibration for a ΔTA of 600 µmol/kg would take **~1,050 days** (doi:10.5194/bg-23-137-2026). That empirically corroborates the modeling result that most sites plateau at 0.6–0.8 mol CO₂ per mol TA only after **3–4 years**, and up to **8–10 years** in places like Hawaii (He & Tyka, doi:10.5194/bg-20-27-2023).
That last point is the crux for anyone buying OAE credits: **the CO₂ uptake occurs over years, at a signal-to-noise ratio far below what any campaign can observe.** The tracer paper says the quiet part plainly — direct observations are needed *to validate the model-based MRV approaches currently in use.* Today, OAE credits are fundamentally model-issued, with in-water data constraining the near-field only.
**The lab literature adds a live efficiency risk:** commercially available ultrafine Ca(OH)₂ generally caused a *net loss* of alkalinity in seawater; Mg(OH)₂ caused net loss above ΔTA ~600 µmol/kg (Hartmann et al., doi:10.5194/bg-20-781-2023). Runaway CaCO₃ precipitation can cut uptake efficiency from ~0.8 to **0.1 mol CO₂ per mol TA** (Moras et al., doi:10.5194/bg-19-3537-2022).
**And a realism problem cuts the other way on impacts:** a 2026 JGR-Oceans study argues that ΔTA >1,000 µmol/kg — the perturbation common in OAE experiments — occurs in the real world only for minutes in a minuscule fraction of perturbed water, while the ubiquitous real range is **1–100 µmol/kg**, which is rarely the focus of environmental assessment (doi:10.1029/2025jc023598). So much of the ecological literature may be over-testing, not under-testing.
---
## 4. The evidence-to-commitment gap
Against that ~4-paper open-water base, the corpus holds **4 OAE offtakes totalling 560,691 tonnes**:
| Buyer / supplier | Tonnes | Value | Date |
|---|---|---|---|
| Frontier / Planetary (Halifax) | 115,211 | $31.3M (~$272/t) | 2025-08-26 |
| Frontier / CREW Carbon (New Haven) | 71,878 | $32.1M (~$447/t) | 2024-12-17 |
| Microsoft / Ebb Carbon (Sequim Bay) | 350,000 | undisclosed | 2024-10-24 |
| Microsoft / CREW Carbon | up to 23,602 | undisclosed | 2026-08-04 |
Two caveats the corpus flags itself: the Ebb deal is **1,333 t firm plus options up to 350,000 t**, so it is mostly optional volume, not contracted delivery; and the two CREW deals are wastewater alkalinity enhancement, classed under `oae` by a corpus owner decision, not open-ocean OAE. Also notable: **no tracked policy instrument in the corpus names OAE** across the 14 jurisdictions covered — this pathway is being financed by voluntary buyers well ahead of any compliance framework.
---
## 5. What's happened since the corpus cutoff (web, off-corpus)
The single most important OAE field result is newer than the paper fetch and **not in the corpus**: WHOI's **LOC-NESS** project ran the first EPA-permitted OAE field trial in US federal waters in the Gulf of Maine in **August 2025**, with preliminary findings presented at the Ocean Sciences Meeting in Glasgow in late February 2026 (WHOI via ecomagazine.com, oceannews.com, newswise.com). Per Chemistry World's May 2026 report, the team dispersed 65,000 litres of sodium hydroxide over 1 km² alongside 760 litres of rhodamine dye, tracking a 3,600 km² zone with three vessels and four AUVs plus gliders, drifters and satellite imagery. Secondary coverage (bode-living.com, Feb 2026) reports **2–10 tonnes of CO₂ removed in the four days after the release**, with the team estimating up to ~50 tonnes total and no significant adverse effects observed.
Read that carefully: it is a genuine milestone — the largest, best-instrumented deliberate open-ocean alkalinity release yet — and the observed removal is **single-digit tonnes over the observation window**, against contracted volumes in the hundreds of thousands. These are preliminary, conference-presented results; I was unable to confirm peer-reviewed publication status before exhausting my search budget.
---
## 6. Honest assessment and gaps
**How strong is the field evidence?** Weak in the specific sense that matters commercially: *no published open-water experiment has yet closed the carbon budget on an intentional alkalinity addition.* What field work has established is (a) alkalinity can be added without detectable acute local harm at conservative doses, (b) plumes dilute below detection within metres to hundreds of metres, and (c) the resulting CO₂ uptake signal is a few percent of background variability and unfolds over 1–10 years. Points (b) and (c) mean OAE MRV is structurally model-dependent, and the models are largely unvalidated against in-water carbon uptake.
**Where the evidence *is* reasonably strong:** carbonate chemistry limits and secondary-precipitation thresholds (well-replicated lab work), ecological response at mesocosm scale (several 39–53 day multi-level experiments now with consistent "moderate effects" findings), and regional capacity/equilibration modeling.
**Gaps I could not close:**
- The corpus returned 10 of 16 field-trial papers; the other 6 are unseen, so my "3–4 genuine open-water releases" is a floor.
- 58 OAE papers (19%) are unclassified for study type — the true field-trial share could be modestly higher.
- Study-type labels are LLM-assigned from abstracts and conflate mesocosm with open-water; that's my reclassification, not the corpus's.
- I could not verify Planetary's or Ebb's most recent *delivered and verified* tonnage under Isometric's OAE protocol, nor any peer-reviewed LOC-NESS paper, because my web search allowance ran out. Anyone underwriting this pathway should check those two things directly.
Judge’s note — All three gold checklist items are surfaced precisely (308 papers, 16 field trials at 6.4%, lab 89 / modeling 80 / review 52), with a clean breakdown table including the 58 unclassified records and a caveat that classification was LLM-derived. The answer goes well beyond bean-counting: it deflates the headline '16 field trials' by reading the actual records and separating genuine open-water alkalinity additions (~3) from mesocosms, dye-only proxy releases, and natural-analogue time series — exactly the kind of expert-relevant distinction the question is probing. Modeling-vs-observation tension is handled substantively (Ho et al. MRV synthesis on why observation alone cannot close the accounting; the CESM2/MARBL vs ECCO-Darwin divergence in the credit-relevant first 7 years; sensor-network underestimation of 35–55%; the ΔTA >1000 µmol/kg mis-specification critique). Citations are near-uniformly DOI-level with journal and date, so claims are checkable. Beyond-corpus additions are real and material: LOC-NESS preliminary results (Ocean Sciences Meeting, Feb 2026; 65,000 L NaOH, 2–10 t CO2 estimate, explicitly flagged as conference-stage, not peer-reviewed), Planetary's 138 net tonnes delivered to Frontier (Nov 2024), and the 625.6 t Isometric-verified issuance (June 2025) — with the sharp observation that the first verified credits are model-quantified rather than field-measured. Coverage honesty is exemplary: post-2026-08-14 gap, no preprints, non-English absence, search-tool limit, pending review status on one deal, derived-not-quoted prices. Minor deductions only for the offtake/policy section being somewhat tangential to the question as asked, and for unverifiable DOI strings that a reader must spot-check.
I'll start with the coverage tool, then pull the OAE evidence profile. Let me pull the full field-trial set and MRV literature, then check what's happened outside the corpus. Now the required step: checking what the corpus says it cannot cover — field trial results and literature after the 2026-08-14 paper fetch. ## Bottom line The OAE literature is large and growing fast, but the part of it that consists of **intentional alkalinity addition to open water with carbon uptake actually measured** is very small — roughly a dozen studies worldwide, most of which detected a signal only within metres of the discharge point, and only one of which (LOC-NESS, preliminary and unpublished as of the corpus cut) has attempted to quantify CO₂ uptake in the open ocean. Everything above that scale rests on models. That is not an accident of immaturity; the field's own MRV literature says direct observation alone *cannot* close the accounting. --- ## 1. The corpus counts: how the literature splits From the Carbon Curve evidence profile for `oae` (papers fetched 2026-08-14; OpenAlex CC0 metadata, 2010+, English; pathway queries are complete result sets, so these are field statistics rather than a sample): **308 OAE papers in scope**, of which 250 carry a study-type classification: | Study type | Count | % of classified | % of all 308 | |---|---|---|---| | Lab | 89 | 35.6% | 28.9% | | Modeling | 80 | 32.0% | 26.0% | | Review | 52 | 20.8% | 16.9% | | **Field trial** | **16** | **6.4%** | **5.2%** | | Survey | 8 | 3.2% | 2.6% | | LCA | 5 | 2.0% | 1.6% | | Unclassified | 58 | — | 18.8% | So the crude answer to your question: **modeling outnumbers field work 5:1**, and lab-plus-modeling is 68% of the classified literature. Field trials are about 1 in 16 papers. Two caveats on these numbers. Study types were assigned by an LLM from title and abstract (classifier v1, 2026-08-12), not by human coding, so edge cases are noisy. And 58 papers are unclassified largely because no rights-cleared abstract exists — the true field-trial count could be modestly higher. ## 2. The important deflation: "field trial" ≠ open-water alkalinity release This is where the headline 16 overstates the evidence. Reading the actual records, the field-trial set contains at least four different things: **(a) Genuine intentional in-water alkalinity addition — about 3 studies** - **Tasmania, Australia** — the first Australian OAE field trial, continuous NaOH addition at a coastal site. At the release point pCO₂ fell by up to 370 µatm with TA up ~545 µmol/kg, but **maximum downstream pCO₂ decreases were only 22–77 µatm, signal strengths of <1–5%**, and the plume dispersed "within meters of the site of addition" (Sci Rep, 2025-12-12, [10.1038/s41598-025-31606-w](https://doi.org/10.1038/s41598-025-31606-w)). - **Sequim Bay, Washington** — Ebb Carbon's electrochemical alkalinity released via a wastewater outfall at PNNL. pH in the mixing tank held at the permit ceiling of 9.0; at the outfall the **measurable signal was limited to ~2.5 m from the discharge pipe**. The paper is explicit that this is "proof-of-concept for a conservative small-scale release" (Front. Environ. Eng., 2025-09-18, [10.3389/fenve.2025.1641277](https://doi.org/10.3389/fenve.2025.1641277)). - **Mg(OH)₂ into wastewater discharged to the coastal ocean** — 74% reduction in aqueous CO₂ in the wastewater (efficiency 0.73 mol CO₂/mol alkalinity), but in the receiving ocean only a 27.2 µatm pCO₂ reduction and 0.017 pH units, and the authors state plainly that **"this proof of concept field trial did not achieve CDR due to its small scale"** (Commun. Earth Environ., 2024-06-28, [10.1038/s43247-024-01506-4](https://doi.org/10.1038/s43247-024-01506-4)). **(b) Mesocosms — enclosed water, not open ocean.** Bergen fjord, Norway (47 days, ΔTA 0–600 µmol/kg, calcium- and silicate-based, [10.5194/bg-23-137-2026](https://doi.org/10.5194/bg-23-137-2026)); Helgoland North Sea (39 days, six alkalinity levels to ΔTA 1250 µmol/kg, [10.5194/bg-23-4691-2026](https://doi.org/10.5194/bg-23-4691-2026)); the *Oikopleura dioica* zooplankton study (53 days, Norwegian fjord, [10.1371/journal.pone.0344503](https://doi.org/10.1371/journal.pone.0344503)). These are the strongest ecological evidence in the field, but they are closed systems and they measure biology, not verified net CDR. **(c) Proxy releases with no alkalinity at all.** The Martha's Vineyard study released **rhodamine dye only**, as a stand-in for alkalinity, then constructed a "hypothetical alkalinity release" — and even in that idealised construction, the calculated CDR signal over 36 hours amounted to **10% of the total potential CDR** ([10.5194/bg-22-5511-2025](https://doi.org/10.5194/bg-22-5511-2025)). **(d) Natural analogues and baseline surveys.** A 26-year South China Sea time series showing natural TA rising 0.56 µmol/kg/yr ([10.1029/2025gl117872](https://doi.org/10.1029/2025gl117872)); oyster-farming mesocosms; bivalve shell dissolution in Narragansett Bay; a Hawaiian sediment microbial baseline for a future olivine campaign. Useful context, not intervention evidence. **Net: of 16 "field trials," roughly three are intentional open-water alkalinity additions, and none of those measured net carbon removal — they measured whether the chemical perturbation was detectable at all.** The honest reading is that the field has demonstrated *detectability of the alkalinity signal*, not *quantification of the CO₂ uptake*. ## 3. Why modeling isn't a substitute — and isn't yet converged The 32% modeling share is not laziness. The field's own MRV synthesis (Ho et al., *State of the Planet*, 2023-11-27, [10.5194/sp-2-oae2023-12-2023](https://doi.org/10.5194/sp-2-oae2023-12-2023)) states that because equilibration takes months or longer and added alkalinity dilutes below background variability, **"comprehensive quantification of carbon removal via OAE will be impossible through observational methods alone, and numerical simulations will be required."** But the models disagree with each other. A 2026 comparison of CESM2/MARBL against ECCO-Darwin, using precisely matched injection patches and seasons, found **"substantial differences… well beyond the expected intrinsic variation of each model"** in subtropical regions, mostly attributable to vertical transport representation, with the largest divergence in the first 7 years post-injection ([10.5194/bg-23-4943-2026](https://doi.org/10.5194/bg-23-4943-2026)). That is the credit-relevant window. Sensing is also a binding constraint. A blind mesocosm plus Sequim Bay field test found sensors *could* distinguish an mCDR signature from natural variability under optimal conditions, but that idealised networks of **3 to 100 sensors would underestimate the net change in surface pH by at least 35–55%** ([10.3389/fclim.2025.1649723](https://doi.org/10.3389/fclim.2025.1649723) — this record is unclassified by study type in the corpus). And the mesocosm evidence base may be systematically mis-specified: a 2026 modeling paper argues **ΔTA >1,000 µmol/kg — "a perturbation common in OAE experiments" — occurs only for minutes in a minuscule fraction of real-world perturbed water**, while the realistic range of 1–100 µmol/kg is under-studied, risking overestimation of ecological effects ([10.1029/2025jc023598](https://doi.org/10.1029/2025jc023598)). For the solid-feedstock variant the gap is starker still: a 2025 review of olivine-based marine enhanced rock weathering concluded that **no peer-reviewed results from larger-scale coastal field studies existed**, and that "the current knowledge base is not sufficient to predict the outcome of in situ mERW applications" ([10.5194/bg-22-355-2025](https://doi.org/10.5194/bg-22-355-2025)). ## 4. What I found by searching beyond the corpus The corpus warns it holds nothing published after 2026-08-14 and no preprints or conference results. That matters here, because the single most important field result is exactly that kind of item: - **LOC-NESS (Gulf of Maine), preliminary results announced 25 February 2026.** WHOI presented findings at the Ocean Sciences Meeting in Glasgow from the **first US EPA-permitted OAE field trial**, conducted in federal waters in August 2025 (whoi.edu press release; Newswise, 25 Feb 2026). Per Chemistry World's reporting (May 2026), the team dispersed **65,000 litres of sodium hydroxide over 1 km²** plus 760 litres of rhodamine dye, tracked across a 3,600 km² zone using three vessels, four AUVs and NASA's PACE satellite, and estimated that **an additional 2–10 tonnes of atmospheric CO₂ were drawn into the ocean over the four-day monitoring period**. Note the width of that range — a factor of five on a single, heavily instrumented trial — and note that it is preliminary and conference-stage, not peer-reviewed. This is the first open-ocean OAE trial to put a number on uptake at all, and it is not in the corpus. - **Commercial verification has run ahead of the peer-reviewed record.** Planetary Technologies delivered **138 net tonnes** of OAE removal from Halifax Harbour to Frontier (96 t to Shopify, 42 t to Stripe) in November 2024, and in June 2025 announced **625.6 tonnes of independently verified credits** to Stripe, Shopify and British Airways — verified by Isometric under its Ocean Alkalinity Enhancement from Coastal Outfalls protocol with quantification informed by [C]Worthy, from 1,103.2 tonnes of alkalinity dosed at Tufts Cove (planetarytech.com; Carbon Herald, 17 Jun 2025). So the *first* verified OAE credits are quantified by a registry protocol leaning on models, not by a published field measurement of uptake. I was unable to complete further searches (search tool limit reached), so I could not check for developments after mid-2026 — including whether LOC-NESS results have since been published, Planetary's cumulative issuance, or the status of EU CRCF methodology work on OAE. ## 5. Commercial and policy standing, for contrast The corpus holds **4 OAE offtakes totalling 560,691 tonnes**, 2 with disclosed value: | Buyer → Supplier | Date | Tonnes | Value | Derived $/t | |---|---|---|---|---| | Frontier → Planetary (Halifax, CA) | 2025-08-26 | 115,211 | $31.3M | **$272** | | Frontier → CREW Carbon (New Haven, US) | 2024-12-17 | 71,878 | $32.1M | **$447** | | Microsoft → Ebb Carbon (Sequim Bay, US) | 2024-10-24 | 350,000 | undisclosed | — | | Microsoft → CREW Carbon | 2026-08-04 | up to 23,602 | undisclosed | — | Both prices are **derived by division, not quoted**. The Microsoft–Ebb deal is 1,333 t firm plus options up to 350,000 t, so most of that volume is optional, not contracted. The Microsoft–CREW record is **`review_status: pending`** — not yet human-reviewed, its pathway assignment is flagged as needing a human call (wastewater alkalinity enhancement has no dedicated taxonomy code), and "up to 23,602" is a ceiling. Corpus-wide, only 21 of 59 deals disclose value, covering 4.5% of contracted tonnes, so these two prices describe the disclosing minority. The contrast is the point: **~560,000 tonnes contracted against a delivered-and-verified record in the hundreds of tonnes**, and a field literature that has not yet directly measured net removal. On policy, the corpus coverage note is blunt: **no tracked policy instrument names OAE.** The corpus does hold a CDR Policy Scoop episode (2025-10-01) on the European Commission's first technical workshop on ERW and OAE under the CRCF, so process is underway — but no certified methodology is recorded, and policy events are only current to 2026-07-17. ## 6. What neither source covers - **Post-2026-08-14 literature, preprints, and non-English work** are structurally absent from the corpus, and I could not fill that gap by search. - **Peer-reviewed LOC-NESS results**, if published, would materially change the picture — a conference press briefing is the weakest tier of evidence for a number as consequential as 2–10 t. - **Registry-verified quantification** (Isometric protocol, [C]Worthy modeling) sits outside both the paper corpus and the deal corpus; its underlying assumptions are on the Isometric Registry, not in peer review. - **Long-term ecological effects and secondary CaCO₃ precipitation at deployment scale** — the lab work is clear that runaway precipitation can cut efficiency from ~0.8 to 0.1 mol CO₂ per mol TA ([10.5194/bg-19-3537-2022](https://doi.org/10.5194/bg-19-3537-2022)), but no field study has tested this at operational scale. - **Non-OECD and Global South siting**, where much modeled potential is concentrated (North Madagascar, Peru, Brazil per [10.5194/bg-20-27-2023](https://doi.org/10.5194/bg-20-27-2023)) but where both the trial record and the corpus's jurisdiction coverage are thin. The database offers no ratings or rankings and I have not invented any; the framing above is my reading of the records, and where I've inferred (e.g. reclassifying "field trials" by what they actually did) I've shown the studies so you can check the call.
Judge’s note — All three gold checklist items are hit precisely and quantitatively: 308 in-scope papers with the explicit note that this is a complete result set as of 2026-08-14; 16 field trials at 5.2% (matching ~6%); and the lab (89) / modeling (80) / review (52) distribution laid out in a table with shares. The answer goes well beyond bookkeeping by interrogating the 'field trial' label itself — decomposing the 10 visible records into 2 genuine open-water alkalinity releases (Tasmania NaOH, Scientific Reports 2025-12-12, with the crucial detail that downstream pCO2 signals were only 22–77 µatm / <1–5% signal strength; Sequim Bay/Ebb Carbon, Frontiers 2025-09-18, measurable signal within ~2.5 m of outfall), 3 enclosed mesocosms (Bergen, Helgoland with the silica ballasting 50–60% decline), and 4 that are not alkalinity additions at all (rhodamine tracer proxy, natural-analogue trends, oyster mesocosms, bivalve shell dissolution). That is exactly the kind of critical reading an expert practitioner needs and it strengthens the 'only ~2 real open-water trials' conclusion. The structural explanation for modeling dominance (He & Tyka 3–4 yr to 0.6–0.8 mol/mol plateau, half the CDR lost in downwelling regions; ~1,050 days to 95% equilibration; Martha's Vineyard detecting only 10% of potential CDR after 36 h) is well chosen and correctly framed as 'model-based MRV is not a shortcut, it is the only way to close the mass balance.' Lab constraints (Moras 0.8→0.1 mol/mol under runaway precipitation; Hartmann net TA loss above ~600 µmol/kg and for ultrafine Ca(OH)2) are load-bearing and correctly attributed. The 'Lethal by Design?' dose-mismatch argument is a sophisticated addition. Citation quality is excellent: nearly every quantitative claim carries journal, date, and often citation count, making claims checkable. Coverage honesty is exemplary — it flags auto-classification by title/abstract (v1), 58 unclassified papers (19%) as a source of drift, that it could only see 10 of 16 field-trial abstracts, English-only/2010+/no-preprint/no-post-cutoff coverage, exclusion of grey literature (Ocean Visions, [C]worthy, NOAA, Frontier OAE Primer), the deal layer being a curated floor rather than a census, and explicitly labels the Frontier 'first verified OAE credits' claim as a buyer's assertion rather than verified record. It also names the right next step (Isometric registry filings). Clarity is high: headline answer in the first paragraph, tables, and a memorable framing ('three review articles for every field trial'). The main shortfall is on the beyond-corpus probe: the answer stays essentially inside the corpus. Every named trial (Tasmania, Sequim Bay, Bergen, Helgoland, Martha's Vineyard, Lethal by Design) is pre-2026-08-14 corpus content, no post-cutoff papers or preprints are surfaced, and the only near-beyond-corpus material (Frontier's 2026 Innovation page claims about Planetary's verified 2025 deliveries and gas-exchange evidence) is itself presented as a corpus-flagged uncorroborated signal rather than independently verified. It correctly identifies registry documentation as the missing layer but does not go get it, so the single most decision-relevant open question — what Planetary's and Ebb's actual delivered/verified tonnes show — remains unresolved.
I'll start by checking what the corpus actually covers, then pull the OAE evidence profile. The profile gives me the headline split. Let me pull the full field-trial set, the MRV literature, and check policy/registry standing. Two searches errored on query complexity. Retrying with simpler terms. ## Headline OAE has a large and fast-growing literature, but the in-water evidence base is thin, recent, and — critically — has never measured the thing being sold. In the Carbon Curve corpus, which holds a **complete result set for the OAE pathway query as of 2026-08-14**, there are **308 in-scope OAE papers**, split as: | Study type | Count | Share of 308 | |---|---|---| | Lab | 89 | 28.9% | | Modeling | 80 | 26.0% | | Review | 52 | 16.9% | | **Field trial** | **16** | **5.2%** | | Survey | 8 | 2.6% | | LCA | 5 | 1.6% | | Unclassified | 58 | 18.8% | So the direct answer: **roughly 1 paper in 20 involves field work of any kind; lab plus modeling is 55%, and reviews alone outnumber field trials more than 3:1.** A literature that produces three review articles for every field trial is a literature still arguing about what it expects to find. ## The 16 "field trials" are softer than the label suggests The evidence profile lists 10 of the 16. Sorting them by what was actually done: **Genuine intentional alkalinity release into open water — 2:** - **Tasmania, NaOH release** (*Scientific Reports*, 2025-12-12) — "the first OAE field trial in Australia." At the release point, alkalinity rose ~545 µmol kg⁻¹ and pCO₂ fell by up to 370 µatm. But **downstream pCO₂ decreases were only 22–77 µatm, described by the authors as "signal strengths of <1–5%,"** with the plume dispersing "within meters of the site of addition." - **Sequim Bay, Washington — Ebb Carbon electrochemical alkalinity via a wastewater outfall** (*Frontiers in Environmental Engineering*, 2025-09-18). pH raised to the permitted maximum of 9.0 in the mixing tank; **"the region with a measurable signal was limited to within ~2.5 m of the discharge pipe."** No measurable change in temperature, salinity, turbidity or oxygen. **Enclosed mesocosms (in the sea, but not open-system) — 3:** the Bergen fjord campaign (47-day, ΔTA 0–600 µmol kg⁻¹, calcium- and silicate-based, *Biogeosciences* 2026-01-08), its zooplankton companion paper (53-day *Oikopleura dioica* study, *PLoS ONE* 2026-03-30), and the Helgoland North Sea mesocosms (39-day, ΔTA up to 1,250 µmol kg⁻¹, *Biogeosciences* 2026-07-08). **Not alkalinity additions at all — 4:** a rhodamine **tracer** release south of Martha's Vineyard used as a *proxy* for OAE (*Biogeosciences*, 2025-10-10); a 26-year natural-alkalinity trend in the South China Sea (*GRL*, 2025-09-02); oyster-farming field mesocosms (*PNAS*, 2025-09-02); and natural bivalve-shell dissolution in Narragansett Bay (*L&O Letters*, 2025-06-30). Plus one instrumentation paper on high-frequency in situ TA sensors (*Communications Engineering*, 2026-04-22). On a strict reading — deliberate alkalinity added to open water, with in-water tracking of the resulting plume — the published field record in this corpus is **about two trials, both from 2025, both at metres-to-hundreds-of-metres scale.** ## The structural reason modeling dominates This is not laziness; it is physics. The CO₂ uptake that constitutes the removal happens over **years and hundreds of kilometres**, far outside any measurable plume: - He & Tyka (*Biogeosciences*, 2023-01-03, 145 citations) model an uptake plateau of **0.6–0.8 mol CO₂ per mol alkalinity reached only after 3–4 years**, up to **8–10 years off Hawaii**, with **up to half the CDR potential lost to bottom waters in downwelling regions such as around Iceland.** - The Bergen mesocosm measured real gas exchange (up to −15 mmol C m⁻² d⁻¹) and then estimated **~1,050 days to 95% equilibration** for ΔTA of 600 µmol kg⁻¹. - The Martha's Vineyard tracer study found that after 36 hours the detectable CDR signal — a 4 µatm fCO₂ rise, 0.004 pH drop, 1.8 µmol kg⁻¹ DIC increase — represented **only 10% of the total potential CDR.** Detection thresholds were TA >10 µmol kg⁻¹, pH >0.01, fCO₂ >10 µatm; a "dynamic baseline" cut baseline variability by 25% for fCO₂ and 60% for TA. So model-based MRV is not a shortcut around measurement — it is currently the only way to close the mass balance. In-water data can verify the near-field perturbation and validate model initial conditions; it cannot yet verify a tonne. ## Where lab work is genuinely load-bearing The strongest quantitative constraints on OAE efficiency come from tanks, not oceans, and they are cautionary: - **Moras et al.** (*Biogeosciences*, 2022-08-01, 177 citations): runaway CaCO₃ precipitation can cut uptake efficiency **from ~0.8 to 0.1 mol CO₂ per mol TA added**. Safe below ΩA ≈ 5; aragonite precipitation above ΩA ≈ 7. - **Hartmann et al.** (*Biogeosciences*, 2023-02-20, 150 citations): reactive alkaline solids cause **net alkalinity loss above ΔTA ≈ 600 µmol kg⁻¹** for Mg(OH)₂, and **commercially available ultrafine Ca(OH)₂ caused net TA loss across all tested additions** — a direct challenge to ship-based slurry designs. These are the numbers that discipline the field, and they are laboratory numbers. ## Ecological evidence: reassuring, but tested at the wrong doses Mesocosm results are mostly benign — *Oikopleura dioica* showed no response across ΔTA 0–600; a Tasmanian coastal phytoplankton community showed "generally moderate" effects at ΔTA ~500 (*Biogeosciences*, 2022-12-01). One negative signal stands out: the Helgoland mesocosms found **silica ballasting ratios fell by up to 50% (suspended) and 60% (exported)** at pH >9.25, which the authors say could shoal remineralisation and **"weaken net CO₂ removal."** But a 2026 paper — "Lethal by Design?" (*JGR Oceans*, 2026-08-01) — argues the whole experimental design is misaligned: **ΔTA >1,000 µmol kg⁻¹, "a perturbation common in OAE experiments," occurs in the real world "only for minutes in a minuscule fraction" of perturbed water**, whereas the ubiquitous real-world range of **1–100 µmol kg⁻¹ is "not commonly in focus."** The ecological literature may therefore be systematically overestimating risk while the efficacy literature underestimates scale. ## Evidence versus money and policy Contracting has run ahead of the field record. Four OAE deals in the corpus, **560,691 tonnes contracted**, two with disclosed value: - **Frontier ↔ Planetary**, 2025-08-26: 115,211 t for **$31.3M (~$272/t derived)**, pre-dissolved Ca/Mg oxide at Tufts Cove, Halifax, delivery 2026–2030 — largest OAE offtake at announcement. - **Frontier ↔ CREW Carbon**, 2024-12-17: 71,878 t for **$32.1M (~$447/t derived)** — wastewater alkalinity enhancement, mapped to `oae` by an explicit owner decision on 2026-08-14, not an open-ocean intervention. - **Microsoft ↔ Ebb Carbon**, 2024-10-24: headline 350,000 t, but the record notes the structure is **1,333 t firm plus options** — Sequim Bay, verified under Isometric's OAE protocol. - **Microsoft ↔ CREW Carbon**, 2026-08-04: up to 23,602 t (ceiling, not floor; record still pending review). Policy standing is close to nil. The corpus states plainly: **"No tracked policy instrument names this pathway."** The July 2026 EU ETS review proposal would centrally purchase 250 Mt of CRCF-certified removals over 2031–2040 but **initially covers only DACCS and BioCCS**. Puro's active methodology list does not include OAE. Isometric — the one registry with an OAE protocol in use here — became CORSIA-eligible on 2025-11-26, which is the nearest thing to a compliance channel. The Commission held its first technical workshop on ERW and OAE under the CRCF around October 2025 (*CDR Policy Scoop*, 2025-10-01). Notably, **Frontier itself still lists the core questions as open.** In an uncorroborated signal from its 2026 Innovation program page, Frontier names "determining how to add alkalinity to the ocean efficiently and safely" and "building precise scalable measurement tools for open systems" as target funding gaps — while also claiming OAE air-sea gas exchange "is now shown to be a net sink with appropriate site selection" (citing Planetary deployment data plus Dalhousie and [C]worthy modeling) and that **Planetary delivered the world's first verified OAE credits in 2025** under the Isometric protocol. Treat that as a buyer's claim, not a verified record — the corpus flags it uncorroborated, and I found no peer-reviewed paper in the corpus reporting those delivery results. ## Caveats on these numbers - **Study types were auto-classified** by Claude from title and abstract (v1, 2026-08-12), not human-curated. The exact split is directional. **58 papers (19%) are unclassified**, mostly where no rights-cleared abstract exists — if those skew toward any category the percentages shift. - The `field-trial` label conflates **enclosed mesocosms, open-water releases, natural-analogue observation and tracer proxies.** My decomposition above is my own reading of the 10 abstracts returned; **I could not see the other 6** of the 16 and cannot characterise them. - Coverage is **OpenAlex metadata, 2010+, English only, no preprints, nothing after 2026-08-14.** Non-English work and the substantial grey literature (Carbon to Sea, [C]worthy, NOAA and Ocean Visions technical reports; the 2025 Frontier *OAE Primer*) are outside the paper layer. - Deal coverage is a curated floor, not a census — CDR.fyi is the fuller transaction scoreboard. Only 21 of 59 deals corpus-wide disclose value, covering 4.5% of contracted tonnes, and price disclosure is concentrated in Frontier. - **Notable absence:** I found no paper in this corpus reporting in-water verification of a *commercially delivered* OAE tonne. That gap between two metre-scale trials and 560,691 contracted tonnes is the single most important fact here, and it should be read as a genuine gap rather than a search artefact — though a targeted search of registry documentation (Isometric's published Planetary and Ebb project records) would be the right next step, since registry filings are not a layer this corpus holds.
Asked, verbatim: “What does the research literature say about enhanced rock weathering's MRV challenges, and which companies are selling ERW removals anyway?”
Judge’s note — Substantively strong on the science: it surfaces six distinct, concretely-described MRV failure modes (cation- vs inorganic-carbon accounting divergence, lysimeter/centrifuge sampling bias, radiocarbon/Sr-isotope cost-accuracy tradeoffs, expert-elicitation uncertainty on loss pathways and deep-soil/aquatic validation, CarbonPlan's protocol-comparison critique of Puro v1.0 vs CSI v0.9, and entanglement with trace-metal/environmental monitoring), plus named modeling references (Calabrese 2022, Baek 2023). Commercial coverage is broad and dated: Frontier–Eion ($33M, ~78.7 kt, 2027–2030), Lithos Carbon and UNDO as Microsoft counterparties, Terradot's acquisition of Eion and the combined 400k+ t book, Google's 200 kt Terradot commitment, InPlanet–Microsoft 28.5 kt under Isometric, Carbon180 market-share figures, DOE pilot procurement. The registry/protocol layer (Puro, Isometric, CSI) is explicitly handled, and Frontier–Eion/Lithos are both named. It misses the corpus-level meta point entirely — no acknowledgement of the ERW literature being held as a complete/large result set or that quantification is a tracked topic, so no sense of how exhaustive the underlying evidence base is. Citation quality is middling: journals and organizations are named and quoted verbatim, but most papers lack authors, titles, or years, making verification effortful; a couple of claims (InPlanet's 'world's first issued credits' vs Eion's Puro issuance) are internally inconsistent, though the answer flags this itself. Coverage honesty is exemplary — it cleanly separates verified from background knowledge, lists five specific unverified gaps, names the sources it would need (CDR.fyi, SCEPTER, Beerling 2024, Clarkson 2024), and calls out strong-acid weathering as the most consequential unchecked mechanism. Structure and decision-usefulness are excellent for a practitioner.
I'll research this thoroughly — both the scientific literature on ERW measurement challenges and the current commercial landscape. Good starting material. Let me dig into the specific scientific papers and the current market data. My search budget is exhausted, so below I distinguish clearly between claims backed by the sources I retrieved (cited) and points from my background knowledge that I could **not** verify in this session (flagged as unverified). --- # Enhanced rock weathering: the MRV problem, and who is selling anyway ## 1. The literature's core verdict: MRV is *the* bottleneck The framing in the peer-reviewed literature is unusually blunt. In *Environmental Science & Technology*, a study of a 40-year-old weathering site states that the main challenge associated with enhanced weathering is monitoring, reporting, and verifying the amount of carbon removed as a result of enhanced weathering reactions — a sentence repeated nearly verbatim in a separate ES&T paper on agricultural-land impacts, which adds that although a geochemical modeling framework for ERW in soils is emerging, there is a scarcity of experimental and field trial data exploring potential environmental impacts, risks, and monitoring strategies associated with this practice . The structural reason, as framed by researchers at Oxford: quantification of uncertainty in open-system carbon removal pathways presents a particular challenge, and although enhanced weathering has gained attention for its scalability, low cost and co-benefits for farmers, several outstanding questions remain in quantifying net carbon removal . Unlike DAC + geologic storage, where you meter CO₂ into a well, ERW's product — dissolved bicarbonate alkalinity — leaves the field of measurement and travels through soil, groundwater, rivers and eventually the ocean over years to decades. ## 2. Specific, documented failure modes in measurement **(a) Cation-based vs. carbon-based accounting disagree.** A 2024 *Frontiers in Climate* study compared the two dominant approaches and concluded that despite being a promising CDR method, enhanced weathering has been limited by uncertainty surrounding the measurement of CO₂ sequestration; the study compared cation-based calculations of CDR against inorganic carbon-based calculations, and soil results against leachate results . Its title is the finding: inorganic carbon-based approaches *may be required to complement* cation-based approaches — i.e., counting released Ca/Mg/K/Na is not on its own sufficient evidence that CO₂ was consumed. **(b) Pore-water sampling hardware is itself a source of bias.** A study using a forest ERW trial in Wales found that MRV of ERW remains challenging due to existing limitations of aqueous-phase sampling methodologies, such as passive and tension lysimeters, which may not fully capture weathering fluxes across varying soil moisture conditions; the study assessed centrifugation-based pore water extraction and compared lysimeter- and centrifuge-derived chemistry from feedstock-amended and control plots . Practically: two labs sampling the same plot with different instruments can get different answers. **(c) Isotopic/tracer methods are powerful but costly.** The Consett (Co. Durham, UK) study used radiocarbon, δ¹³C, ⁸⁷Sr/⁸⁶Sr and major element data at a site where steel slags weathered in a landscaped deposit for over 40 years . A companion *Frontiers* measurement review notes that radiocarbon is expensive to analyse and requires specialist sample preparation and accelerator mass spectrometry, but can provide valuable insights into the age of enhanced-weathering products and pin-point modern atmospheric carbon sources with lower uncertainty than traditional stable carbon and oxygen isotopes . That is a real cost/accuracy trade-off for any project trying to verify thousands of hectares. **(d) Loss pathways are the least-constrained term — per experts themselves.** A formal expert elicitation on agricultural enhanced weathering (published in *Communications Earth & Environment*) found that uncertainty in the magnitude of loss pathways is still high , and that this reflects the challenge of measuring weathering processes directly; many respondents indicated the need for either improved and well-validated models or further empirical study, especially in-field trials encompassing deep soils, with validation of models for processes occurring in deep soils and the aquatic system considered relatively poor . On downstream transport specifically, the elicitation notes river transport may not significantly limit EW's CDR potential in most US rivers, though these estimates are chemistry-focused and do not always include the full picture . **(e) Protocols disagree about what evidence is even required.** CarbonPlan's assessment is that one persistent challenge around enhanced weathering MRV is the lack of common understanding about what different quantitative methods — measurements, models, and forms of record keeping — can actually tell us about net carbon removal . At the time of that analysis, they compared the only two public enhanced weathering protocols available — the Puro Enhanced Rock Weathering Methodology v1.0 and the Carbon Standard International Global Rock C-Sink Guideline v0.9 — alongside the methods used in the Beerling et al. field-trial estimate for context . CarbonPlan has separately published a synthesis of carbon removal estimates from the enhanced weathering literature under the title "Does enhanced weathering work? We're still learning," drawing on work including Calabrese et al. (2022), *Nano- to Global-Scale Uncertainties in Terrestrial Enhanced Weathering*, and Baek et al. (2023) on climate controls on global ERW capacity . **(f) MRV is entangled with environmental-risk monitoring.** The ES&T agricultural paper identifies potential negative consequences and positive co-benefits of ERW scale-up and suggests mitigation approaches — the practical point being that trace-metal and soil-chemistry monitoring is a separate obligation layered on top of carbon accounting. ### Additional mechanisms I could not verify in this session (treat as unconfirmed) From background knowledge, the literature also emphasizes: **strong-acid weathering** (nitric/sulfuric acid from fertilizer nitrification dissolving rock with no net CO₂ drawdown, which would inflate cation-based estimates on cropland); **secondary clay/carbonate formation** trapping cations and alkalinity; **CaCO₃ re-precipitation downstream**, which returns roughly a portion of the captured carbon; **baseline/counterfactual noise** from background weathering and soil heterogeneity relative to a small signal; **multi-year lag** between application and realized removal, versus buyer demand for near-term delivery; and **grinding/transport life-cycle emissions** trading off against reactivity at finer grain sizes. I was unable to retrieve primary citations for these in this session and would not present the numbers without re-checking Kanzaki/Zhang/Reinhard (SCEPTER modeling), Beerling et al. (2024, Illinois trial), Clarkson et al. (2024 review) and Dietzen/Rosing directly. ## 3. Who is selling ERW removals anyway Commercial activity has run well ahead of methodological consensus. Puro launched the first carbon crediting methodology for ERW in 2022 . **Market size and delivery.** Carbon180 puts the scale in perspective: ERW currently makes up less than half a percent of purchased carbon removal on the voluntary market, but just over 2% of delivered carbon removal — i.e., small in contracted volume, disproportionately large in what has actually been handed over. On policy, the US DOE has listed enhanced rock weathering as a carbon removal technology under its pilot procurement plan, though there aren't currently any purchases planned or in place at the federal level . **Terradot (US/Brazil) — now the largest player.** Terradot announced in early February 2026 that it had agreed to acquire the assets of Eion, a US ERW company known for olivine-based deployments and some of the first issued ERW credits . The two combined hold carbon removal contracts totalling 400,000+ tons with buyers including Microsoft, Google and Frontier , and Eion had recently secured its first carbon removal credits through Puro.earth from part of its 2023 deployment . Terradot's own MRV posture: it is investing in AI, synthetic data and process-based models, has partnered with registries including Puro and Isometric, and says it aims to exceed standard verification protocols . Google reportedly made a record 200,000-ton commitment to Terradot . **Eion (pre-acquisition).** In March 2025, Eion announced a $33 million carbon removal purchase agreement through the buyer coalition Frontier, with purchases by Google, Shopify, JPMorgan Chase and H&M Group among others, under which Eion will remove more than 78,700 tons of CO₂ between 2027 and 2030 . Note a reporting discrepancy worth flagging: one secondary source describes this as "Google's $33 million offtake agreement with Eion Carbon to purchase 79,000 tons of carbon credits by 2030" , whereas the contemporaneous trade coverage attributes it to the Frontier coalition with Google as one participant. Methodologically, Eion deploys olivine on farmland across the Mid-Atlantic US, Kenya and elsewhere, using a patented "soil fingerprinting" method . **UNDO (UK/Canada).** Microsoft announced it would purchase carbon removal from UNDO, to be carried out over the following couple of decades ; UNDO reports that Microsoft purchased 28,900 tonnes of permanent carbon removal from UNDO, supported by debt financing from Inlandsis , and separately that Barclays signed an agreement . Earlier Microsoft deals included UNDO for 15,000 tons in Canada and the UK, Eion for 8,000 tons in the US, and Lithos Carbon for 11,400 tons . UNDO describes its accounting as including operational emissions inputs so it can report net carbon removal over time, using conservative methods designed to support independent verification as standards evolve . **InPlanet (Brazil).** InPlanet announced in June 2026 what it called the world's first enhanced rock weathering carbon removal credits issued , noting that early purchases from catalytic buyers such as Frontier (Stripe & Shopify), Milkywire (Klarna & WRLD Foundation) and Klimate were critical in enabling its first deployments . It has also inked a deal to sell 28,500 ERW credits to Microsoft, certified under the UK-based standard Isometric and generated from its Brazilian programme . Third-party assessment: with Microsoft, Adyen and AXA Switzerland as named buyers and credits rated 'A' by BeZero Carbon . **Caution:** the "world's first issued credits" claim sits awkwardly against Eion being described as having some of the first issued ERW credits via Puro — competing "first" claims here are registry-specific marketing and should not be read as settled fact. **Others.** Lithos Carbon is a named Microsoft counterparty (above). Frontier has signed $3.7 million in contracts for enhanced weathering projects with eight startups, with neither Vesta nor UNDO involved in those particular contracts — Vesta being a coastal/marine variant rather than cropland ERW. ## 4. How to read the gap The honest summary is that buyers are underwriting a pathway whose central measurement question — how much of the alkalinity generated actually becomes durable, atmosphere-derived carbon storage — the field itself describes as poorly validated, particularly for processes occurring in deep soils and the aquatic system . The commercial response has been to lean on registry protocols (Puro, Isometric, CSI) plus conservative deduction factors, but CarbonPlan's point stands that there is not yet common understanding of what these quantitative methods can actually tell us about net carbon removal . **Gaps in this analysis I want to be explicit about:** I could not verify (i) the specific per-hectare removal rates from the major field trials, (ii) the current version numbers and deduction structures of the Puro and Isometric protocols as of today, (iii) total tonnes *issued* versus *contracted* across the sector (CDR.fyi would be the source), (iv) whether the Terradot–Eion transaction has closed, and (v) pricing. Several ERW-specific mechanisms I listed in §2 as "unconfirmed" — especially strong-acid weathering on fertilized cropland — are, in my judgement, the most consequential open items for anyone diligencing these credits, and deserve primary-source checking.
Judge’s note — Gold item 1 is only partially served: the answer does characterize a corpus-derived ERW literature set with a study-type breakdown (88 lab, 65 modeling, 60 field, etc.) and notes the fetch cutoff, but it reports 356 'in-scope' papers rather than the ~1,500+ complete result set, and it does not surface quantification/MRV as a tracked corpus topic — the topical framing it gives is method-type, not the tracked-topic layer. Items 2 and 3 are clearly hit: Frontier–Lithos (first ERW offtake, $57.1M) and Frontier–Eion ($33M, ~$419/t) appear in the deal table, and Section 3 covers the certification layer in depth (Isometric Enhanced Weathering in Agriculture protocol, 1,000-yr durability floor, CORSIA eligibility; Puro ERW methodology with the sharp caveat that CCP eligibility is programme-level not methodology-level; EU CRCF delegated-act status). Beyond-corpus content is unusually rich and specific: named MRV criticisms (the ES&T 'overlooked soil acidity' overestimation claim with the correct mechanism — nitric/sulfuric acid dissolution releasing cations without CO2 consumption, breaking cation-based mass balance; Schiedung et al. Nature Reviews on the unverified land–ocean continuum and Ni/Cr in ultramafics; two 2026 null-result field/lysimeter trials with the excellent framing that absence of signal is ambiguous between 'no removal' and 'wrong instrument'; SOC destabilization confounding single-pool accounting; site-dependence defeating generic application rates; QXRD as genuine methodological progress) plus off-corpus seller/market facts (InPlanet's 235.53 Isometric-issued credits to Adyen via ClimeFi, Microsoft–InPlanet, Mati Carbon XPRIZE, Tambora–Wild Assets, Boeing/Supercritical, Mitsui O.S.K., CDR.fyi's 55%-sold/24%-delivered/12%-retired delivery gap, and the EU ETS 250 Mt proposal covering only DACCS/BioCCS). Citations are precise and checkable (journals, dates, DOIs, author lists, registries, regulation numbers). Honesty is exemplary: it flags the ES&T item as metadata-only (title claim, not verified magnitudes), notes derived rather than quoted $/t, names the corpus inclusion thresholds that hide small ERW sellers, states all three cutoff dates, and lists four specific unresolved questions it could not verify due to a search-quota limit rather than passing silence off as absence. The synthesis ('conflicting evidence, not settled doubt'; literature moving opposite to deal flow; acid-source critique would be a correction to issued tonnes) is decision-useful for an expert. Docked one point for the corpus-scale/tracked-topic miss and for a slight risk that some very recent DOIs are unverifiable.
I'll start by mapping what the corpus covers, then dig into the ERW evidence base. Now let me pull the MRV-specific literature and the seller landscape. Now let me verify against the current web, since the corpus stops at mid-August 2026. ## Short version The literature's core finding is that ERW's problem isn't whether basalt weathers — it's that **you cannot measure the CO₂ in an open system**, so every commercial tonne rests on an inference chain with several unresolved links. Meanwhile roughly **560,000 tonnes of ERW have been contracted in tracked offtakes alone at disclosed prices of $300–419/t**, with deliveries already running since 2024. That gap is the story. --- ## 1. What the research says about MRV **Scale of the evidence base.** The Carbon Curve corpus holds 356 in-scope ERW papers, split as 88 lab, 67 unclassified, 65 modeling, 60 field-trial, 57 review, 12 LCA, 7 survey (papers fetched to 2026-08-14). Field trials are only ~17% of the literature — the pathway's ambition rests disproportionately on modeling. The most-cited works are still projections, not measurements: Beerling et al. (*Nature*, 2020) on cropland potential (675 citations) and Strefler et al. (*ERL*, 2018), which put costs at ~$60/tCO₂ for dunite but ~$200/tCO₂ for basalt and flagged that "the main uncertainties about the potential relate to weathering rates and rock mass that can be integrated into the soil." **The specific MRV failure modes now in the literature:** **(a) Dissolution ≠ removal — the acid-source problem.** This is the most damaging recent result. A paper in *Environmental Science & Technology* (5 Aug 2026, DOI 10.1021/acs.est.6c05139) is titled, bluntly, "Overlooked Soil Acidity Leads to Significant Overestimation of Carbon Sequestration in Enhanced Rock Weathering." The mechanism: if silicate dissolution is driven by nitric or sulfuric acid from fertiliser rather than by carbonic acid, cations are released — and cation-release-based mass balance counts them — but no atmospheric CO₂ is consumed. Note the corpus holds this as metadata only (closed access, no licensed abstract), so I'm reporting the title claim, not verified magnitudes. **(b) The land–ocean continuum is unverified.** Schiedung, Harrington, Dupla, Don, Hilton, Doetterl and Hemingway (Thünen Institute of Climate-Smart Agriculture + ETH Zurich), *Nature Reviews Earth & Environment*, 24 Feb 2026, DOI 10.1038/s43017-026-00761-7, is the sharpest recent synthesis. Its verification-relevant conclusion is that the fate of weathering products moving through rivers to the ocean remains poorly constrained — meaning the endpoint that ERW credits claim (dissolved bicarbonate, durable for ~10,000 years) is not the thing being measured at the field. The same paper argues feedstock supply is a binding constraint, that the most reactive (ultramafic) rocks carry the highest nickel and chromium loads, and that dedicated mining looks unavoidable at scale. The Thünen press framing was that ERW is not yet reliable for climate protection. **(c) Signal-to-noise: several 2026 field trials detected nothing.** An urban field trial published in *Biogeosciences* (3 Mar 2026, DOI 10.5194/bg-23-1681-2026) reported that basalt addition did not increase soil inorganic carbon, Tessier-extractable base cations, or dissolved inorganic carbon in leachates — the authors attribute this to coarse, albite-rich basalt in an alkaline soil, or to weathering products sitting in pools their extraction scheme couldn't see. That last clause is the MRV problem in one sentence: absence of signal is ambiguous between "no removal" and "wrong instrument." A nine-week lysimeter study (*Soil and Tillage Research*, 22 Apr 2026) likewise reported no visible basanite weathering. **(d) ERW perturbs the carbon pools used as its own reference.** A subtropical cropland trial (*Agriculture*, 30 Jan 2026, DOI 10.3390/agriculture16030338) found basalt raised soil inorganic carbon stocks 30–37% and organic carbon stocks 16–23% after one year — but *reduced* SOC stability by 17–26%, with the labile fraction rising sharply. A KIT paper (11 Aug 2026) finds ERW alters mineral-associated organic carbon formation. If ERW simultaneously adds inorganic carbon and destabilises organic carbon, single-pool accounting overstates net removal. **(e) Site-dependence defeats generic protocols.** A comparative analysis across 19 global studies (*J. Plant Nutrition and Soil Science*, 12 Aug 2026) found clay content correlated with yield response (r = 0.510, p = 0.03) and that CEC and pH drive most of the variance — the authors call for site-specific recommendations rather than generic application rates. **(f) Genuine methodological progress, though.** Quantitative X-ray powder diffraction produced direct evidence of enhanced mineral weathering in cropland (*Applied Geochemistry*, 9 May 2026) — a measurement of the mineral itself rather than a downstream proxy. And there is early work on Earth-system digital twins for verifiable gigatonne-scale ERW (30 Jul 2026), which is where the modelling-plus-sparse-measurement approach is heading. The 2024 *Frontiers in Climate* review of measurement options for terrestrial EW (DOI 10.3389/fclim.2024.1345224) remains the standard map of the toolkit. --- ## 2. Who is selling it anyway **Tracked offtakes in the corpus (6 deals, 559,867 t, 3 with disclosed value):** | Buyer | Supplier | Tonnes | Value / price | Announced | Geography | Delivery | |---|---|---|---|---|---|---| | Frontier | **Lithos Carbon** | 154,240 | $57.1M / ~$370/t | 2023-12-07 | US, 80+ farms, 9 states (basalt) | 2024–2028 | | Google | **Terradot** | 200,000 | undisclosed | ~Dec 2024 | S. Brazil (basalt) | early 2030s | | Frontier | **Terradot** | 90,000 | $27M / $300/t | 2024-12-12 | S. Brazil | 2025–2029 | | Frontier | **Eion** | 78,707 | $33M / ~$419/t | 2025-03-25 | US (olivine) | 2027–2030 | | Microsoft | **Alt Carbon** | 36,920 | undisclosed | 2026-06-11 | Darjeeling, West Bengal, India | by 2029, 3-yr deal + option | | Wild Assets | **Stack Carbon** | "tens of thousands" | undisclosed | 2026-08-03 | Uganda | multi-year | The Frontier–Lithos deal is recorded as the world's first enhanced weathering offtake. The Stack Carbon record is still pending review and its source (Argus) claims it is the largest ERW offtake from Africa to date; the separate "1 Mt by 2035" figure is Stack Carbon's company-wide ambition, not this contract. **Sellers active but below the corpus's inclusion thresholds (≥50kt or ≥$10M for non-Frontier buyers), from web checks:** - **InPlanet** (Brazil) — holds the first-mover credential: Isometric issued 235.53 independently verified ERW credits in January 2025, delivered to Adyen via ClimeFi. Microsoft signed for >28,500 tonnes over 2026–2028 (reported May 2026). - **Mati Carbon** (India, Africa; smallholder-focused basalt) — won the $50M XPRIZE Carbon Removal grand prize on 22 April 2025, with Shopify, Stripe and H&M among its buyers and J.P. Morgan blended finance announced May 2025. - **Tambora** — 14,000 t sold to Wild Assets in Q2 2025 (per CDR.fyi). - Others named in market coverage include **UNDO**, **Silicate**, **Vaulterra**, **ClimeRock** and **Zumwalt Acres**; **Boeing** bought 20,000 credits across six suppliers spanning biochar and enhanced weathering via Supercritical in April 2026, and **Mitsui O.S.K. Lines** signed a 10,000 t enhanced weathering agreement in Q2 2025. **The delivery gap is the number to hold onto.** CDR.fyi's January 2026 investment analysis found that of CDR companies that have raised private capital, ~55% have sold credits but only **24% have delivered tonnes and 12% have retired any**; enhanced weathering accounts for ~3% (~$121M) of private CDR investment across 9 companies and 12 deals. CDR.fyi's June 2026 market update states plainly that enhanced weathering, alkalinity enhancement and marine CDR have high potential but that MRV and regulatory confidence have created friction. --- ## 3. The certification layer that is meant to close the gap - **Isometric's Enhanced Weathering in Agriculture Protocol** (operational, UK registry) sits under the Isometric General Standard with a 1,000+ year durability floor. Isometric was approved by the ICAO Council on 26 Nov 2025 as a CORSIA-eligible programme — the first durable-CDR-only registry to get a compliance channel. - **Puro.earth** lists an Enhanced Rock Weathering methodology; Puro was confirmed CCP-Eligible by ICVCM on 11 Dec 2025, but that is **programme-level, not methodology-level** approval, so no ERW methodology has itself been CCP-assessed. - **EU CRCF** (Reg. (EU) 2024/3012) is pathway-neutral, with methodologies arriving as delegated acts; the Commission held its first technical workshop on ERW and OAE around October 2025. - **The important negative:** the Commission's 17 July 2026 EU ETS review proposal would centrally purchase 250 Mt of CRCF-certified permanent removals over 2031–2040 — but **initially only DACCS and BioCCS**. ERW is not in that compliance channel as drafted. Likewise the UK's interim BSI Flex quantification standards (July 2025) cover BECCS and DACCS only. The US DOE CDR Purchase Pilot Prize does include an enhanced weathering category. --- ## 4. My read, and what I can't tell you The honest characterisation is **conflicting evidence, not settled doubt**. Buyers with real diligence budgets (Frontier, Google, Microsoft, JPMorgan) have kept contracting through 2026, and the first verified credits have been issued and retired — that is not nothing. But the 2026 literature is moving in the opposite direction to the deal flow: a *Nature Reviews* perspective saying the pathway isn't ready, an *ES&T* paper alleging systematic overestimation from an overlooked acid source, and multiple field trials that couldn't detect weathering at all within a year. If the acid-source critique holds up quantitatively, it is a correction to *issued* tonnes, not just future ones. **Explicit gaps:** - Price signal is thin and biased: 3 of 6 tracked ERW deals disclose value, all Frontier, and the per-tonne figures are derived (value ÷ tonnes), not quoted. Across the whole corpus only 4.5% of contracted tonnes carry a price. - The corpus threshold (≥50kt or ≥$10M outside Frontier) systematically hides the smaller ERW sellers — InPlanet's and Mati's actual sold and delivered volumes are not in it. - Papers stop at 2026-08-14; policy events at 2026-07-17; deals at 2026-08-04. - My web verification hit a search-quota limit before I could confirm: current cumulative ERW tonnes *issued* on the Isometric and Puro registries, the status of a CRCF ERW delegated act, the current version of Isometric's ERW protocol, and the operational status of Lithos and Terradot against their 2023–2024 delivery schedules. Treat those four as open questions rather than as absences.
Judge’s note — All three gold items are substantively covered. (1) The answer explicitly states the ERW paper query is held as a complete result set as of 2026-08-14 and that `mrv-quantification` is a tracked topic tag (132 of the in-scope set) — though it reports 356 in-scope papers rather than the ~1,500+ fetched works, suggesting it surfaced a filtered subset and slightly understates corpus depth. (2) Both Frontier–Lithos Carbon (154,240 t, $57.1M) and Frontier–Eion (78,707 t, $33M) appear in the deal table, plus Terradot–Google/Frontier, Alt Carbon–Microsoft, Stack Carbon–Wild Assets. (3) The Isometric Enhanced Weathering in Agriculture Protocol is named as the de facto standard with its 1,000-year durability floor, alongside Puro.earth ERW methodology, EU CRCF delegated-act status, and CORSIA eligibility. Beyond-corpus yield is strong and specific: Nature Reviews Earth & Environment (2026-02-24) four-class uncertainty framing; PNAS (2026-06-24) 'better as farming practice than CDR'; Power et al. (Front. Clim. 2025) geochemical pitfalls; Kanzaki et al. (ERL 2025) soil cation storage; Vienne et al. (Soil 2026) weathering without inorganic CO2 removal; Buma et al. expert elicitation on loss pathways; Communications Sustainability (2026-02-16) revised 0.35–0.76 / 0.7–1.1 GtCO2/yr potentials; InPlanet's Jan-2025 first verified ERW credits + BeZero 'A' + Microsoft 2025-12-17 >28,500 t; Mati Carbon's 492.4 t first issuance under Isometric V1.2; UNDO/Jim Mann in the 2026 MRV debate. That is ~10 distinct findings, several of which are precisely the 'named MRV criticism' the probe asks for. Citation quality is excellent — journals, dates, DOIs, citation counts, and clear [DB]/[Web] provenance marking. Coverage honesty is exemplary: it flags `review_status: pending` records, warns that all disclosed prices are derived by division and all from Frontier, admits Stack Carbon has no tonnage figure, concedes the ES&T 2026-08-05 overestimation paper is metadata-only, explicitly says it could not characterise the PNAS findings, refuses to assert Silicate/Carbonaught status, and names the contract-vs-issuance gap as the number it could not close. It even self-diagnoses that its own keyword retrieval likely missed papers the corpus holds rather than claiming a coverage gap. Clarity and decision-usefulness are high: the MRV problems are organised into nine mechanistically distinct failure modes (open system, tracer loss, non-carbonic acid, indistinguishable controls, model divergence, SOC contamination, energy penalty, trace-metal caps, commercial mass-balance methods), and the closing synthesis — that the biases all point toward overestimation while nearly every deal is coded durable-1000y by convention — is exactly the kind of judgement an underwriter needs. Minor deductions only for the paper-count discrepancy and the number of quantitative claims a reader would have to spot-check.
I'll start by checking what the corpus covers, then pull the ERW evidence profile. Now let me pull the MRV-specific literature and the wider deal/supplier picture. Now the critical step — checking what the corpus says it cannot cover: literature after 2026-08-14, deals after 2026-08-04, and ERW suppliers outside the Frontier/hyperscaler deal threshold. ## Summary Two things are true at once: the ERW literature has spent five years documenting why its carbon accounting is hard to close, and roughly half a million tonnes have already been contracted against it — with the first verified credits issued in January 2025. Below, corpus records are marked **[DB]** and things I found by searching the open web are marked **[Web]**, because the corpus has hard freshness boundaries (papers fetched 2026-08-14, deals through 2026-08-04, policy events through 2026-07-17) and an explicit deal-inclusion threshold that excludes most ERW suppliers. --- ## 1. How large the evidence base is **[DB]** The `erw-ag` evidence profile holds **356 in-scope papers**: 88 lab, 65 modeling, **60 field trials**, 57 reviews, 12 LCAs, 7 surveys, 67 unclassified. Of these, **132 are tagged `mrv-quantification`** — the single largest topic cluster for the pathway. That is a mature-ish literature by CDR standards, but note the composition: lab and modeling still outnumber field trials, and many field trials are one-to-four years long against a process the industry credits as 1,000-year durable. ## 2. What the literature actually says the MRV problems are ### (a) It is an open system with no meter, and the measurement approaches disagree The core reference is Knapp et al.'s review, *A review of measurement for quantification of carbon dioxide removal by enhanced weathering in soil* (Frontiers in Climate, 2024-06-19, doi:10.3389/fclim.2024.1345224, 53 citations) **[DB]**. It splits methods into solid-phase (including exchangeable pools), aqueous-phase and gas-phase, notes gas-phase flux measurement is dominated by short-term organic carbon cycling, and concludes that because final storage happens in the ocean while measurement happens at the field, quantification requires deliberately redundant "surplus measurement strategies." That is a polite way of saying no single method closes the budget. ### (b) The tracer disappears before you can measure it Amann et al., *Enhanced Weathering and related element fluxes – a cropland mesocosm approach* (Biogeosciences, 2020-01-08) **[DB]**: dunite on Belgian agricultural soil produced calculated weathering rates **up to three orders of magnitude lower** than other studies, with CO₂ consumption of just 2.3–4.9 t CO₂ km⁻² a⁻¹. The authors attribute the gap to Mg being adsorbed, mineralised, taken up by plants, concentrated by evapotranspiration or moved by preferential flow — Mg in upper soil layers ran ~10× higher than in outlet water. Their conclusion is the MRV thesis in one line: careful tracking of weathering indicators in the field is essential, or global potential estimates are unanchored. ### (c) Non-carbonic acids inflate the number If weathering is driven by nitric or sulphuric acid rather than carbonic acid, cation release happens without atmospheric CO₂ removal. Dietzen & Rosing (*Quantification of CO₂ uptake by enhanced weathering of silicate minerals applied to acidic soils*, IJGGC, 2023-03-16) **[DB]** set a working threshold: soils below **pH 6.3 need correction**, and soils below **pH 5.2 may not be suitable candidates at all**. Their measured result — **728 kg CO₂/ha from 50 t/ha of Greenlandic glacial rock flour over three years** — is roughly an order of magnitude below headline modelling estimates. This thread is still live: the corpus holds *Overlooked Soil Acidity Leads to Significant Overestimation of Carbon Sequestration in Enhanced Rock Weathering* (Environmental Science & Technology, **2026-08-05**, doi:10.1021/acs.est.6c05139) **[DB]** — but it is closed-access, so the corpus holds metadata only, no abstract, and no study type. I could not verify the magnitude of the claimed overestimate. ### (d) The control plot sometimes matches the treated plot The most sobering field result in the set is Larkin et al., *Quantification of CO₂ removal in a large-scale enhanced weathering field trial on an oil palm plantation in Sabah, Malaysia* (Frontiers in Climate, 2022-08-30) **[DB]**. Across three amended and three paired reference catchments over ~3 years, CO₂ drawdown via alkalinity export was **3.8 ± 0.8 t CO₂/ha in amended catchments and 3.7 ± 0.6 in the references** — statistically indistinguishable. Only one of three pairs showed a difference (~1 t CO₂/ha), and the high background rates were driven mainly by weathering of **carbonate fertilisers**, not the applied silicate. Signal-to-noise, not chemistry, is the binding constraint. Recent field trials reinforce it **[DB]**: an urban field trial (*Biogeosciences*, 2026-03-03) found basalt addition produced **no increase in soil inorganic carbon, no increase in Tessier-extractable base cations and no DIC increase in leachates** in an alkaline soil; a 2026 lysimeter study (*Soil and Tillage Research*, 2026-04-22) reported no visible basanite weathering after nine weeks. ### (e) Models and measurements diverge in both directions Vienne et al. (*Frontiers in Climate*, 2022-05-17) **[DB]** found PHREEQC **underestimated pore-water Mg by a factor of three** while **overestimating calcite precipitation** in a potato mesocosm. Lewis et al. (*Applied Geochemistry*, 2021-06-21) **[DB]** showed CDR potential across six commercially-used basalts spans **1.3–8.5 t CO₂/ha after 15 years** purely from mineralogy and specific surface area — and that estimates built from rock *chemistry* (cheap, widely available) must be converted to mineralogy-based estimates to be realistic. Feedstock characterisation is therefore a first-order MRV variable, not a procurement detail. ### (f) Organic carbon contaminates the inorganic signal — in both directions Rock dust changes soil organic carbon, which is exactly the pool ERW accounting must exclude. *Global Change Biology* (2023-11-22) **[DB]** found rock dust raised mineral-associated organic matter by **22%**, partly counteracted by plants. A 2026 subtropical cropland trial (*Agriculture*, 2026-01-30) **[DB]** found basalt raised SOC stocks by 16–23% and SIC stocks by 30–37%, but **reduced the SOC stability index by 17–26%** — carbon gained in a labile form. A German repository item (KITopen, 2026-08-11) is tagged `mrv-quantification` on the same question **[DB]**. ### (g) The energy penalty can exceed the measured capture Taylor et al.'s 15-year wollastonite watershed study in New Hampshire (*Biogeosciences*, 2021-01-12) **[DB]** measured cumulative capture by carbonic acid weathering of **0.025–0.13 t CO₂/ha** against a **logistical penalty of 0.8–2.4 t CO₂/ha** from mining, grinding, transport and spreading. Net CDR was only positive (8.5–11.5 t CO₂/ha) once increased forest productivity was counted — a different carbon pool with different durability. Separately, Life-cycle work (*Communications Earth & Environment*, 2022-05-05) **[DB]** finds **grinding, not mining, dominates** supply-chain impacts. ### (h) Side effects cap the application rate that MRV assumes Dupla et al. (*European Journal of Soil Science*, 2023) **[DB]** find that at the commonly modelled **40 t/ha/yr**, regulatory limits for **copper and nickel would be exceeded after 6 and 10 years** respectively. Olivine specifically is discouraged on Ni/Cr grounds by the mesocosm literature above. ### (i) The methods that commercial MRV actually runs on Two corpus papers underpin current protocols: *Initial Validation of a Soil-Based Mass-Balance Approach for Empirical Monitoring of Enhanced Rock Weathering Rates* (ES&T, 2023-11-14, 66 citations) **[DB]** — the immobile-tracer/mass-balance approach used commercially — and, newly, *Direct evidence for enhanced mineral weathering in cropland verified by quantitative X-ray powder diffraction* (Applied Geochemistry, 2026-05-09) **[DB]**, an independent solid-phase verification route. On the downstream side, *River chemistry constraints on the carbon capture potential of surficial enhanced rock weathering* (Limnology & Oceanography, 2022-10-11) **[DB]** estimates riverine carbon transport potential of 7.1–21.3 GtCO₂/yr globally, concluding transport is **unlikely to be the primary bottleneck** — one of the few reassuring findings, though it flags loss processes along the land–ocean continuum as unresolved. ### (j) What has landed since — and what the corpus retrieval missed **[Web]** - **Nature Reviews Earth & Environment, 2026-02-24**, *Uncertainties of enhanced rock weathering for climate-change mitigation* (doi:10.1038/s43017-026-00761-7) — a Perspective structured around four uncertainty classes: feedstock availability, plant–soil system impacts, CDR efficiency along the land–ocean continuum, and socio-economics. - **PNAS, 2026-06-24**, *Enhanced rock weathering has greater promise as a sustainable farming practice than a CO₂ removal technology* (doi:10.1073/pnas.2602421123). The title carries the thesis. I could only retrieve its metadata and reference list, not its findings, so treat the framing as indicative rather than characterised. Its citations point to a cluster of 2025–26 critiques I have **not** independently verified: Power et al., *Are enhanced rock weathering rates overestimated? A few geochemical and mineralogical pitfalls* (Front. Clim. 6, 2025); Kanzaki et al., *Soil cation storage is a key control on the carbon removal dynamics of enhanced weathering* (ERL 20, 2025); Vienne et al., *Weathering without realizing inorganic CO₂ removal revealed through base cation monitoring* (Soil 12, 2026); and Buma et al., an expert elicitation on agricultural enhanced weathering reporting CDR potential and **loss-pathway uncertainties**. - **Communications Sustainability, 2026-02-16** (covered by Cornell): global ERW potential of **0.35–0.76 GtCO₂/yr by 2050 and 0.7–1.1 GtCO₂/yr by 2100** — an order of magnitude below the 2–4 Gt/yr in the 2018 Minx et al. review the corpus lists as ERW's most-cited paper, and far below Strefler et al.'s 2018 theoretical 4.9 Gt/yr for basalt. **Honesty note:** the corpus states the ERW paper query is a *complete* result set as of 2026-08-14, so the Feb and June 2026 items above are probably held; my keyword searches simply didn't surface them (a retrieval limitation on my side, not a proven coverage gap). I could not confirm either way. --- ## 3. Who is selling ERW removals anyway ### Corpus records **[DB]** — `erw-ag` aggregate: **6 deals, 559,867 tonnes contracted, 3 with disclosed value** | Announced | Supplier | Buyer | Tonnes | Value | Derived $/t | Where | Delivery | |---|---|---|---|---|---|---|---| | 2023-12-07 | **Lithos Carbon** | Frontier | 154,240 | $57.1M | **$370** | US (80+ farms, 9 states) | 2024–2028 | | 2024-12-01 | **Terradot** | Google | 200,000 | undisclosed | — | Brazil | by early 2030s | | 2024-12-12 | **Terradot** | Frontier | 90,000 | $27M | **$300** | S. Brazil | 2025–2029 | | 2025-03-25 | **Eion** | Frontier | 78,707 | $33M | **$419** | US, olivine | 2027–2030 | | 2026-06-11 | **Alt Carbon** | Microsoft | 36,920 | undisclosed | — | Darjeeling, W. Bengal, India | by 2029 | | 2026-08-03 | **Stack Carbon** | Wild Assets | "tens of thousands" | undisclosed | — | Uganda | multi-year | Caveats that matter: - **Prices are derived by division, not quoted**, and all three disclosing deals are Frontier. Across the whole corpus only 21 of 59 deals disclose value, covering 4.5% of contracted tonnes. The $300–$419/t band describes Frontier's purchasing, not the ERW market. - The Alt Carbon and Stack Carbon records are **`review_status: pending`** — extracted but not human-reviewed. The Stack Carbon record has no tonnage figure at all (the source said only "tens of thousands"), and the extraction note explicitly warns that Stack Carbon's "1mn credits by 2035" figure is company ambition across Uganda *and* Madagascar, not this contract. Microsoft also holds an **option on additional Alt Carbon volumes contingent on delivery and verification milestones** — a structure worth noting, since it prices the MRV risk described above. - A durability inconsistency sits in the data: Stack Carbon is coded `conventional` while Alt Carbon is coded `durable-1000y` "per ERW convention." Both are pending review. This is the accounting question — how long bicarbonate stays out of the atmosphere — showing up as a data-entry problem. ### Sellers the corpus structurally cannot hold **[Web]** The corpus only ingests Frontier deals plus Microsoft/Google/JPMorgan/Meta/Amazon deals at **≥50 kt or ≥$10M**. Most of the ERW supply side sits below that line: - **InPlanet** (Brazil) — delivered what it and Isometric announced on **2025-01-06** as the **world's first verified ERW carbon removal credits**, under Isometric's Enhanced Weathering protocol, with a post-issuance **'A' rating from BeZero Carbon**. It signed a deal with **Microsoft on 2025-12-17 for >28,500 t between 2026 and 2028** — a Microsoft ERW deal that falls below the corpus's tonnage threshold and so does not appear in the table above. - **Mati Carbon** (India/Africa smallholder ERW) — no entity record in the corpus at all. Carbon Herald reported (June 2026) that it delivered **492.4 tonnes** as the **first company to issue under Isometric's V1.2 ERW protocol**. Note the scale gap: 492 tonnes delivered against six-figure contracts. - **UNDO** (UK) — founder Jim Mann appears in May 2026 trade coverage of the ERW MRV debate; the corpus holds no record of the company. - Others active in the market (Silicate, Carbonaught, and similar) I could **not** verify in this session — my web-search budget was exhausted before I could confirm them, so I am not asserting their status. **The delivery-versus-contract gap is the number I could not close.** Neither the corpus (which tracks contracts, not issuances) nor my searches produced an authoritative total of ERW tonnes *issued and retired*. CDR.fyi is the scoreboard for that, and the corpus itself names it as the fuller transaction record. --- ## 4. The certification and policy standing that MRV feeds into **[DB]** - **Isometric Enhanced Weathering in Agriculture Protocol** (UK registry, operational) — the de facto ERW standard, sitting under the Isometric General Standard with its 1,000-year durability floor. The corpus records it as "v1.x, confirm current version"; **[Web]** it was at **V1.2** by June 2026. - **Puro.earth** — General Rules **v4.4 approved 2026-05-07**; Enhanced Rock Weathering is among its active methodologies. Puro was confirmed **CCP-Eligible by ICVCM on 2025-12-11 at program level** (individual methodologies not yet CCP-assessed). - **EU CRCF** (Regulation 2024/3012, in force 2024-12-26) is a pathway-neutral umbrella; ERW methodologies arrive as delegated acts. The Commission held its first technical workshop on ERW and OAE under the CRCF (covered in a 2025-10-01 CDR Policy Scoop episode). - **ERW is not in the EU ETS proposal.** The Commission's 2026-07-17 ETS review would centrally purchase **250 Mt of CRCF-certified permanent removals over 2031–2040 — initially DACCS and BioCCS only**, with biochar explicitly excluded on evidentiary grounds (limited long-term field studies, degradation-rate uncertainty). ERW faces the same evidentiary bar and is not named. **The MRV literature above is precisely why.** - **UK**: interim BSI Flex quantification methodologies cover **BECCS and DACCS only**; UK ETS GGR integration (government response 2025-07-21) requires **200-year minimum storage**, UK-based removals initially, legislating by end-2028 and operational by end-2029. - **US**: the **DOE CDR Purchase Pilot Prize** does include enhanced weathering as a category (24 Phase 2 semifinalists, up to 10 winners at up to $3M each, preserved with a $45M FY26 appropriation). - **CORSIA**: Isometric's approval as an eligible programme (announced 2025-11-26) is the first durable-CDR-only registry admitted — potentially a compliance channel for ERW credits, though actual availability of eligible units with corresponding adjustments remains unquantified. **Relevant listening [DB]:** *Is Enhanced Rock Weathering Ready for Scale?* with Dirk Paessler and Mel Murphy (CDR Policy Scoop, 2026-08-09) addresses exactly this question — whether ERW can be measured reliably enough to scale. --- ## 5. Where the two sources disagree, and what neither covers **Disagreement:** the corpus's most-cited papers carry 2018-era potential estimates of **2–4 GtCO₂/yr** (Minx et al.) and up to **4.9 Gt/yr for basalt** (Strefler et al.); the 2026 literature I found by searching lands at **0.7–1.1 GtCO₂/yr by 2100**. This is not a data error in either source — it is the field revising downward as field measurement replaces modelling, and it is the single most important trend for anyone underwriting ERW offtakes. **Neither source covers:** - **Registry issuance and retirement totals by supplier** — the gap between ~560 kt contracted (corpus) and the hundreds-of-tonnes deliveries visible in trade press is unmeasured here. - **Prices for any non-Frontier ERW deal.** Google's 200 kt, Microsoft's two purchases and both 2026 deals are all undisclosed. - **Whether the 2025–26 overestimation findings have forced protocol changes.** The corpus does not hold Isometric version history; I could not verify what changed between V1.0 and V1.2. - **Company-held field data.** Most commercial ERW deployment generates unpublished measurement; the peer-reviewed field-trial set (60 papers) is largely academic plots, not the commercial farms credits are issued against. - Non-English literature, preprints, and anything published after 2026-08-14 or announced after 2026-08-04. **One structural point worth stating plainly:** the pattern across (c), (d) and (e) — non-carbonic acid weathering, indistinguishable control plots, and model–measurement divergence — all push in the same direction, toward **overestimation** rather than random error. Meanwhile every corpus deal except one is coded `durable-1000y` by convention rather than by source statement. The commercial claims and the measurement literature are not yet reconciled, and the EU's decision to admit only DACCS and BioCCS into the ETS is the first hard policy consequence of that.
Judge’s note — Strong, expert-grade answer. All three checklist items are addressed: it characterizes the ERW literature holdings and treats quantification/MRV as a tracked theme (though it reports 356 in-scope papers and explicitly calls the set a 'curated floor, not a census,' which conflicts with the gold's ~1,500+ complete-result-set framing — a notable provenance discrepancy); it names Frontier–Lithos (154,240 t, $57.1M, first ERW offtake) and Frontier–Eion (78,707 t, $33M) plus Terradot–Google/Frontier, Alt Carbon–Microsoft, Stack Carbon–Wild Assets; and it lays out the registry/protocol layer (Isometric 'Enhanced Weathering in Agriculture' under the General Standard with 1,000-yr floor, CORSIA eligibility, Puro.earth methodology and ICVCM program-level CCP status, CRCF workshop, EU ETS Article 9c exclusion of ERW, DOE Purchase Pilot Prize). On the beyond-corpus probe: no post-2026-08-14 deals or results are surfaced (the answer explicitly declares that outside its freshness window, which is honest but yields nothing fresh). However it delivers a dense set of specifically named MRV criticisms with journals, dates, DOIs and direct quotes — open-system multi-phase measurement needing 'surplus measurement strategies'; the carbon- vs cation-accounting schism and unsuitability of titration-based TA in agricultural porewaters; accessory-carbonate misattribution and fast initial dissolution inflating rates; pH 6.3/5.2 non-carbonic-acid correction thresholds; nitrification-driven 67% overestimate; the Aug 2026 soil-acidity overestimation paper; lab–field discrepancy synthesis; conflicting SOC evidence; biased/poorly characterized feedstocks and 1.3–8.5 tCO2/ha mineralogy spread; grinding-dominated LCA; Cu/Ni regulatory-limit exceedance. The field-trial comparison table (Swiss vineyards 100±30 kg/ha/yr vs US Midwest 3.7–8.6 t/ha/yr vs indistinguishable Sabah catchments vs NH wollastonite net-negative logistics) is exactly the decision-relevant evidence an expert wants. Citation quality is excellent (venue, date, DOI, quoted language, flagged closed-access and uncorroborated claims). Coverage honesty is exemplary: it flags pending-review records, that all disclosed prices are Frontier's, names absent sellers (Mati, UNDO, InPlanet, Silicate) as a coverage gap rather than absence of activity, points to CDR.fyi and Carbon Gap, and closes with a crisp 'what I can't tell you' including the unanswered question of whether registries have revised ERW crediting downward. Clarity docked one point for the residual tool-process narration at the top ('The MRV paper search errored...'), which adds noise for an expert reader. Main risk not verifiable here is whether the numerous precise figures/DOIs are all accurate; the corpus-scale mismatch with the gold is a mild warning sign.
I'll start by checking what the corpus covers, then dig into ERW specifically. Now let me pull the MRV-specific literature and the commercial/policy layers. The MRV paper search errored on query complexity; let me retry with simpler queries and pull the supplier entities. Those returned only deal aggregates. Let me query the paper layer explicitly. ## Bottom line The ERW literature in this corpus is unusually candid about its own measurement problem: the field has 356 in‑scope papers (60 field trials, 88 lab, 65 modeling, 57 reviews, 12 LCAs), and a large share of the recent field and review work is specifically about how *not* to over‑credit ERW. Meanwhile, at least six public offtakes covering ~560,000 tonnes have been signed since December 2023, with disclosed prices of $300–419/tonne. The commercial layer is running well ahead of the measurement consensus — and several 2025–26 papers say so explicitly. *(Source: Carbon Curve corpus, ERW evidence profile; papers fetched 2026‑08‑14, deals through 2026‑08‑04. This is a curated floor, not a census.)* --- ## What the research says about ERW's MRV challenges **1. It's an open system, and that is the root problem.** The reference review is Clarkson-era work summarised in *"A review of measurement for quantification of carbon dioxide removal by enhanced weathering in soil"* (Frontiers in Climate, 19 Jun 2024, 53 citations, [DOI](https://doi.org/10.3389/fclim.2024.1345224)). Its framing: "as an open-system pathway, the measurement (monitoring), reporting, and verification (MRV) of carbon removal and storage is challenging," with three competing measurement families — solid-phase (incl. exchangeable), aqueous-phase, and gas-phase — each with distinct failure modes. Gas-phase flux measurement, it notes, "can be dominated by short-term organic carbon cycling." Its conclusion is that MRV requires "surplus measurement strategies," i.e. redundancy, not a single proxy. **2. There is an unresolved methodological schism: carbon-based vs cation-based accounting.** *"Reviews and syntheses: Carbon vs. cation based MRV of Enhanced Rock Weathering and the issue of soil organic carbon"* (Biogeosciences, 5 Jan 2026, [DOI](https://doi.org/10.5194/bg-23-53-2026)) argues carbon accounting "is ill-suited to close the balance of all carbon fluxes" in arable land, and — notably — that titration-based total alkalinity, the standard method borrowed from oceanography, "is unsuitable to record TA for porewaters in agricultural soils." It proposes cation accounting instead. That a foundational analytical choice is still being contested in 2026 is itself the headline. **3. Multiple independent papers say published CDR rates are systematically overestimated.** - *"Are enhanced rock weathering rates overestimated? A few geochemical and mineralogical pitfalls"* (Frontiers in Climate, 13 Jan 2025, 24 citations, [DOI](https://doi.org/10.3389/fclim.2024.1510747)) names three: initially fast dissolution of reactive phases inflating long-run rates; **accessory carbonate dissolution being misread as silicate weathering**; and cation-based methods misattributing cations that were never balanced by dissolved inorganic carbon (e.g. strong-acid weathering). Its blunt framing: "Overestimating CDR risks undermining ERW if meaningless carbon credits are counted." - *"Quantification of CO2 uptake by enhanced weathering of silicate minerals applied to acidic soils"* (Int. J. Greenhouse Gas Control, 16 Mar 2023, [DOI](https://doi.org/10.1016/j.ijggc.2023.103872)) quantifies the non-carbonic-acid correction: soils below **pH 6.3** need correction, and soils below **pH 5.2 may not be suitable at all** for ERW crediting. - *"Overlooked Soil Acidity Leads to Significant Overestimation of Carbon Sequestration in Enhanced Rock Weathering"* (Environmental Science & Technology, 5 Aug 2026, [DOI](https://doi.org/10.1021/acs.est.6c05139)) — the newest paper in the corpus on this theme; it is closed-access, so I have title-level evidence only and cannot report its magnitude. - *"Silicate weathering and CO2 consumption within agricultural landscapes, the Ohio-Tennessee River Basin"* (Biogeosciences, 2012) had already found CO₂ consumption estimates assuming carbonic-acid-only weathering were **67% overestimated** in a tilled corn watershed, because nitrification from N fertiliser was driving the weathering. **4. Field results disagree with each other by an order of magnitude or more.** This is the clearest empirical signal of MRV immaturity: | Trial | Reported result | |---|---| | Swiss vineyards, 20 t/ha basalt, 1,000 days (ES&T, 24 Nov 2025) | **100 ± 30 kg CO₂/ha/yr** — "10 to 30 times lower than the upper rates reported in some previous modeling and experimental studies"; most standard proxies (pH, EC, DIC, Ca, Mg) *did not* rise significantly; only Na did | | US Midwest maize/soy and miscanthus, 50 t/ha/yr (Global Change Biology, 17 Aug 2023) | **3.7 t CO₂/ha/yr** (maize/soy) and **8.6 t CO₂/ha/yr** (miscanthus) | | Sabah oil palm catchments, 3 yr (Frontiers in Climate, 30 Aug 2022) | Amended and reference catchments statistically indistinguishable (3.8 ± 0.8 vs 3.7 ± 0.6 tCO₂/ha); only one of three pairs showed ~1 tCO₂/ha excess; high rates were "primarily driven by weathering of carbonate fertilizers" | | New Hampshire forested watershed, wollastonite, 15 yr (Biogeosciences, 12 Jan 2021) | Carbonic-acid weathering capture of only **0.025–0.13 tCO₂/ha** against a **0.8–2.4 tCO₂/ha logistical carbon penalty** | | Kimberlite residues, Ontario, 3 yr (Frontiers in Climate, 1 Jul 2025) | Up to 1.4 tCO₂/ha, but partitioned as **~75% carbonate / ~25% silicate** weathering; "CO₂ fluxes and soil inorganic carbon proved ineffective for CDR quantification" | | Belgian soil mesocosm, olivine (Biogeosciences, 8 Jan 2020) | Calculated weathering rates "up to 3 orders of magnitude lower than found in other studies" | *"Understanding the Lab-Field Discrepancy in Mineral Dissolution From Flasks to Enhanced Rock Weathering"* (Reviews of Geophysics, 28 Nov 2025, [DOI](https://doi.org/10.1029/2025rg000881)) is the dedicated synthesis of that gap. **5. Soil organic carbon confounds the signal, and the evidence conflicts.** - Reduced MAOM accrual after 2 years of crushed rock in three California croplands (Biogeochemistry, 23 Jul 2024) — though no net SOC loss. - Rock dust *increased* mineral-associated organic matter by 22% in a pot trial, but plants "partially counteracted" it via cation loss (GCB, 22 Nov 2023). - ERW "increased soil carbon but reduced SOC stability" in subtropical croplands (Agriculture, 30 Jan 2026). - A global meta-analysis concluded ERW *promotes* SOC accumulation (GCB, 1 Sep 2025). This is genuine conflicting evidence, not noise: if ERW moves the SOC pool in either direction, carbon-based MRV double-counts or under-counts. **6. Feedstock characterisation is weak.** *"Biased selection and incomplete characterization of feedstock materials in enhanced rock weathering experiments"* (Applied Geochemistry, 15 Nov 2025): research is "dominated by a narrow range of commercially available rock powders, arguably often selected more for convenience than for relevance," and these are "often poorly characterized, neglecting core methods and principles of petrography." Relatedly, RTM work on six mined basalts used in large field trials found CDR potentials spanning **1.3 to 8.5 tCO₂/ha over 15 years** for the same 50 t/ha application, purely from mineralogy and particle-size differences (Applied Geochemistry, 21 Jun 2021). **7. Supply-chain and lifecycle emissions are a first-order term, not a footnote.** LCA across twelve nations found **grinding, not mining, dominates** environmental impacts, and results swing on the national energy mix (Communications Earth & Environment, 5 May 2022). *"Life Cycle Assessment in the MRV of Land-Based CDR: Gaps and Opportunities"* (ES&T, 27 Mar 2025) compares registry protocols against LCA practice and flags baselines, additionality, uncertainty, multifunctionality and "overlooked carbon pools" as unresolved. **8. Co-pollutant risk is an MRV-adjacent constraint.** Modeling in the European Journal of Soil Science (2023) found that at the frequently modeled 40 t/ha/yr rate, regulatory limits would be exceeded after **6 years for copper and 10 years for nickel**. Olivine specifically is flagged: the Belgian mesocosm found Ni concentrations exceeding drinking-water limits and recommended against olivine feedstock. **9. What's improving.** Direct-detection methods are maturing: quantitative X-ray powder diffraction giving "direct evidence for enhanced mineral weathering in cropland" (Applied Geochemistry, 9 May 2026); soil-based mass-balance monitoring validated using REE/base-cation ratios (ES&T, 14 Nov 2023); and proposals for Earth-system digital twins for verifiable gigaton-scale ERW (Jul 2026, closed-access, low-visibility venue — treat cautiously). --- ## Who is selling ERW removals anyway Six ERW offtakes in the corpus, **559,867 tonnes contracted**, three with disclosed values totalling **$117.1M across 322,947 tonnes** (volume-weighted ~$363/t): | Supplier | Buyer | Announced | Tonnes | Value / price | Geography & delivery | |---|---|---|---|---|---| | **Lithos Carbon** (US) | Frontier | 2023-12-07 | 154,240 | $57.1M → **$370/t** (derived) | Superfine basalt, 80+ US farms in 9 states; 2024–2028. Recorded as the world's first enhanced weathering offtake | | **Terradot** (US/Brazil) | Google | ~2024-12 | 200,000 | undisclosed | Basalt on Brazilian farmland; delivery by early 2030s. Google's largest single CDR purchase at the time | | **Terradot** | Frontier | 2024-12-12 | 90,000 | $27M → **$300/t** (derived) | Southern Brazil; 2025–2029 | | **Eion** (US) | Frontier | 2025-03-25 | 78,707 | $33M → **$419/t** (derived) | Olivine into southern/midwestern US farm soils; 2027–2030 | | **Alt Carbon** (India) | Microsoft | 2026-06-11 | 36,920 | undisclosed | Darjeeling Revival Project, West Bengal; 3-year agreement, delivery by 2029; Microsoft holds an option on additional volumes contingent on delivery *and verification* milestones | | **Stack Carbon** (Uganda) | Wild Assets | 2026-08-03 | "tens of thousands" (no figure) | undisclosed | Basalt on Ugandan farmland; multi-year. Argus reports it as the largest ERW offtake from Africa to date | Two caveats on these numbers. First, **all three disclosed prices are Frontier deals** — corpus-wide, only 21 of 59 deals carry a value, covering 4.5% of contracted tonnes, and the largest-volume buyers publish nothing. So "$300–419/t" describes Frontier's disclosure practice, not the ERW market. Second, the Alt Carbon and Stack Carbon records are **pending human review**, and their extraction notes are themselves revealing about commercial looseness: the Microsoft/Alt Carbon source "says only that CO2 becomes 'stable bicarbonates' and never names an endpoint"; the Stack Carbon source "calls the credits 'durable' but states no duration." Durability of 1,000 years was assigned by corpus convention, not by the seller. **Named ERW sellers not in this corpus.** Mati Carbon, UNDO, InPlanet and Silicate returned no records. That is a coverage gap (the deal layer only captures Frontier plus Microsoft/Google/JPMorgan/Meta/Amazon above a ≥50kt or ≥$10M threshold), **not** evidence they aren't selling. CDR.fyi is the fuller transaction scoreboard. --- ## The institutional layer that lets this happen - **Isometric's "Enhanced Weathering in Agriculture" protocol** (GB, registry authority) is operational, sitting under the Isometric General Standard with a 1,000-year durability floor. Version is recorded as v1.x, unconfirmed at review. - **Puro.earth** lists Enhanced Rock Weathering among its active methodologies (General Rules v4.4 approved 2026-05-07). Puro became ICVCM CCP-Eligible at *program* level on 2025-12-11 — individual methodologies have not been CCP-assessed. - **Isometric became CORSIA-eligible** on 2025-11-26, the first durable-CDR-only registry approved, opening a compliance channel in principle. - **But ERW is outside the biggest new compliance demand signal.** The Commission's 17 July 2026 EU ETS review proposal (Article 9c, 250 Mt of central purchases 2031–2040) covers **only DACCS and BioCCS** with permanent geological storage. ERW is not in it. The **US DOE CDR Purchase Pilot Prize** does include an enhanced weathering category (up to $3M per winner, up to 10 winners; preserved with a $45M FY26 appropriation). - Frontier's own **Verification Confidence Levels** framework (with CarbonPlan, since 2022) explicitly prices uncertainty: `Delivered tons = Net tons × (1 − uncertainty discount)`, with a stated philosophy that "it's better to remove 10,000 tons with 60% confidence than to remove 1,000 tons with 90% confidence." That is the honest answer to "why buy anyway." *(Recorded as an uncorroborated source claim from Frontier's own writing.)* A separate Frontier signal claims 2025 saw "the first verified tons from complex open systems, with enhanced weathering using the Cascade Foundations framework," and asserts that "enhanced weathering kinetics are commercially viable within the first few years" — also uncorroborated, and in direct tension with the Swiss vineyard and Sabah results above. Two podcast episodes bear directly on this: *"Is Enhanced Rock Weathering Ready for Scale?"* with Dirk Paessler and Mel Murphy (CDR Policy Scoop, 9 Aug 2026) and *"Are ERW and OAE ready for Carbon Removal Certification?"* with Rachel Smith (1 Oct 2025), on the European Commission's first CRCF technical workshop for ERW and OAE. --- ## What I can't tell you - **Delivery and issuance.** The corpus tracks contracts, not tonnes verified and retired against these ERW deals. Whether Lithos, Terradot or Eion have issued credits at contracted volumes is not held here. - **The Aug 2026 ES&T overestimation paper's magnitude** — closed access, metadata only. Same for several field trials without rights-cleared abstracts. - **Post-2026-08-14 literature and post-2026-08-04 deals** are outside the freshness window entirely. - **CRCF's ERW methodology status.** A technical workshop happened (Oct 2025) but no adopted ERW delegated act appears in the corpus. Absence here is not proof of absence — check Carbon Gap's European carbon removal policy tracker. - **Whether any registry has revised ERW crediting downward** in response to the 2025–26 overestimation literature. That's the question I'd most want answered, and this corpus doesn't answer it.
Asked, verbatim: “I'm diligencing Vaulted Deep. What do we know about their deals, and how does their pathway's evidence base look?”
It found every checklist fact on this question — not every question favors the database, and this page reports those too.
Judge’s note — All three gold items are hit substantively: the Frontier offtake with exact tonnage/value (and derived ~$382/t), the Microsoft 4.9 Mt 2025 deal (framed as roughly second-largest CDR deal rather than explicitly 'largest in pathway', but its scale-dominance is made clear via the >95% concentration point), and a clear pathway framing (wet organic waste slurried and injected into deep disposal wells, geologic storage, durability claims). Well beyond the corpus: Advantek spinout origin, $32M Series A (Nov 2024), $8M XPRIZE runner-up (Apr 2025, with Mati Carbon grand prize), direct Google 50,000 t deal (Sept 2025) plus the Google/Isometric avoided-methane methodology workstream, Isometric Biomass Geological Storage protocol issuance history (1,666 t Apr 2024; >18,000 by Jul 2025; 25,215 cumulative), Hutchinson KS single-site concentration, ICVCM CCP approval of the protocol, 350Solutions verification and ~2% buffer pool, and a 2025 Chemical Reviews BiCRS durability reference — roughly ten distinct substantive findings, all checkable and mostly accurate. Citations are named and date-stamped, though a few are loosely attributed ('ESG News, Sept 2025', 'Carbon Credits, Jul 2025', 'Later 2025 ICVCM announcement') and one aggregator source (American Bazaar) is weak. Coverage honesty is excellent: an explicit note that searches were truncated, a dedicated gaps section, and 'not evidence of absence' framing on the missing critical commentary. Analytically strong for an expert reader: the ~25 kt issued vs. 4.9 Mt contracted arithmetic, the counterfactual/decay-fraction sensitivity for biosolids land application, downhole methanogenesis as distinct from avoided surface methane, the 1,000 vs 10,000-year durability inconsistency, and the additionality tension with PFAS-driven tipping-fee economics are exactly the right diligence pressure points. Minor deductions for a couple of unverified table cells left as placeholders and slight over-hedging in places, but overall this is close to a model diligence answer.
I'll research Vaulted Deep's deal history and the scientific evidence base for their pathway. ## Vaulted Deep — diligence snapshot **What they are:** A Houston-based BiCRS (biomass carbon removal and storage) company that takes wet, low-value organic waste — biosolids, paper sludge, manure, food and ag residues — turns it into a slurry, and injects it into deep disposal wells. It spun out of **Advantek Waste Management Services**, a ~10-year-old industrial waste injection business, and uses Advantek's core slurry-injection technology (Frontier Climate, May 2024). Operations began September 2023. Note on dates: figures below are stamped with the date I could verify them. I was unable to complete several follow-up searches (tool limit), so the "gaps" section is materially incomplete rather than exhaustive. --- ## 1. The deal book | Counterparty | Announced | Volume | Value | Term | |---|---|---|---|---| | Frontier buyers (pre-purchase) | Sept 2023 | Not verified | Not verified | — | | Frontier buyers (offtake) | 1 May 2024 | 152,480 t | **$58.3M** (~$382/t) | 2024–2027, with options on future project tonnes at **lower** prices | | Google (via Frontier) | Mar 2024 | Included in above | — | — | | **Microsoft** | 17 Jul 2025 | **up to 4.9 Mt** | Undisclosed | 12 years, through **2038** | | **Google** (direct follow-on) | 17 Sep 2025 | **50,000 t** | Undisclosed | by **2030** | Other capital: **$32M Series A** (ESG Today, Nov 2024) and **$8M as second runner-up in the XPRIZE Carbon Removal competition**, awarded 23 April 2025 (PRNewswire; Mati Carbon took the $50M grand prize). **Reading the deal book:** - The Frontier offtake is the only one with a disclosed price. **~$382/t** is a real number to anchor on, but Frontier explicitly structured options on future tonnes at lower prices — so $382 is a ceiling, not a run-rate. Microsoft and Google prices are undisclosed and, at Microsoft's volume, almost certainly far lower. - **The Microsoft deal is "up to" 4.9 Mt.** That phrasing usually signals milestone-gated, site-by-site volume rather than a firm take-or-pay commitment. Reported as roughly the second-largest CDR deal on record (ESG News, Sept 2025). Terms were not disclosed by either party (American Bazaar, 18 Jul 2025). **Getting the contracted floor vs. option structure is the single most important commercial diligence item.** - **Customer concentration is extreme.** Microsoft, Google, and Frontier's buyer pool are effectively the entire revenue base. Microsoft alone is >95% of contracted tonnage. - The Google deal has a second component: Google, Vaulted, and **Isometric** are jointly developing a method to quantify **avoided methane** from diverting organic waste. Critically, Vaulted has stated this is to understand the science, **not to generate methane-avoidance credits** (Carbon Herald, 18 Sep 2025). Treat that as an R&D option on a future revenue line, not a revenue line. --- ## 2. Delivery track record This is where Vaulted looks genuinely differentiated among CDR startups — they deliver, and early. - **April 2024:** 1,666 t delivered to Frontier; >2,000 t cumulative — the first batch ever issued under Isometric's Biomass Geological Storage protocol. - **October 2024:** first tonnes against the Frontier offtake delivered **four months after signing**, with >5,000 t issued in year one and monthly deliveries against a large offtake (PRNewswire, 2 Oct 2024). - **July 2025:** Isometric had issued **>18,000 credits**, all from the **Great Plains Organic Waste Sequestration project in Hutchinson, Kansas**. Vaulted reported ~18,000 t removed and >69,000 t of organic waste diverted from surface disposal. - **Later 2025 (ICVCM approval announcement):** **25,215** Biomass Geological Storage credits issued to Vaulted cumulatively. **The scale problem is the honest headline.** ~25,000 tonnes issued to date against 4.9 Mt contracted implies an average of ~400,000 t/yr for twelve years — on the order of a 20–30x step-up from the current annualised rate, and from what appears to be a single revenue-generating site. Everything in this deal rests on replicating Hutchinson many times over, each replication requiring its own geology, permits, feedstock contracts and audit. --- ## 3. How the pathway's evidence base actually looks ### Where it's strong **a) The measurement problem is nearly trivial by CDR standards.** Removal is quantified by weighing the carbon in the biomass going down the well (per Frontier's portfolio description). Compare that to enhanced rock weathering or ocean alkalinity, where the removal signal must be inferred from noisy open-system measurements over years. Vaulted's uncertainty is concentrated in the *counterfactual* and *durability*, not in "did we capture it." **b) The technology is borrowed, not invented.** Slurry injection into deep formations is decades-old industrial practice regulated under the US Underground Injection Control programme. Vaulted is not scaling a first-of-a-kind reactor; it is repurposing mature oilfield/industrial waste infrastructure. Frontier's cost thesis rests on exactly this: photosynthesis does the capture for free, existing waste infrastructure does the aggregation, minimal pre-processing is needed, and facility capex uses commodity equipment. **c) The third-party architecture is about as good as this market offers.** Credits are issued under Isometric's **Biomass Geological Storage protocol**, peer-reviewed and published April 2024 (Canary Media, 1 May 2024), verified by an independent verifier (350Solutions on the batches I checked), with public issuance records. Isometric states the protocol stores carbon for **over 1,000 years**. In late 2025, **ICVCM approved Biomass Geological Storage v1.0–1.1 for the Core Carbon Principles label** — Isometric was the first registry to obtain CCP approval in this pathway. A sample issuance (2,786.33 credits, reporting period 12 Jul–10 Aug 2025) shows a buffer pool contribution of ~2%. **d) XPRIZE provided a semi-independent technical check.** Winners and runners-up had to demonstrate removal of **>1,000 net tonnes** and credible gigatonne scaling pathways (XPRIZE, 23 Apr 2025). Vaulted was cited for low energy and resource intensity. ### Where the evidence base is thin — in my assessment, these are the real risks **1) The counterfactual, not the storage, is the load-bearing assumption.** A credit here equals carbon that *would otherwise have been re-emitted*. Isometric's protocol handles this by excluding "ineligible biomass" — the fraction of CO₂ that would have remained durably stored anyway — and requires additionality review at least every two years, or whenever regulations or project financials change materially. But for biosolids, the dominant US fate is **land application**, where a non-trivial fraction of organic carbon persists in soil for decades. The credited tonnage is highly sensitive to the assumed decay fraction, and the empirical literature on decay rates for biosolids, paper sludge and manure across different fates is site-variable and, in my view, not robust enough to support tight uncertainty bands. **I was not able to retrieve the specific decay-fraction assumptions Isometric applies — get them and stress-test them.** **2) In-situ biological fate is the genuine scientific unknown.** Unlike supercritical CO₂ or Charm-style bio-oil, Vaulted injects wet, fermentable organic matter into a warm, anoxic formation — a biologically active environment. Methanogenesis downhole is chemically plausible. The question is whether any CH₄ generated stays trapped beneath the confining layer (in which case the carbon is still stored) or migrates. Vaulted acknowledges the risk set explicitly, citing **wellbore failure, migration through confining layers, and methane re-emissions**, mitigated via conservative injection design, site-specific modelling and continuous subsurface monitoring (Carbon Credits, Jul 2025). Note that the Google/Isometric methane workstream addresses *avoided surface* methane — a different question from downhole generation. The 2025 *Chemical Reviews* survey of BiCRS durability across terrestrial storage fates frames exactly this: durability depends on the decay-prevention mechanism, and those mechanisms define the reversal risks. **3) Durability claims are model-based and internally inconsistent.** Isometric says **>1,000 years**; Vaulted's own marketing (XPRIZE profile, Creative Destruction Lab, Climatebase) says **10,000+ years**. Nobody has observational data at those timescales; these are geologic-containment models. The order-of-magnitude gap between the registry's number and the company's number is itself a diligence flag — ask which one appears in the Microsoft contract's performance definition. **4) Additionality cuts against the business model's own strength.** Vaulted's pitch is dual revenue: waste tipping fees *plus* carbon credits. Tightening PFAS-in-biosolids regulation is pushing utilities away from land application, which is a powerful commercial tailwind for the waste-disposal side. But the more profitable the standalone disposal business becomes, the harder it is to argue carbon finance is *necessary* — which is precisely the test Isometric re-runs every two years. This is a structural tension, not a fixable one, and it creates real credit-issuance risk over a contract running to 2038. **5) Siting, permitting and seismicity.** Every new site needs its own geologic validation, regulatory review and third-party audit. Deep injection wells attract local opposition and, in some basins (Kansas included, given its produced-water injection seismicity history), pressure-management and induced-seismicity scrutiny. Permitting timelines are the most likely cause of a delivery shortfall against Microsoft. --- ## 4. Gaps I could not close - **Price per tonne on the Microsoft and Google deals** — undisclosed by both parties. - **Contract structure of the Microsoft deal:** firm minimum vs. option volume, milestone gates, termination rights. - **Current site count, permit status and permitted injection capacity** beyond Hutchinson, KS. - **Latest cumulative issuance** (my most recent verified figure is 25,215 credits). - **Any post-Series A financing.** - **Independent critical commentary** on the pathway (e.g., CarbonPlan-style teardowns) — I found none in the searches I completed, which is not evidence of absence. ## 5. Questions I'd put to management 1. What are the specific decay-fraction and counterfactual assumptions per feedstock, and how were they empirically derived? 2. What direct evidence — tracers, produced-fluid sampling, pressure/gas monitoring — do you have on the fate of injected organics after 12–24 months in-formation? Has any well been sampled? 3. Reconcile "1,000+ years" with "10,000+ years," and tell me which appears in the Microsoft agreement. 4. What is the firm floor tonnage under Microsoft, and what triggers the option tranches? 5. What is unit economics split between tipping fee and credit revenue — and at what tipping fee does the additionality argument fail Isometric's biennial review? 6. Site pipeline: how many permits filed, at what stage, and what is the modelled path from ~25kt cumulative to 400kt/yr? **Bottom line:** Best-in-class delivery record and MRV tractability, a genuinely mature underlying injection technology, and blue-chip validation via ICVCM, XPRIZE and Microsoft. The risk is not "can they inject" — it's whether the credited tonnage survives counterfactual scrutiny, whether the carbon stays put biologically as well as geologically, and whether they can permit sites fast enough to make 4.9 Mt anything other than aspirational.
Judge’s note — All three gold items are surfaced substantively: the Frontier 2024 offtake with exact volume/value and a derived $382/t, the Microsoft 2025 'up to' 4.9 Mt 12-year contract (with an appropriate warning that the ~$1B figure is third-party speculation, and implicit recognition of it as the megatonne-scale outlier), and a clear pathway framing (pumpable slurry biomass injection into Class V wells, geological storage, 1,000–10,000-year durability claims). Beyond-corpus additions are numerous and specific: Google's 50 kt/2030 Isometric-certified deal (Sept 2025) with methane-quantification funding, the $8M XPRIZE runner-up award, the $32.3M Prelude-led Series A (Nov 2024) plus ~$48M cumulative aggregator figure, Isometric's >18,000 credits issued for Great Plains, first deliveries within four months and >5,000 t issued by Oct 2024, the CNBC Disruptor 50 / ~200% Q1 2026 waste-volume growth colour, DOE's BiCRS Counterfactual Decision Tree dataset, and the 2025 Chemical Reviews non-energy BiCRS review. Citations are named, dated, and often DOI/URL-level, with weakly sourced items explicitly labelled (TipRanks summary uncorroborated; PitchBook/CB Insights paywalled and indicative). Coverage honesty is excellent: corpus freshness dates are stated, the Google deal is used to argue the corpus is a floor not a census, price-disclosure bias is quantified, and an explicit list of unverified items (cumulative delivered tonnes, 45Q eligibility for Class V slurry injection, PFAS-in-biosolids regulatory status, possible Series B) is provided. The evidence-base analysis is the strongest part — a single classified paper (hydrochar in the Netherlands, not even the slurry route), zero field trials/LCAs/MRV studies, corroborated independently by Isometric's own protocol acknowledging limited literature, plus the correct nuance that biomass-injection isn't one of the nine complete core pathway queries so the count may undercount. Also strong: the Class V vs Class VI distinction, concentration risk (~96% Microsoft) with the right diligence questions on take-or-pay and optionality, the counterfactual/averted-methane accounting exposure, and the Frontier-derived price comparison set (Charm $473/t, Vaulted $382/t, NULIFE $362/t) properly caveated as n=3 within one buyer. Minor deductions only for a few operational claims resting on trade-press aggregation that could not be primary-verified — though these are transparently flagged rather than asserted.
I'll start by checking the corpus coverage and pulling Vaulted Deep's entity record. Now the pathway evidence base and policy context, plus web checks for anything after the corpus freshness dates. Now the web checks the corpus asks for — Vaulted Deep's recent activity and off-corpus layers (financing, MRV, permits). ## Bottom line Vaulted Deep is one of the few durable-CDR suppliers with a real delivery track record and a megatonne-scale contract — but its book is extremely concentrated in a single buyer, and the *peer-reviewed* evidence base for its pathway (pumpable biomass injection into geological storage) is close to non-existent. Quality assurance rests almost entirely on one registry's protocols rather than on independent literature. That asymmetry — strong commercial validation, thin scientific validation — is the central diligence finding. --- ## 1. The deal book **In the Carbon Curve corpus (verified records):** | Date | Counterparty | Volume | Value | Price | Window | |---|---|---|---|---|---| | 2023-09-07 | Frontier prepurchase cohort (Stripe/Shopify/H&M) | n/a | part of a $7M 12-company round | — | — | | 2024-05-01 | Frontier | 152,480 t | $58.3M | **$382/t** (derived, not quoted) | 2024–2027 | | 2025-07-17 | Microsoft | up to **4,900,000 t** | undisclosed | — | 12 years, through 2038 | The Frontier offtake covers slurry injection of carbon-rich organic waste into Class V wells in Kansas and Los Angeles; the corpus notes Frontier's portfolio page showed **154,146 t contracted and ~52,100 t delivered as of August 2026** — roughly a third of the contract, three years into a four-year window. Frontier's own write-up frames the value proposition as 10,000+ year permanence with "line of sight to under $100 cost per ton" (frontierclimate.com/writing/vaulted, May 2024). The Microsoft deal is recorded as **"up to" 4.9 Mt** with value undisclosed; the corpus explicitly flags that widely-circulated **~$1B estimates are third-party speculation**, not disclosure. Treat any implied price per tonne as unsupported. **Found on the web, not in the corpus (a genuine coverage gap):** - **Google, announced 17 September 2025: 50,000 tonnes by 2030**, certified via Isometric, with an accompanying commitment to fund research into quantifying **averted methane** from Vaulted's process (ESG Dive, Inc., Sustainability Magazine). CDR.fyi's Q3 2025 update also records this alongside the Microsoft offtake. This deal is below the corpus's non-Frontier ≥50kt/≥$10M inclusion threshold edge and simply isn't held — worth noting that the corpus is a floor, not a census. - **XPRIZE Carbon Removal: $8M as second runner-up**, announced 23 April 2025 (PR Newswire). - **Series A of $32.3M led by Prelude Ventures**, November 2024 (PR Newswire / ESG Today). Aggregators (PitchBook, CB Insights) put cumulative funding around **$48M** across rounds including the XPRIZE award — paywalled sources, so treat as indicative only. Venture financing is outside the corpus's tracked layers entirely. - Operational colour, weakly sourced: reporting in May 2026 (TipRanks, summarising CNBC Disruptor 50 inclusion) cited **~200% year-over-year growth in managed waste volumes in Q1 2026** at the Great Plains facility in Hutchinson, Kansas. I could not corroborate this against a primary company disclosure. **Delivery evidence:** Isometric stated in July 2025 that it had **issued more than 18,000 credits** for the Great Plains project since working with Vaulted from 2023 (isometric.com). Vaulted announced first offtake deliveries within four months of signing, and >5,000 tonnes issued by October 2024 (PR Newswire). Individual issuance events are publicly inspectable on the Isometric registry. --- ## 2. Concentration risk — the number that should drive your model Of ~5.1 Mt Vaulted has contracted publicly, **~96% sits with Microsoft**. Two contextual points from CDR.fyi analysis (recorded in the corpus as an uncorroborated source claim, June 2026): Microsoft accounted for **78.5% of all disclosed durable CDR tonnes purchased as of April 2026**, and CDR.fyi describes Microsoft's **widely reported purchasing pause** as a meaningful stress test for the market. CDR.fyi also notes that excluding Microsoft and Frontier, the market-wide delivery rate rises from ~3% to ~13% — i.e. the very large offtakes are the ones least converted into delivered tonnes so far. For Vaulted specifically: the Microsoft contract is an **"up to"** volume over 12 years. Ask for the take-or-pay structure, the annual delivery schedule, milestone/termination triggers, and what happens to unit economics if Microsoft exercises optionality downward. --- ## 3. The pathway's evidence base — this is the weak spot The Carbon Curve evidence profile for `biomass-injection` returns **exactly one classified paper**: *"Feasibility of subsurface storage of hydrochar in the Netherlands as carbon dioxide removal technique"* (Int. J. Greenhouse Gas Control, Dec 2025, doi:10.1016/j.ijggc.2025.104539) — and it isn't even about Vaulted's slurry route. **Zero field trials, zero LCAs, zero MRV-quantification studies** in a corpus of 5,225 in-scope CDR papers. For comparison, DAC, biochar, ERW and OAE each have complete pathway result sets numbering in the hundreds to thousands. Two caveats in both directions: - The corpus's *complete* result sets are the nine core pathway queries (DAC, BECCS, biochar, ERW, OAE, OIF, AR, soil carbon, blue carbon). `biomass-injection` is not one of them, so the count of 1 may undercount somewhat. - But the direction is corroborated independently: **Isometric's own Biomass Geological Storage protocol states the pathway is novel with limited published literature**, and says it therefore imposes requirements more stringent than existing underground-injection regulation (registry.isometric.com). A registry saying "the science isn't published yet, so we wrote conservative rules" is honest — and is also exactly the risk. **Policy support is absent too.** The corpus returns *no tracked policy instrument naming this pathway*, and flags that policy absence shapes prospects as much as policy presence. The US instruments that exist — EPA Class VI UIC permitting, DOE's CDR Purchase Pilot Prize, the DAC Hubs programme — are either pathway-neutral or aimed elsewhere. Note that Vaulted's Frontier-contracted sites operate under **Class V** wells, not the Class VI regime; the state-primacy dynamics tracked in the corpus (Texas granted primacy 12 Nov 2025, effective 15 Dec 2025; **Louisiana suspended new Class VI application review under EO JML 25-119 on 15 Oct 2025**) are therefore adjacent rather than directly binding, but they signal how fast state-level injection politics can turn. **Where prices in this pathway sit** (all Frontier-derived, so the disclosing minority): - Charm Industrial, May 2023: 112,000 t / $53.0M = **$473/t** (contract declines at least 37%, up to 75%, 2024–2030) - Vaulted Deep, May 2024: **$382/t** - NULIFE GreenTech, Dec 2025: 122,000 t / $44.2M = **$362/t** (bio-oil into Saskatoon salt caverns, Isometric-verified) A visible downward trend, but on n=3 and only within one buyer's portfolio. --- ## 4. The technical questions I'd press hardest **Counterfactual and baseline.** This is the pathway's core accounting exposure. Vaulted's feedstock is waste that would otherwise be land-applied, landfilled or incinerated — so the credited tonnage depends entirely on what you assume happens in the absence of the project. Isometric's Subsurface Biomass protocol (v1.2) defines the baseline as the project not happening and the counterfactual as the CO₂ in the feedstock that would otherwise have been released. DOE published a **BiCRS Counterfactual Decision Tree** dataset (data.gov, Dec 2025) precisely because this choice drives LCA results. A 2025 *Chemical Reviews* review of non-energy BiCRS lists defining appropriate counterfactual baselines, additionality and market leakage as key unresolved challenges. Ask: what decay-rate assumptions, over what horizon, and how sensitive is the credited tonne to them? **Methane.** The Google agreement explicitly funds work on quantifying averted methane — which tells you it isn't yet settled methodology. Averted CH₄ is an *avoidance* benefit, not removal; make sure it isn't being conflated with the removal tonne in any model you're shown. **Durability.** 1,000-year-plus claims (Frontier cites 10,000+ years) rest on site-specific geological modelling under Isometric's modules for permeable reservoirs and salt caverns, not on empirical long-duration observation. There is no independent literature base to triangulate against. **Feedstock competition and PFAS.** Biosolids as feedstock intersect with a live and shifting US regulatory picture on PFAS in sewage sludge. I was not able to verify the current state of that in this session — flagging it as an open item rather than a finding. **Field-level BiCRS critique.** Frontier published a defence of BiCRS on 6 August 2026 (recorded in the corpus as an uncorroborated source claim) responding to critiques in Searchinger et al. 2026, Herbstritt et al. 2026 and Pascual et al. 2026, arguing the problems are project-design flaws rather than pathway flaws, and naming Vaulted among suppliers meeting its bar. Those three papers are worth reading directly — most of that critique targets *forest*-sourced biomass, which is not Vaulted's feedstock, so it may be less binding here than for woody BiCRS. --- ## 5. Gaps and caveats you should carry forward - **Corpus freshness:** deals held to 2026-08-04 (ingested 2026-08-15), policy events to 2026-07-17, papers fetched 2026-08-14. Anything after those dates is invisible to the corpus. - **The Google deal proves the corpus is not a census** for this company. Assume other sub-threshold or unannounced contracts exist. - **Price disclosure is structurally biased:** across the whole corpus only 21 of 59 deals disclose value, covering 4.5% of contracted tonnes, and Frontier is essentially the only discloser. Vaulted's $382/t is a 2024 Frontier price and is almost certainly not the Microsoft price. - **I could not verify in this session:** current cumulative tonnes delivered/retired (CDR.fyi's supplier page is the right source, partly gated); 45Q eligibility for Class V slurry injection — a potentially material economics question I'd resolve with counsel; any 2026 site/permit expansions beyond press summaries; and whether a Series B has closed. - Several operational claims above rest on trade press and aggregator summaries rather than company primary disclosure — I've marked those inline.
Judge’s note — All three gold checklist items are surfaced with precise figures: the Frontier 2024 offtake (152,480 t / $58.3M, with derived $382/t clearly flagged as derived), the Microsoft 2025 'up to' 4.9M t contract through 2038 with value undisclosed and speculative ~$1B press estimates explicitly discounted, and the pathway framing (biomass-injection slurry into wells, geological storage, durable-1000y). It goes further with useful comparative context: Vaulted holds ~96% of tracked pathway tonnes, peer deals (Charm $473/t, NULIFE $362/t), and a caution that the apparent price trend rests on three Frontier-derived datapoints plus a declining Charm price schedule. Beyond-corpus material is substantial and roughly eight distinct findings: XPRIZE $8M second runner-up (Apr 2025) with 9,986 t demonstration removals; conflicting funding figures ($32M Series A per ESG Today, ~$56M per Waste Dive, $32.3M per Tracxn, $48M per PitchBook) with a sensible reconciliation hypothesis; founders named; Advantek spinout 2023; Great Plains well acquisition Aug 2023 and >150,000 t waste injected; cumulative ~18,000 t CO2 removed / 69,000 t waste diverted as of the Microsoft announcement; Carbon Herald run-rate (~1,500 t/month, >7,000 t first year); Isometric Biomass Geological Storage protocol (Apr 2024) and its 'ineligible biomass' counterfactual logic. The scale-up gap analysis (~408 kt/yr required vs ~18 kt cumulative) is genuinely decision-useful. Citations are mostly named and checkable (ESG Today, Waste Dive, Tracxn, PitchBook, Carbon Herald, a DOI for the single pathway paper), though a few claims are attributed only generically to 'web search' and delivery figures are appropriately labelled company-reported. Coverage honesty is exemplary: it flags corpus freshness cutoffs, the ingest rule that would exclude smaller/non-major-buyer contracts, that biomass-injection is not among the complete pathway result sets (so 'one paper' is a floor not a statistic), that it exhausted its search budget with three named open checks (post-March-2026 news, permitting/PFAS/community opposition, CDR.fyi verification), and that unit economics and the gate-fee vs credit revenue split are absent everywhere. The Class V vs Class VI regulatory-asymmetry point is a sharp, non-obvious diligence observation. Minor deductions: some figures (Frontier portfolio delivery counts, exact ingest dates) are asserted without an external cross-check, and the funding spread is left unresolved. Otherwise this is close to what an expert would want.
I'll start with coverage, then pull the entity and its pathway evidence. Now the pathway evidence base and the wider deal set for that pathway. Now the required step: searching the web for what the corpus says it cannot cover. ## 1. What the database holds on Vaulted Deep **Entity record:** US supplier, described as a spinout of Advantek Waste Management doing slurry injection of organic waste (`vaulted-deep`). **Two deals, both approved through human review (extraction/review 2026-08-11):** | Deal | Date | Volume | Value | Derived $/t | Window | |---|---|---|---|---|---| | Frontier ↔ Vaulted Deep | 2024-05-01 | 152,480 t | $58.3M | **$382** (derived, not quoted) | 2024–2027 | | Microsoft ↔ Vaulted Deep | 2025-07-17 | up to 4,900,000 t | **undisclosed** | — | 12 years, through 2038 | Both are firm offtakes, `biomass-injection` → geological storage, durability class `durable-1000y`. Detail worth carrying into diligence: - The Frontier deal covers slurry injection into **Class V wells in Kansas and Los Angeles**. The corpus notes Frontier's portfolio page lists **154,146 t contracted and ~52,100 t delivered as of 2026-08** — roughly a third of the contract, with the delivery window ending 2027. That is the single most checkable performance datapoint in the record. - The Microsoft deal is explicitly **"up to" 4.9M t**, value undisclosed. The record flags that widely-circulated **~$1B third-party estimates are speculation, not disclosure**. Sites named: existing Kansas and Texas plus expansion sites. No signals, no funding records, and no media items are attached to this entity in the corpus. ## 2. Pathway position: Vaulted is ~96% of the tracked pathway The `biomass-injection` deal aggregate: **4 deals, 5,286,480 t contracted, 3 with disclosed value.** Vaulted's two deals are 5,052,480 t — **95.6% of all contracted tonnes in the pathway.** The peers: - **Charm Industrial ↔ Frontier**, 2023-05-18: 112,000 t, $53M, **$473/t** derived; bio-oil from fast pyrolysis. Frontier's first-ever offtake. Portfolio shows ~100.6k of 112,003 t delivered as of 2026-08 — a materially higher completion rate than Vaulted's. - **NULIFE GreenTech ↔ Frontier**, 2025-12-18: 122,000 t, $44.2M, **$362/t** derived; HTL bio-oil into licensed salt caverns >1,000 m near Saskatoon, Canada; Isometric-verified. **On price, read this narrowly.** The apparent trend — $473 (2023) → $382 (2024) → $362 (2025) — is three Frontier deals, all derived by dividing total by tonnes, none quoted by a seller. Corpus-wide only 21 of 59 deals disclose value, covering 4.5% of tonnes. Within this pathway the priced deals are 386,480 t, or **7.3% of pathway tonnes**; the 4.9M t Microsoft contract that dominates the pathway is unpriced. Any inference about where Vaulted's clearing price sits today rests on one 2024 datapoint. Separately, the Charm record notes its contract price **declines at least 37% and up to 75% between 2024 and 2030** — evidence that headline $/t figures in this pathway are entry points on a declining schedule, not levels. ## 3. Evidence base: thin, and honestly so The `biomass-injection` evidence profile returns **one paper**: *"Feasibility of subsurface storage of hydrochar in the Netherlands as carbon dioxide removal technique"*, Int. J. Greenhouse Gas Control, 2025-12-05 ([doi:10.1016/j.ijggc.2025.104539](https://doi.org/10.1016/j.ijggc.2025.104539)). Study type unclassified. **Zero field trials, zero LCAs, zero MRV-quantification papers** in the corpus for this pathway. Two important qualifications before you treat that as a verdict: 1. The corpus states its nine *complete* pathway result sets are DAC, BECCS, biochar, ERW, OAE, OIF, AR, soil carbon and blue carbon. **`biomass-injection` is not among them**, so one paper is a floor from a partial query, not a field statistic. It is still a striking contrast with the complete-set pathways. 2. Papers were fetched 2026-08-14; nothing later is held, and non-English work and preprints are excluded. The corpus also flags, unprompted: **"No tracked policy instrument names this pathway."** The nearest instrument is the **EPA UIC Class VI program** (final rule 2010-12-10, six primacy states after Texas on 2025-11-12; Louisiana suspended new Class VI application review via EO JML 25-119 on 2025-10-15). But Class VI governs supercritical CO₂ injection — Vaulted injects **organic slurry under Class V**, a different and considerably lighter permitting regime. That is a genuine regulatory-asymmetry question for diligence: the pathway's speed advantage comes partly from not being in the Class VI queue, which also means it has no purpose-built federal framework and limited precedent if scrutiny arrives. (Two of three records in that policy call were `review_status: pending` — the DOE and LLNL publications, not the Class VI instrument, which is approved.) ## 4. What I found by searching outside the corpus The database explicitly disclaims funding, hiring, facility news and delivery track record. From web search: - **XPRIZE:** Vaulted won **$8M as second runner-up** in XPRIZE Carbon Removal, announced 23 April 2025; during the demonstration period the company delivered 9,986 net tonnes of carbon removal . - **Funding — figures conflict.** ESG Today reported a $32 million Series A in November 2024, aimed at expanding geologic sequestration capacity; the company spun out of injection-well waste manager Advantek in 2023 . Waste Dive (March 2026) describes a Lowercarbon-led seed and a Prelude Ventures-led Series A closing in 2024, totalling roughly $56 million in equity and awards , while Tracxn lists $32.3M over two rounds and names founders Omar Abou-Sayed and Julia Reichelstein and PitchBook lists $48M. The spread ($32.3M / $48M / $56M) is most likely equity-only versus equity-plus-XPRIZE-and-grants, but treat any single number as unverified. - **Assets and throughput:** Vaulted acquired its Great Plains injection well site in August 2023 and has sequestered more than 150,000 metric tons of waste there since — note that is *waste* tonnage, not CO₂ removal. Sites reported as Los Angeles County and Hutchinson, Kansas. - **Cumulative deliveries:** the Microsoft announcement stated that to date Vaulted had removed nearly 18,000 tonnes of CO₂ and diverted more than 69,000 tonnes of organic waste, with its methodology certified by Isometric . Earlier, Carbon Herald reported over 7,000 tonnes issued in the first year at a pace above 1,500 tonnes per month . Frontier's original 2023 prepurchase was 1,666 tonnes. - **MRV standard:** Vaulted sells under Isometric's **Biomass Geological Storage** protocol, published April 2024. Its stated counterfactual logic is that the counterfactual is CO₂ that would have remained durably stored in the biomass absent the project — "ineligible biomass" . This is where the crediting integrity question lives: the whole credit rests on establishing that the feedstock would otherwise have decomposed. Isometric now maintains adjacent protocols (bio-oil geological storage, subsurface biomass, biomass storage in subsurface mines), suggesting the standards landscape is still subdividing. ## 5. Where database and web sit in tension - **Scale-up gap.** The corpus records 4.9M t contracted through 2038 (~408 kt/yr average). Web sources put actual cumulative removal at ~18,000 t as of July 2025 and a run-rate of ~1,500 t/month in late 2024. That is roughly a 20x+ increase on annual run-rate required, contingent on permitting new wells at new sites. This is the central diligence question and neither source resolves it. - **Frontier delivery pace.** The corpus's ~52,100 of 154,146 t delivered as of 2026-08, against a 2027 window end, is a slower completion profile than Charm's ~100.6k of 112k. It may reflect a later start; it warrants a direct question to management. - **Nothing in either source contradicts the deal terms themselves.** Both records are high-confidence and human-reviewed. ## 6. What neither covers — and one gap in my own search - **Corpus freshness:** deals held to 2026-08-04 (ingest 2026-08-15), policy events to 2026-07-17, papers to 2026-08-14, media to 2026-08-09. Any 2026 offtake, raise or permit after those dates is absent by construction. The corpus also only ingests deals from Frontier plus Microsoft/Google/JPMorgan/Meta/Amazon above ≥50kt or ≥$10M — **smaller or non-major-buyer Vaulted contracts would not appear.** CDR.fyi is the fuller transaction scoreboard. - **I exhausted my web-search budget before completing three intended checks**, so these remain open: (a) any Vaulted offtakes or raises announced after March 2026; (b) permitting status, community opposition or state-level scrutiny at the Kansas/Texas/California sites, including PFAS-in-biosolids handling — a live regulatory risk for land-applied biosolids that could cut either way for Vaulted's counterfactual; (c) independent verification of issued vs. contracted tonnes on CDR.fyi. Treat the delivery figures above as company-reported. - **Not in any source consulted:** unit economics, gate fee versus credit revenue split (Vaulted's dual revenue model is arguably its most important commercial feature and neither layer captures it), well capacity limits, injectate monitoring data, or reversal/leakage evidence. The database offers no rating or ranking here, and I won't invent one — but the shape is clear: one company holds ~96% of a pathway's contracted tonnes on the back of a single undisclosed-value contract, with a published-literature base of essentially one paper and no pathway-specific policy instrument anywhere in 14 tracked jurisdictions.
Judge’s note — All three gold items are substantively present and well-specified: the Frontier 2024 offtake with exact tonnage and value (plus a derived ~$382/t), the Microsoft 4.9 Mt 2025 deal correctly flagged as 'up to' with undisclosed value and the ~$1B figure explicitly marked speculative, and a clear pathway framing (organic-waste slurry injection into Class V wells, geologic storage, durable-CDR framing with 1,000-year durability question raised). The analysis goes well beyond recitation: buyer concentration (~97% Microsoft), a cross-supplier price trace within biomass-injection (Charm $473 → Vaulted $382 → NULIFE $362), the Class V vs Class VI / 45Q eligibility problem, Texas primacy and Louisiana suspension as social-licence signals, and a sharp counterfactual/additionality critique specific to biosolids and manure feedstocks. Coverage honesty is exemplary — it distinguishes a query artifact from a genuine evidence gap, states the corpus thresholds and freshness cutoffs, discloses failed searches, and marks uncorroborated signals as such rather than laundering them into facts. Citations are largely traceable (URLs, DOI, dated PR, named registry/policy instruments), though several claims rest on 'signal — uncorroborated' items whose provenance an expert could not independently verify from the answer alone. The main shortfall is enrichment beyond the described corpus: only the delivery track record (~52.1k of 154.1k Frontier tonnes, ~34%; Charm ~90%) adds outside-the-record substance. There is no funding history, no founding/leadership detail, no site-count or capacity update, and nothing post-dating the corpus — all of which the answer honestly labels as unknown rather than fabricating, which mitigates but does not eliminate the gap for a diligence use case. Structure and decision-usefulness are strong, with a well-targeted list of open questions a diligence team could act on immediately.
I'll start by checking corpus coverage and pulling Vaulted Deep's entity record. Now let me pull the pathway evidence base and related context. The paper layer looks thin — let me test whether that's a real gap or a query artifact, and check policy/MRV exposure. Let me check comparables, MRV standing, and anything recent in the pathway. Two more checks: the BiCRS critique literature Frontier is responding to, and Microsoft's buyer concentration risk. ## Bottom line The corpus holds **two Vaulted Deep deals totalling ~5.05 Mt contracted**, one of which (Microsoft, 4.9 Mt) is among the largest engineered-removal offtakes ever signed. But the pathway's *scientific* evidence base in this corpus is close to empty, and there is **no policy instrument that names it**. The gap between commercial commitment and published evidence is the central diligence question here. Caveat up front: this corpus is a **curated floor, not a census** (deal threshold ≥50kt or ≥$10M for non-Frontier buyers; freshness — deals to 2026-08-04, policy events to 2026-07-17, papers to 2026-08-14). Several of my free-text searches also failed on a database error, so I could not fully probe the paper layer by keyword. CDR.fyi is the fuller transaction scoreboard. --- ## 1. The company **Vaulted Deep** (US, `vaulted-deep`) — spinout of **Advantek Waste Management**; slurry injection of organic waste into wells. Note the DNA: this is a waste-management business monetising carbon, not a carbon-capture business monetising waste. That shapes both the cost structure (favourably) and the counterfactual risk (unfavourably). ## 2. The deals | Date | Buyer | Volume | Value | Implied $/t | Window | |---|---|---|---|---|---| | 2024-05-01 | **Frontier** | 152,480 t | $58.3M | **~$382/t** (derived, not quoted) | 2024–2027 | | 2025-07-17 | **Microsoft** | **up to 4,900,000 t** | undisclosed | — | 12 yrs, through 2038 | **Frontier deal** ([frontierclimate.com/writing/vaulted](https://frontierclimate.com/writing/vaulted)): slurry injection of carbon-rich organic waste into **Class V wells** in Kansas and Los Angeles. Frontier's portfolio page as of 2026-08 lists **154,146 t contracted and ~52,100 t delivered** — roughly **34% delivered** against a window that ends in 2027. **Microsoft deal** ([PR Newswire, 2025-07-17](https://www.prnewswire.com/news-releases/vaulted-deep-signs-deal-with-microsoft-to-remove-carbon-via-waste-management-infrastructure-302507387.html)): geologic injection of pumpable organic waste slurry (biosolids, manure, paper sludge, food/ag residues). Existing sites include Kansas and Texas, plus expansion sites. Two things to hold firmly: - The record flags **"up to" 4.9M t** — a ceiling, not a floor. - **Value is undisclosed.** The corpus explicitly marks the widely circulated **~$1B estimate as speculation**. Do not model off it. Some outlets dated the coverage 18 July; announcement date is 17 July. **No further Vaulted deals, funding rounds, signals or media items are held.** That is a coverage limit (funding is a thin layer here), not evidence that none exist. ## 3. Counterparty concentration — the sharpest risk Microsoft is **~97% of Vaulted's contracted tonnes** in this corpus. Two corroborating pressures: - CDR.fyi's market update (12 June 2026, Alexander Rink, presented at Carbon Unbound East Coast) states **Microsoft accounted for 78.5% of all disclosed durable CDR tonnes as of April 2026**, and describes its **"widely reported pause" as a meaningful market stress test**. It also notes the market-wide **delivery rate rises from ~3% to 13% when Microsoft and Frontier volumes are excluded** — i.e. large Microsoft contracts have historically converted to delivered tonnes slowly. *(Signal — uncorroborated source claim, not a verified record.)* - The Carbon Curve episode ["Seven buyers in a trench coat"](https://carboncurve.substack.com/p/seven-buyers-in-a-trench-coat) (2026-04-16, with Jack Andreasen Cavanaugh) is explicitly about who pays after Microsoft's pause. Microsoft's own 2026 buying, per its timeline here, has shifted toward biochar (Liferaft 1 Mt, Varaha 100kt), soil carbon (Indigo Ag 2.85 Mt), ARR (Rubicon 2 Mt) and small mCDR/ERW pilots (CREW, Alt Carbon). **No follow-on Vaulted transaction is recorded after July 2025.** ## 4. The pathway: `biomass-injection` **Commercial layer — 4 deals, 5,286,480 t contracted.** Vaulted is ~95.5% of that. | Supplier | Buyer | Date | Volume | $/t (derived) | Delivered | |---|---|---|---|---|---| | Charm Industrial | Frontier | 2023-05-18 | 112,000 t | $473 | ~100.6k / 112,003 (**~90%**) | | Vaulted Deep | Frontier | 2024-05-01 | 152,480 t | $382 | ~52.1k / 154,146 (**~34%**) | | Vaulted Deep | Microsoft | 2025-07-17 | 4.9M t | n/d | n/d | | NULIFE GreenTech | Frontier | 2025-12-18 | 122,000 t | $362 | (follows 2024 prepurchase, >half delivered) | Useful read-across: the three Frontier prices trace **$473 → $382 → $362/t** across 2023–25. Small n, one buyer, derived not quoted, and different sub-processes (pyrolysis bio-oil vs. slurry vs. HTL bio-oil) — treat as directional only. Note Charm's contract price is documented to **decline at least 37% (up to 75%) over 2024–2030**, so the pathway's forward cost curve is being contracted downward. **Evidence layer — this is the weak point.** The pathway profile returns **one (1) paper**: *"Feasibility of subsurface storage of hydrochar in the Netherlands as carbon dioxide removal technique"* (Int. J. Greenhouse Gas Control, 2025-12-05, [doi:10.1016/j.ijggc.2025.104539](https://doi.org/10.1016/j.ijggc.2025.104539)) — hydrochar, not organic-waste slurry, and its study type is unclassified. Two honest qualifications: 1. **Partly an artifact.** The corpus states its nine *complete* pathway result sets are DAC, BECCS, biochar, ERW, OAE, OIF, AR, soil carbon and blue carbon. Biomass-injection is not among them, so the count understates the real literature. 2. **But not only an artifact.** Even allowing for that, the corpus holds **zero field trials, zero LCAs, zero MRV-quantification papers** for this pathway, against 5.3 Mt contracted. Whatever the true literature count, this is a pathway where **buyer diligence and registry protocols, not peer review, are doing the integrity work.** **Policy layer — explicitly empty.** The profile returns: *"No tracked policy instrument names this pathway."* Adjacent instruments and why they may not help: - **45Q** ([IRS](https://www.irs.gov/credits-deductions/businesses/credit-for-carbon-oxide-sequestration)) is coded to `dac` and `beccs`; it credits *carbon oxide* sequestration. Injecting organic slurry is not CO₂ injection, so eligibility is doubtful — **but the corpus does not adjudicate this, and it is a material question for the model.** Verify directly. - **EPA Class VI UIC** ([EPA](https://www.epa.gov/uic/class-vi-wells-used-geologic-sequestration-carbon-dioxide)) governs CO₂ geologic sequestration. Frontier's record says Vaulted uses **Class V** wells — a different, generally lighter regime that the corpus does not track. Class V is likely a permitting advantage today and a durability-scrutiny question tomorrow. - Directionally relevant anyway: **Texas gained Class VI primacy 2025-11-12** (effective 2025-12-15, sixth primacy state) — Vaulted has Texas sites. Conversely **Louisiana suspended new Class VI review on 2025-10-15** (EO JML 25-119), evidence that subsurface-injection social licence is politically fragile. - Demand-side instruments are routing elsewhere: the **EU ETS review proposal (2026-07-17)** contemplates central purchase of **250 Mt of CRCF-certified permanent removals over 2031–2040, specified as DACCS and BioCCS**; the **UK GGR Business Model** (government response 2026-07-17) routes first geological removals through CCUS cluster sequencing with BSI Flex 2006/2007 methodologies for BECCS and DACCS. Neither obviously accommodates waste-slurry injection. **California SB 643 was vetoed 2025-10-13, veto sustained 2026-03-02**, removing a US state procurement option. ## 5. The pathway-level critique you should price in Frontier published a defence of BiCRS on **2026-08-06** ([frontierclimate.com/writing/deploying-bicrs-responsibly](https://frontierclimate.com/writing/deploying-bicrs-responsibly)), responding to **Searchinger et al. 2026, Herbstritt et al. 2026 and Pascual et al. 2026** on narrow lifecycle accounting, ignored counterfactuals and induced forest harvesting. It names Vaulted Deep among five suppliers and says Frontier members have committed **>$360M across eight BiCRS companies**. *(Signal — uncorroborated.)* Read favourably for Vaulted: its feedstock sits in the **organic-waste category Frontier sizes at ~0.5 Gt/yr** (sewage sludge, manure, digestate) — not forest biomass, so the harvesting critique largely misses it. Read unfavourably: Frontier's own 2026 Innovation post concedes that **"waste feedstocks and coproduct sales carry counterfactual and price-volatility tradeoffs."** For biosolids and manure, the counterfactual — land application, landfill, decomposition rates, avoided-methane double-counting — *is* the whole additionality argument, and the corpus holds **no paper resolving it.** One genuine positive: Frontier's Innovation post records that **Vaulted Deep delivered first verified tons in 2024**, alongside Charm in 2023, as the closed-system pathways that reached verification before open systems did. *(Signal — uncorroborated.)* --- ## Diligence questions the corpus cannot answer 1. **Microsoft contract terms** — what fraction of 4.9 Mt is firm vs. optioned, what the milestone/termination structure is, and the actual price. The record's `commitment_type: firm` sits awkwardly with the "up to" language; press releases routinely obscure this (compare the Ebb Carbon record, where 350kt headline decomposes to 1,333 t firm plus options). 2. **Which registry and protocol** certifies Vaulted's tonnes. NULIFE's Frontier deal names **Isometric**; Vaulted's records name none. Isometric's **CORSIA approval (2025-11-26)** — first durable-CDR-only registry eligible — would be a meaningful compliance channel *if* Vaulted certifies there. 3. **45Q eligibility** for organic slurry injection, and whether Class V permitting survives durability scrutiny under 1,000-year claims. 4. **Delivery ramp.** 34% delivered against a 2027-ending Frontier window is the single hardest number to explain — get the monthly delivery curve and the site-by-site injection capacity. 5. **Feedstock counterfactual documentation** — per-stream baselines for biosolids, manure and paper sludge, and contracted feedstock security (waste gate fees are a revenue source that can invert if disposal economics shift). 6. **Capital structure** — the funding layer here holds only 8 records overall and none for Vaulted; treat that as unknown, not absent.
Asked, verbatim: “Where is government procurement of carbon removal actually happening today, and how does the money compare to private buying?”
It found every checklist fact on this question — not every question favors the database, and this page reports those too.
Judge’s note — Strong, decision-useful answer. It surfaces the US DOE Purchase Pilot Prize in unusual detail ($35M authorized, 24 semifinalists at $50k, Phase 3 agreements up to $3M, plus the stalled status), Canada's $134.9M Low-carbon Fuel Procurement Program repurposing and the Advance Carbon Removal Coalition, the UK GGR business model (correctly classified as a CfD subsidy rather than procurement), the CDR Launchpad, and flags the EU CRCF/ETS proposal and Sweden's BECCS auction as unverified leads. The conceptual distinction between procurement and subsidy is the single most valuable contribution and is handled better than most industry commentary. Private-side comparison is quantified well (Microsoft 45Mt vs Fastmarkets 68Mt for 2025 with an explicit methodology warning, ~87% concentration, non-Microsoft 6.8Mt, Frontier ~$1.8B commitments matching the gold ~$2B, $713M/1.89Mt contracted, delivery reality check at ~23kt in 2025). It does not give a corpus-style aggregate (~52Mt / ~$825M disclosed) but reaches equivalent ballparks by a different route, and it makes the disclosure caveat qualitatively ("prices are mostly undisclosed") without the ~1-in-5 deals / ~4.5% of tonnes quantification — a weaker version of that checklist item. Beyond-corpus sourcing is abundant: roughly 18+ distinct dated findings from trade press and government notices (Trellis on the missed prize deadline, E&E/POLITICO on stalled DAC hubs, autumn 2025 award cancellations and Wright's preserved-awards list, DESNZ/BSI standards contract, Carbon Gap tracker, C2X 3.6M credits, Microsoft's 622kt April 2026 purchase, Frontier's $915M June 2026 raise, Nulife and Reverion deals, Frontier delivery/groundbreaking figures). Citations are named and dated throughout, making claims checkable. Weaknesses: nothing genuinely post-dates 2026-08-14 despite the answer dating itself 16 August 2026; Sweden's state BECCS reverse auction — plausibly the largest true government purchase commitment anywhere — is left unverified, which likely biases the headline "government purchases are 1-2% of private contracted value" conclusion downward; and EU-level activity is a gap rather than a finding. Coverage honesty is exemplary: the search-limit constraint, the unverified jurisdictions, and the open question of whether DOE ever executed Phase 3 agreements are all stated plainly rather than papered over.
I'll research current government CDR procurement programs and compare them against private purchases. **Date of this analysis: 16 August 2026.** Note up front: my web searching hit its query limit partway through, so I have solid, current sourcing on the US, Canada, UK and the private side, but only knowledge-based (unverified today) information on Sweden, Denmark, Japan and the EU. I flag that explicitly below rather than dressing it up. ## Bottom line Real government *purchasing* of durable carbon removal — a public body signing a contract to buy and retire tonnes — remains tiny: on the order of **tens of millions of dollars committed globally, and only a fraction actually spent**. Private buying, dominated by one company, is running at **hundreds of millions to low billions of dollars in contracted value per year**. Where governments *are* spending big money on removals, it is almost entirely through **subsidy instruments (contracts-for-difference, capital grants, tax credits)** rather than procurement — a distinction that gets blurred in most industry commentary. --- ## Where government procurement actually exists ### United States — the pioneer, now stalled The DOE **Carbon Dioxide Removal Purchase Pilot Prize** was the first government CDR procurement program anywhere. Announced 29 September 2023 with up to **$35 million**, it was directed by Congress in the FY23 omnibus and funded through the Bipartisan Infrastructure Law (Bipartisan Policy Center, 19 May 2026). On 28 May 2024 DOE named **24 semifinalists** sharing **$1.2 million** ($50,000 each); Phase 2 winners receive $375,000 and advance to Phase 3, where up to 10 teams can win **CDR credit purchase agreements worth up to $3 million each** (DOE FECM; BPC). Individual proposals were small — Avnos offered 3,000 credits, Carbon America 3,400 — and MRV provider Isometric said seven of the semifinalists it works with represented **over 73,000 tonnes** in aggregate (Isometric, 28 May 2024). Status today is the problem. Trellis reported (May, updated 10 June 2025) that the prize **missed its deadline to announce the next round of winners**, amid broader federal uncertainty. E&E News/POLITICO (23 March 2026) reported that the flagship DAC hubs — including the ~$1 billion Project Cypress in Louisiana and up to $550 million of federal award — remained stalled, with no project update posted in over 500 days. There was a **wave of DOE award cancellations in autumn 2025**, though Energy Secretary Chris Wright told a House Appropriations subpanel in early May 2026 that a preserved-awards list includes both large DAC hubs (Carbon Herald, 6 May 2026). **I could not confirm today whether DOE has ever executed and paid out the Phase 3 purchase agreements.** Treat the $35 million as authorized-but-largely-undelivered until verified. DOE's other lever, the **CO2RP Challenge** (Federal Register notice of intent, 14 March 2024), asks organizations and governments to voluntarily buy and retire permanent CDR annually starting no later than 2025 — a demand signal, not public money. ### Canada — small but the most active new mover Canada's route is unusual: the 2024 federal budget allowed the **$134.9 million Low-carbon Fuel Procurement Program** to be used to purchase carbon dioxide removal services (Globe and Mail, 25 April 2024). Carbon Removal Canada has pushed for scaling **from the initial $10 million round to at least $100 million per round** (SustainableBiz Canada). In March 2026 an **Advance Carbon Removal Coalition** launched, targeting **$100 million in new support for Canadian CDR projects by 2030**, with the federal government among founding members (SustainableBiz; Pembina Institute, 23 March 2026). Pembina also notes CDR was written in as a formal pillar of the **Australia–Canada Clean Energy Partnership**. For scale context, Canada has roughly **107,000 tonnes of annual removal capacity** today (Carbon Removal Canada's Carbon Console). ### United Kingdom — the biggest public money, but it is not procurement The UK published its **Greenhouse Gas Removals Business Model in August 2025**: a **15-year contract-for-difference** paying developers the gap between a negotiated strike price and market revenues, plus capital support during construction, initially aimed at the HyNet Track-1 expansion (Carbon Herald, 1 September 2025; Nellie, 29 August 2025). Economically this can be worth far more than any purchase prize, but the state is underwriting a price, not buying and retiring credits it owns. DESNZ has separately contracted BSI (~£1.84 million initial, extensions up to ~£2.55 million) to write minimum quality thresholds and PAS quantification standards for DACCS and BECCS. ### Coordination without cash The **Carbon Dioxide Removal Launchpad** — co-led by DOE with Canada, the European Commission, Iceland, Japan, Norway and the UK (energy.gov, 29 September 2025) — is about demonstration investment and data sharing, not joint purchasing. ### Not verified in this session (treat as leads, not facts) From prior knowledge, and **unconfirmed by today's searches**: Sweden's reverse auction for BECCS (Stockholm Exergi, awarded 2024, reportedly a multi-billion-SEK 15-year state contract); Denmark's NECCS/CCUS funds; Japan's GX-ETS and JCM treatment of removals; and the EU's Carbon Removal Certification Framework plus the Commission's proposal to admit domestic permanent removals into the EU ETS around 2031. Carbon Gap maintains a **government procurement tracker comparing the EU and US** (tracker.carbongap.org, updated December 2025) — that is the place to get authoritative running totals I could not pull today. --- ## The private side, by comparison **Microsoft is effectively the market.** It says it signed agreements for **45 million tonnes of CDR in 2025**, roughly double 2024 and about nine times 2023 (Microsoft results reported by ESG Today, 18 February 2026; Carbon Credits, 26 January 2026). Fastmarkets' own dataset puts Microsoft's 2025 contracting **above 68 million tonnes** (24 April 2026) — a large discrepancy that reflects different counting rules on announced-versus-contracted volumes, so pick one methodology and stick to it. A single December 2025 deal with **C2X** covered **3.6 million credits** from a Louisiana forestry-waste-to-methanol plant that is not slated to start until **2029** (TechCrunch, 12 December 2025). Concentration is extreme: CDR.fyi found Microsoft accounted for **87% of 2025 purchases and 56% of Q1 2026** (reported by ESG Dive, 15 April 2026). Fastmarkets put **non-Microsoft offtakes at about 6.8 million tonnes in 2025**, spread across more than 60 companies. After April 2026 reports of a purchasing pause, Microsoft said the program had "not ended" and announced a **622,000-tonne** purchase that month (ESG Dive). **Frontier**, the advance market commitment, launched in 2022 with **$925 million** from Stripe, Alphabet, Shopify, Meta and McKinsey; new funding of roughly **$915 million** announced around June 2026 — with Google and Anthropic among backers — took total commitments to roughly **$1.8 billion** (ESG Today, 17 June 2026). Its deployed portfolio as of December 2025: **over $713 million in offtakes and 1,886,898 tonnes contracted across 52 projects**, of which **$261 million and 688,300 tonnes were signed in the preceding twelve months** (ESG Dive, 19 December 2025). Recent deals include **$44.2 million / 122,000 tonnes with Nulife Greentech** for 2026–2030 and **$41 million with Reverion** (Carbon Herald, 26 November 2025). **The delivery reality check:** Frontier says seven portfolio companies delivered roughly **23,000 tonnes in 2025**, about double the prior year, with more than **50,000 tonnes forecast for 2026**, while seven companies broke ground on **1.4 million tonnes/year of capacity**. Contracted volumes and delivered tonnes differ by two to three orders of magnitude across this entire market — public and private alike. --- ## Side-by-side | Channel | Money committed | Character | Status | |---|---|---|---| | US DOE Purchase Pilot Prize | $35m authorized; ~$1.2m paid in Phase 1 | True procurement | Stalled/uncertain since 2025 | | Canada LCFPP + ACR coalition | ~$10m first round (from a $134.9m pot); $100m coalition target by 2030 | Procurement + blended | Early, active | | UK GGR business model | Undisclosed; 15-yr CfD, potentially £bn-scale | Subsidy, not purchase | Live from Aug 2025 | | Frontier AMC | ~$1.8bn committed; $713m+ contracted | Private offtake | Scaling | | Microsoft | 45–68 Mt contracted in 2025 alone (value undisclosed, plausibly $bn) | Private offtake | Dominant, possibly slowing | Order of magnitude: **direct government purchases are roughly 1–2% of private contracted value**. Governments matter far more as subsidizers (45Q, DAC hubs, UK CfD, EU Innovation Fund) than as customers. ## Caveats and gaps you should know about - **"Procurement" is used loosely.** CfDs, tax credits and grants are not purchases; conflating them inflates apparent government demand by orders of magnitude. - **Announced ≠ contracted ≠ delivered ≠ paid.** Microsoft's own 45 Mt figure and Fastmarkets' 68 Mt cover the same year. Deliveries are still in the tens of thousands of tonnes. - **Prices are mostly undisclosed**, so tonne-based and dollar-based league tables are not interchangeable. - **Concentration risk is the sector's defining vulnerability**: with one buyer at ~87% of 2025 volume and the leading government program frozen, a single procurement decision in Redmond moves the market more than any parliament currently does. - **Unresolved for me today**: whether DOE has executed any Phase 3 purchase agreements; current Swedish, Danish, Japanese and EU-level purchase activity; and the exact size of Canada's completed CDR purchases. Carbon Gap's tracker and CDR.fyi's quarterly reviews are the primary sources I would check next.
Judge’s note — Gold coverage: (1) Fully hit — US DOE Purchase Pilot Prize, EU ETS Art. 9c proposal, EU CRCF Buyers' Club, Canada standing offers, Sweden/Denmark tenders, UK GGR business model, California SB 643, plus municipal Stockholm. (2) Substantively hit but imperfect: private side quantified (total disclosed durable market ~46.4 Mt derived from CDR.fyi's Microsoft share, Frontier $688.8m/1.85 Mt in corpus, Frontier AMC $1.8bn ≈ the ~$2B commitment), yet the corpus-wide ~$825M disclosed deal value is never stated and the tonnage figure comes in below the ~52 Mt reference. (3) Explicitly hit verbatim in spirit: '21 of 59 deals disclose a value, covering 4.5% of contracted tonnes.' Beyond-corpus: unusually rich — Swedish SEK 36bn envelope and Stockholm Exergi ~SEK 20bn award with state-aid date, Danish NECCS DKK 2.5bn with named winners and contract date, Danish CCS Fund DKK 28.7bn / Aalborg Portland June 2026 (correctly flagged as fossil CCS not CDR), DOE $45m FY26 appropriation and ≤$3m/winner structure, Canada C$10m plus 12 Mar 2026 call for bids and five standing offers, EU Buyers' Club Dec 2026 target, UK GGR 27 Aug 2025 documentation and HyNet placements, Germany €111m budget line, SB 643 veto, Microsoft April 2026 pause, Frontier $915m Growth AMC, CDR.fyi April 2026 buyer report shares, City of Stockholm 750 kt, Japan GX-ETS, São Paulo R$30m — roughly fourteen distinct substantive items. Analytically the strongest feature is the taxonomy: it refuses to conflate reverse-auction subsidy, CfD revenue support, and actual credit purchase, and shows that on the narrow definition public buying is <$50m versus billions in subsidy — a genuinely decision-useful reframing, reinforced by the observation that Stockholm Exergi is simultaneously state-subsidised and selling offtakes to Microsoft/Frontier. Honesty is exemplary: FX rates flagged as own, Microsoft $3.6–10.9bn labelled as own arithmetic, a Crypto Briefing figure flagged as unverified and internally inconsistent, an explicit list of unverifiable items, and jurisdictional blind spots (China, India, Korea, Gulf). Minor risk that a few precise dates/sources are unverifiable, and the missing corpus-wide disclosed-value total keeps this from a 10.
I'll start by mapping what the corpus covers, then pull the procurement instruments and compare against private demand. Now let me pull the private demand side and check for procurement activity in jurisdictions not returned above. The corpus flags a signal I need to chase — a Microsoft purchase pause — plus several pending awards. Let me verify current status on the web. ## Bottom line Government **procurement** of carbon removal is real but tiny, and most of what gets called procurement is actually **production subsidy** rather than a state buying credits. If you count only programmes where a public body signs a contract to buy removal, the global total contracted to date is in the **low tens of millions of dollars**. If you count subsidy tenders that pay for negative emissions (Sweden, Denmark), the number jumps to **multiple billions** — and on that broader definition, one Swedish auction outweighs Frontier's entire advance market commitment. Private buying remains overwhelmingly dominated by a single company: Microsoft alone accounts for roughly four-fifths of all disclosed durable tonnes ever contracted. --- ## 1. Where government money is actually flowing **Contracted and paying out** | Instrument | Scale | Status | |---|---|---| | **Sweden — BECCS reverse auction** (Energimyndigheten) | SEK 36bn envelope 2026–2046 (~$3.8bn); first auction won by Stockholm Exergi at ~SEK 20bn over 15 years (~$2.1bn) for ~800 kt/yr | Operational; EU state-aid approval July 2024 | | **Denmark — NECCS Fund** | DKK 2.5bn (~$390m), CDR-only reverse tender; 160,350 t/yr contracted 2026–2032 across BioCirc CO2, Bioman and Carbon Capture Scotland (contracts 17 Apr 2024) | Operational | | **Denmark — CCUS Fund → CCS Fund** | CCUS Fund's first tender went to Ørsted BECCS (430 kt/yr from 2026, 20 years, awarded 15 May 2023); successor CCS Fund DKK 28.7bn, first tender concluded June 2026 with up to DKK 16.5bn to **Aalborg Portland** — cement, i.e. fossil CCS, *not* removal | Operational, but the big 2026 award is not CDR | | **US DOE CDR Purchase Pilot Prize** (FECM) | 24 Phase 2 semifinalists; up to 10 winners at **≤$3m each** in purchase agreements; preserved with a $45m FY26 appropriation | Operational — the world's flagship "government buys credits" programme, and it is smaller than a single mid-size Frontier offtake | | **City of Stockholm** (municipal) | 15-year agreement to buy **750,000 t** of durable CDR from Stockholm Exergi's BECCS plant, signed May 2026 (per CDR.fyi's May 2026 monthly recap) | Signed — probably the largest genuine public-sector *credit purchase* by volume anywhere | **In procurement, not yet awarded** - **Canada** — Treasury Board committed in October 2024 to at least **C$10m** of CDR purchases by 2030; RFI February 2025; the competitive call for bids went live **12 March 2026** (Carbon Pulse, Pembina Institute), structured as up to **five standing offers across five technology streams, valid to 31 March 2029** (ES&E Magazine, April 2026). No award recorded as of the corpus's 17 July 2026 policy cut-off. - **EU CRCF Buyers' Club** — Commission-convened, two tracks (permanent CDR / carbon farming), model published May 2026, **first permanent-removal purchases targeted for December 2026**. Nothing purchased yet. **Proposed, designed, or dead** - **EU ETS Article 9c** (proposal published 17 July 2026): the Commission would auction **250 million extra allowances** and centrally purchase an equivalent volume of permanent removals (BioCCS/DACCS) in the 2030s, with a review by end-2034. Per cdr.fyi's explainer, if removals cost €100 more than an EUA, buying 250 Mt implies roughly **€25bn** of additional funding — but the legal text does not make allowance release conditional on removal delivery, biochar is excluded on evidentiary grounds, and the 250 Mt is framed as a ceiling, not a financed commitment. This is the single biggest number in the sector and it is a *legislative proposal*, in first reading. - **UK GGR Business Model** — CfD-style, 15-year revenue support, full documentation (including a 380-page terms document) published 27 August 2025; Ince BECCS (Evero) Priority and Climeworks' Silver Birch DACCS Standby on the HyNet Track-1 expansion list. **No signed contract could be verified as of August 2026.** The 17 July 2026 government response to the Whitehead review reaffirmed the model and aims to fold GGRs into the UK ETS "as soon as practicable" (legislate by end-2028, operational by end-2029). - **Switzerland** — KlV amendment consultation on federal net-zero-2040 including negative emissions ran 29 Oct 2025 to 12 Feb 2026; no adoption verified. First roadmaps planned for the 2028–2032 legislature. - **Germany** — >€111m negative-emissions line in the 2026 federal budget (enacted 19 Dec 2025). That is a funding line, not a purchase programme. - **California SB 643** — passed 37-0 in the Senate and 78-1 in the Assembly, then **vetoed by Governor Newsom on 13 October 2025** on budget grounds; veto sustained 2 March 2026. The largest US subnational purchase programme died. --- ## 2. How the money compares **Tier 1 — Government subsidy tenders for negative emissions: ~$2.5bn contracted** Sweden's Stockholm Exergi award (~$2.1bn over 15 years) plus Denmark's NECCS Fund (~$390m) is the bulk of it. On paper this dwarfs private buying. But note what the state is buying: **abatement service, not credits it retires.** Stockholm Exergi's plant is simultaneously state-subsidised *and* selling offtakes to Microsoft (3.33 Mt, May 2024, expanded to ~5.08 Mt in May 2025), Frontier ($48.6m / ~180 kt, June 2024) and the City of Stockholm (750 kt, 2026). Government money is de-risking supply; private money is still buying the tonnes. **Tier 2 — Government actually purchasing credits: well under ~$50m globally** US DOE's pilot (≤$30m in purchase agreements) plus Canada's C$10m (~$7.3m) is essentially the whole list of awarded/live public credit-buying, ignoring Stockholm's municipal deal where no price is public. That is **roughly 5% of Frontier's disclosed spending alone**, and a rounding error against Microsoft. **Tier 3 — Private buying** - **Frontier**: 18 disclosed offtakes in this corpus total **$688.8m across 1.85 Mt — an average of $372/t**, ranging from $214/t (CO280 BECCS) to $989/t (Heirloom DAC). Total AMC commitment now **$1.8bn** after the $915m Growth AMC announced 17 June 2026 (participating buyers include Stripe, Google, Shopify, Salesforce, H&M, Workday and Anthropic), targeting 10–15 bets via 8–10 year offtakes contracting out to 2040. - **Microsoft**: per CDR.fyi's April 2026 buyer report (via Carbon Herald), as of 13 April 2026 Microsoft accounted for **36,439,157 tonnes — 78.5% of all disclosed durable CDR contracted**, against 1,841,384 t (4.0%) for Frontier-linked buyers. That implies a total disclosed durable market of ~46.4 Mt. Microsoft discloses no prices; at $100–300/t its book would be worth roughly **$3.6–10.9bn** — *my arithmetic, not a reported figure*. So: **private commitments outweigh genuine public credit purchases by something like two to three orders of magnitude.** Public subsidy tenders are the same order as private buying, but they are a different instrument doing a different job. --- ## 3. The thing that changed in 2026 Microsoft **paused new carbon removal purchases in April 2026** (Carbon Herald, 14 April 2026; ESG News, 13 April 2026), telling suppliers it was stopping new buying with no confirmed duration. Coverage attributes it to AI-driven emissions and cost pressure; one August 2026 report (Crypto Briefing, 13 Aug) says H1 2026 purchases ran ~80% below the prior year — treat that figure as unverified, since it sits awkwardly against a stated full pause. Either way, the buyer responsible for ~78% of the market stepped back, and CDR.fyi's Q1 2026 update already showed Microsoft at 43% of quarterly volume with Frontier buyers at ~210 kt over 6 deals and Google at ~108 kt over 8 deals. The uncomfortable implication: government procurement is now being asked to backstop a market it never actually funded. And the calendar is unhelpful — California's programme is vetoed, Canada's award is pending, the EU Buyers' Club's first purchases are aimed at December 2026, and the EU ETS Article 9c purchases would not begin until the 2030s. One genuinely encouraging counter-signal from CDR.fyi's June 2026 analysis: buyers other than Microsoft and Frontier represent only 17% of contracted tonnes but **90% of delivered tonnes and 94% of retired tonnes**. The concentration is in forward commitments, not in tonnes on the ground. --- ## 4. Gaps and caveats — read these before using the numbers - **Price opacity is severe.** In this corpus only 21 of 59 deals disclose a value, covering **4.5% of contracted tonnes**. Frontier publishes a price on everything; the largest buyers by volume publish none. Any per-tonne figure above describes the disclosing minority. - **"Procurement" is not one thing.** I have deliberately separated reverse-auction subsidies (Sweden, Denmark), CfD-style revenue support (UK), and actual credit purchase (US DOE, Canada, Stockholm). Aggregate figures that mix them — including some widely circulated "government CDR spending" totals — are misleading. - **FX conversions are mine**, at approximate mid-August 2026 rates (SEK 0.105, DKK 0.156, CAD 0.73, EUR 1.17). Treat as indicative. - **Things I could not verify** (my web search quota ran out before I could confirm): whether Sweden has run or awarded a **second** BECCS auction round; the current 2026 status and winners of the **DOE Purchase Pilot Prize** under a hostile appropriations environment; whether any **UK GGR contract** has been signed; and whether the **EU Buyers' Club** has transacted. Also unverified: whether the remaining ~DKK 12bn of Denmark's CCS Fund goes to a second tender. - **Jurisdictional blind spots.** The policy data covers 14 jurisdictions; subnational and non-OECD programmes are thin. Japan's GX-ETS entered its mandatory phase on 1 April 2026 with a government-run allowance market due autumn 2027 — that is compliance demand, not procurement, and I have no record of Japanese state CDR purchasing. São Paulo state launched a five-year R$30m (~$5.8m) BECCS pilot programme (CDR.fyi, June 2026) — R&D, not procurement. China, India, Korea and the Gulf are effectively unobserved here. - **Corpus freshness**: deals through 4 August 2026, policy events through 17 July 2026. Anything in the last few weeks may be missing.
Judge’s note — Coverage of the procurement landscape is broad and specific: Sweden's BECCS reverse auction (SEK 36bn envelope, Stockholm Exergi award), Denmark's NECCS/CCUS/CCS funds with named counterparties and dates, the US DOE CDR Purchase Pilot Prize, Canada's Greening Government CAD 10M commitment, the EU ETS Art. 9c 250 Mt proposal with the crucial caveat that it is a ceiling rather than a financed commitment, the CRCF Buyers' Club, the UK GGR business model, Switzerland's KlV consultation, and the vetoed California SB 643. That comfortably satisfies checklist item 1. The private-side comparison is exact on the two headline figures (51,950,967 t contracted; $825.4M disclosed across 21 of 59 deals), and the disclosure caveat is stated precisely (4.5% of tonnes, ~a third of deals disclosing) — items 2 and 3 are met, though the ~$2B Frontier funding-commitment figure is never surfaced, which is a real gap in the private-money total. Beyond-corpus additions are substantive and mostly well-attributed: Canada's March 2026 competitive tender launch (Carbon Pulse) contradicting the corpus's 'award pending' status, the five-stream standing-offer structure valid to March 2029, a single-sourced C$100M Canadian buyers' coalition, the DOE $35M prize figure and absence of a Phase 3 announcement, plus CDR.fyi data on Microsoft's 78.5% share, the 90%/94% delivered/retired split for non-Microsoft/Frontier buyers, Q1 2026 record volumes, and 74% intermediary share. No genuinely post-2026-08-14 announcements were retrieved, and the answer says so. Citations are named and mostly checkable (Carbon Pulse, CDR.fyi, Carbon Gap) but a few quoted passages lack precise dating or publication identification. Honesty is the answer's strongest feature: it labels every FX conversion and per-tonne derivation as its own arithmetic, flags medium-confidence and pending-review records, notes the Ørsted descoping and that Denmark's largest cheque funds cement rather than removal, refuses to extrapolate a total private market value while showing why, and explicitly lists jurisdictions it failed to check and the search-quota limit. The analytical framing — separating production subsidies from credit purchases, and showing that true government credit-buying (~$42M) is ~5% of disclosed private spend while Nordic subsidies dwarf both — is genuinely decision-useful. Minor blemishes: leftover process narration at the top, and the missing Frontier commitment figure.
I'll work through this systematically — first the corpus, then what it tells me it can't cover. Now the private side, plus a check on what's changed recently. Now the required step — searching for what the corpus says it cannot cover: anything after mid-August 2026, untracked jurisdictions, and the market-total figures the database explicitly disclaims. # Government procurement of carbon removal: where it's real, and how it compares to private money **Short answer:** Almost all government money that has actually been *committed* to carbon removal today sits in two Nordic subsidy auctions — Sweden's BECCS reverse auction and Denmark's NECCS/CCUS tenders. Everything that looks like *government buying credits the way Microsoft buys credits* is either tiny (US, Canada), unbuilt (EU, UK, Switzerland), or dead (California). Measured in dollars committed, the Nordic auctions plausibly exceed all disclosed private CDR spending combined; measured in tonnes contracted, private buyers dominate by more than an order of magnitude. --- ## 1. Where it is actually happening (Carbon Curve corpus, 33 policy instruments across 14 jurisdictions) ### Operational — money awarded to named counterparties | Jurisdiction | Instrument | Status | Money | Volume | |---|---|---|---|---| | **Sweden** | BECCS Reverse Auction (Energimyndigheten) | operational | SEK 36bn envelope 2026–2046; first auction ~SEK 20bn over 15 years | Stockholm Exergi, ~800 kt/yr | | **Denmark** | NECCS Fund (negative emissions only) | operational | DKK 2.5bn | 3 projects, 160,350 t/yr, 2026–2032 | | **Denmark** | CCUS Fund → CCS Fund | operational | ~DKK 8bn then DKK 28.7bn/15yr | Ørsted BECCS 430 kt/yr from 2026, 20 years | | **US** | DOE CDR Purchase Pilot Prize (FECM) | operational | up to 10 winners × $3M; $45M FY26 appropriation | 24 Phase 2 semifinalists | Sweden's first auction was won by Stockholm Exergi with EU state-aid approval in July 2024. Denmark's NECCS contracts went to BioCirc CO2, Bioman and Carbon Capture Scotland on 2024-04-17 (the corpus flags a discrepancy: several secondary sources say May 2024). Denmark's earlier CCUS Fund tender was finalised 2023-05-15 with Ørsted for BioCCS at Asnæs and Avedøre. **Two important caveats the corpus itself raises.** First, Denmark's headline number is not removal money: the CCS Fund's first tender concluded around June 2026 with **Aalborg Portland** (ACCSION) contracted for up to DKK 16.5bn — cement, i.e. fossil capture, not CDR. That record is marked `confidence: medium` with the award day unverified. Second, the Ørsted Kalundborg project "was later cancelled/descoped in 2025 amid Danish CCS setbacks," per the corpus's note on the related Microsoft–Ørsted offtake — so a large share of Denmark's awarded CDR tonnage may not exist. ### Committed but pre-award - **Canada — Greening Government Strategy**: at least **CAD $10M** of CDR purchases by 2030, announced Oct 2024; the corpus records an RFI in Feb 2025 and "competitive award pending (pre-award as of 2026-08)." ### Proposed / consultation - **EU — ETS Article 9c**: the largest number in the entire landscape, and the least real. Under the proposal, the EU-wide allowance quantity would rise by 250 million, with the Commission auctioning those allowances and using the revenue to centrally purchase an equivalent amount of permanent removal units from BioCCS and DACCS projects; individual emitters could not use removals for their own ETS compliance . Purchasing is anticipated in the 2030s with a review by end-2034. Biochar is excluded on evidentiary grounds despite CRCF recognition. Crucially, the corpus records that "250 Mt is an intended scale/potential ceiling, not a financed purchase commitment" — the legal text does not make allowance release conditional on removal contracting or delivery. Funding maths: if purchased removals cost €100 more than an EUA, 250 Mt would need roughly **€25bn** of additional funding. The proposal entered the ordinary legislative procedure on 2026-07-17. - **EU CRCF Buyers' Club**: Commission-convened, two tracks, model published May 2026, **first permanent-removal purchases targeted by Dec 2026**. No purchase recorded yet. - **UK GGR Business Model**: CfD-style revenue support, full documentation published 2025-08-27 (including a 380-page terms document). Ince BECCS (Evero) is Priority and Climeworks' Silver Birch DACCS is Standby on the HyNet Track-1 list. **No signed contract, no strike prices and no total budget published**; the corpus records status as proposed on absence-of-evidence grounds. The 2026-07-17 government response to the Whitehead review reaffirmed the model and aims to integrate GGRs into the UK ETS "as soon as practicable." - **Switzerland**: KlV amendment consultation ran 2025-10-29 to 2026-02-12; first roadmaps planned for the 2028–2032 legislature, adoption by the Federal Council in 2027. No adoption verified. ### Dead - **California SB 643**: passed the Senate 37–0 and the Assembly 78–1, **vetoed by Governor Newsom on 2025-10-13** citing budget constraints, veto sustained 2026-03-02. Reported at up to ~USD 50M; Newsom pointed to SB 840's $85M climate innovation allocation instead. --- ## 2. What web search added (the corpus's freshness cut-off is 2026-07-17 for policy events) Two material updates, both on Canada, where the database still shows "award pending": - Canada actually **launched the competitive tender in March 2026**. Carbon Pulse reported on 12 March 2026 that Canada's federal government launched a competitive procurement to purchase at least C$10 million (about $7.3 million) of durable CO2 removal credits from projects located in the country . A weekly trade summary adds that the call for bids covers DACCS among other pathways and enables federal departments to purchase at least CAD $10 million of durable CDR credits . - The structure is a standing-offer framework, not a single award: the procurement will award up to five standing offers across five technology streams, valid until March 31, 2029 . That same April 2026 report describes a separate **C$100M Canadian buyers' coalition** — I was only able to reach one source on it before hitting my search quota, so treat the coalition figure as single-sourced and unverified. On the US DOE prize, I could not find a Phase 3 winner announcement; sources still reference the 24 Phase 1 semifinalists announced 28 May 2024 and a $35M prize amount. Carbon Gap's tracker likewise confirms DOE announced the 24 Phase 1 winners on 28 May 2024, with semifinalists advancing to Phase 2 . **Absence of a found announcement is not proof none happened.** --- ## 3. The money comparison ### The private side, from the corpus - **59 deals, 51,950,967 tonnes contracted**, 2023-05-01 to 2026-08-04. - **Only 21 deals disclose a value, totalling $825.4M**, and those cover **2,318,205 tonnes — 4.5% of contracted tonnes.** - Disclosure is concentrated in one buyer: Frontier publishes a price on every deal; Microsoft, Google, Meta and JPMorgan publish none. Across the disclosing minority the derived average is **~$356/t** (my division, not a quoted figure). - The largest private contracts are all price-blind: Microsoft–BTG Pactual 8.0 Mt, Microsoft–Chestnut 7.0 Mt, Microsoft–AtmosClear 6.75 Mt, Microsoft–Vaulted Deep 4.9 Mt, Microsoft–CO280 3.685 Mt, Microsoft–Stockholm Exergi 3.33 Mt. ### The government side, converted The corpus records `monetary_scale_usd: null` for every procurement instrument, so **the following conversions are mine, at roughly 0.0975 USD/SEK and 0.155 USD/DKK, and move with FX**: | | Local | ≈ USD | Note | |---|---|---|---| | Sweden envelope 2026–2046 | SEK 36bn | ~$3.5bn | CDR-only, BECCS | | Sweden, Stockholm Exergi award | SEK 20bn/15yr | ~$2.0bn | ≈ $162/t over 12 Mt (my derivation) | | Denmark NECCS | DKK 2.5bn | ~$390M | ≈ $345/t over ~1.12 Mt (my derivation) | | Denmark CCS Fund | DKK 28.7bn | ~$4.5bn | **CDR-and-CCS scope**; first award is cement | | US DOE prize | — | $35M prize / $45M FY26 | genuine credit purchase | | Canada | CAD 10M | ~$7.3M | genuine credit purchase | | California SB 643 | — | ~$50M | vetoed | ### The comparison that matters **On committed dollars, government is already bigger — but almost none of it is credit buying.** Sweden's single Stockholm Exergi award (~$2.0bn) is roughly 2.4× the entire disclosed private spend in this corpus ($825.4M). Add Denmark's CDR-specific funds and the Nordic total is comfortably north of $2.5bn. But these are 15–20 year production subsidies to named facilities, priced per tonne captured, not purchases of certified credits in a market. Governments buying credits the way corporates do amounts to roughly **$35M (US) + C$10M (Canada) ≈ $42M** — about **5% of disclosed private spend, and about 0.2% of the Nordic subsidy commitments.** **On tonnes, private buying dominates.** 52 Mt contracted privately versus roughly 12 Mt (Sweden) + ~1.1 Mt (Denmark NECCS) + 8.6 Mt (Ørsted, partly descoped) on the government side — and the private figure excludes everything below the corpus's ≥50kt / ≥$10M threshold. **The honest bottom line on total private value: unknown.** Multiplying 52 Mt by the $356/t disclosed average gives ~$18.5bn, and that number is close to meaningless — the disclosers are Frontier's premium early-stage pathways (DAC at $650–989/t, ERW at $370–419/t) while the volume sits in cheap nature-based contracts. I flag the arithmetic only to show why I won't use it. ### The concentration problem underneath both columns Private demand is one buyer. CDR.fyi reported that as of April 13, 2026, Microsoft accounted for 36,439,157 tonnes, or 78.5% of total disclosed durable CDR tonnes contracted, with Frontier-linked buyers at 1,841,384 tonnes (4.0%) and all other buyers at 8,167,880 tonnes (17.5%) . Delivery is a different story: ex-Microsoft and Frontier buyers accounted for only 17% of contracted tonnes but 90% of delivered tonnes and 94% of retired tonnes . Volumes are still growing — CDR.fyi recorded 2.3 million tonnes contracted in Q1 2026, the largest quarter on record and about 560% of Q1 2025 volume, with Microsoft at 43% of contracted volume — and intermediaries facilitated 74% of contracted tonnes in Q1, the highest quarterly share in CDR.fyi's records . Note the CDR.fyi framing published following reports that Microsoft has paused its carbon removal purchases ; I could not verify that pause independently before exhausting my search budget, so treat it as reported context, not established fact. Government demand is similarly concentrated — in two Nordic countries, in one pathway (BECCS), and increasingly in facilities that also serve fossil CCS. --- ## 4. Caveats, disagreements and gaps - **Database vs. web disagreement:** the corpus lists Canada as "RFI issued Feb 2025; competitive award pending." Web sources show the actual tender launched 12 March 2026 with a five-stream standing-offer structure. The corpus is behind on this one; its policy-event freshness is 2026-07-17. - **Pending review:** 14 of 59 deals and 8 of 20 policy-landscape records are pending human review. The Microsoft–CREW Carbon (2026-08-04) and Microsoft–Alt Carbon (2026-06-11) deals are `review_status: pending` and carry unresolved reviewer questions on pathway and storage classification. - **Confidence flags:** the Danish CCS Fund and Swiss KlV records are `confidence: medium` with explicit unverified notes — the Aalborg award day, whether a second Danish tender is running, and whether Switzerland adopted the amendment are all unconfirmed. - **My derivations:** the $162/t Swedish and $345/t Danish figures, all FX conversions, and the $356/t disclosed private average are divisions I performed, not numbers any source quotes. - **Coverage gaps neither source closes:** the corpus is thin on subnational and non-OECD jurisdictions. Japan's GX-ETS entered its mandatory phase on 2026-04-01 but is a compliance instrument, not procurement. I found nothing on procurement in Norway (beyond Longship as infrastructure), the Netherlands' SDE++, Germany, South Korea, Singapore or the Gulf — that is a search gap on my part, not evidence of absence. My web-search quota was exhausted before I could check the EU Buyers' Club's December 2026 purchase target, the Canadian C$100M coalition, or anything after mid-August 2026. - **What no source gives you:** a clean government-vs-private ledger. The two sides are not measured in the same unit — governments mostly subsidise production per tonne captured over 15–20 years; corporates buy delivered credits. Anyone quoting a single "government spend" number is either summing announced envelopes that may never be drawn (the EU's 250 Mt) or ignoring that Denmark's biggest cheque went to a cement plant.
Judge’s note — Strong, expert-grade answer. It nails the conceptual distinction between production subsidy (Sweden BECCS reverse auction SEK 36bn envelope / ~SEK 20bn Stockholm Exergi award, Danish NECCS/CCUS/CCS funds) and true credit purchase-and-retire (DOE Purchase Pilot Prize with $45M FY26 appropriation and ~$30M realistic outlay, Germany's €11.5M certificate line, Canada's ≥CAD$10M, City of Stockholm 50 kt/yr), plus the EU ETS Article 9c 250 Mt proposal with sharp critiques (104% of EUA auction price, ~€25bn funding gap, no conditionality, biochar excluded). The private-side comparison is precise: 51,950,967 t contracted and US$825.4M disclosed across 21 of 59 deals covering 4.5% of tonnes — the required disclosure caveat is front-and-center and correctly attributed to Frontier-publishes/Microsoft-doesn't asymmetry. Missing element: no ~$2B Frontier funding-commitment figure, so the 'total private capital committed' side is understated relative to the checklist. Beyond-corpus contribution is modest: it cites a few primary government pages (Energimyndigheten, Danish Energy Agency, DOE) with budget figures and adds Danish CCS Fund June 2026 tender detail and the California SB 643 veto/veto-sustained sequence, but it explicitly declines to go past the 2026-08-04/07-17 dataset cutoff, so there are no post-2026-08-14 announcements and it leaves open questions (DOE winners named? Canada award?) unresolved rather than checking live sources. Coverage honesty is exemplary — inclusion thresholds, jurisdictional thinness (no Japan/Norway/Gulf procurement records; absence-of-record framing), currency-conversion caveats, pending-review records, and vendor-claim flagging on the CDR.fyi update. Structure and decision-usefulness are excellent, if slightly long.
I'll start by mapping the corpus coverage, then pull the procurement landscape and compare against private deal flow. Now let me pull the private buying side and check for procurement activity in jurisdictions not returned above. Let me get the total deal aggregate, check for procurement instruments in other jurisdictions, and look at recent signals. Two more checks: the US DOE purchase pilot's actual awards, and Germany/other demand-side instruments. ## Bottom line Government "procurement" of carbon removal splits into two very different things, and conflating them badly distorts the money comparison: 1. **Production subsidy dressed as procurement** — reverse auctions and contracts-for-difference where the state pays a project per tonne stored (Sweden, Denmark, the proposed UK model). This is where the *large* public money is, and it is measured in billions. 2. **Actual credit purchase and retirement by a government as a buyer** — where a public body buys tonnes the way Microsoft does. This is where public money is genuinely tiny today: tens of millions of dollars globally. Private buying still dominates the *purchase* market by an order of magnitude or more. But on a per-tonne-of-supply-financed basis, one Swedish auction rivals Microsoft's entire permanent-removal book. --- ## Where government money is actually moving today ### Sweden — the largest operational instrument by far The **Energimyndigheten BECCS reverse auction** carries a **SEK 36bn envelope for 2026–2046**, cleared EU state-aid approval in **July 2024**, and its first auction was won by **Stockholm Exergi at ~SEK 20bn over 15 years for ~800 kt/yr** ([Energimyndigheten](https://www.energimyndigheten.se/en/news/2025/20-billion-to-capture-and-store-over-11-million-tons-of-biogenic-carbon-dioxide/)). The agency's own headline frames that as **"over 11 million tons"** of biogenic CO₂. My arithmetic on those two published figures gives roughly **SEK 1,800/tonne (~US$180–190/t)** of state support. Treat that as a floor on project revenue, not a price: the same plant also sold **179,998 t to Frontier** (US$48.6M, announced 2024-06-18) and **3.33 Mt to Microsoft** (2024-05-06, later expanded ~1.75 Mt to 5.08 Mt total). Public and private money are **stacked on the same tonnes-producing asset**, not competing for it. ### Denmark — three successive funds, one genuinely CDR-only - **NECCS Fund (DKK 2.5bn, negative emissions only)**: contracts awarded **2024-04-17** to BioCirc CO2, Bioman and Carbon Capture Scotland — **160,350 t/yr combined, 2026–2032** (~1.1 Mt total) ([Danish Energy Agency](https://ens.dk/en/supply-and-consumption/ccs-tenders-and-other-funding-ccs-development)). - **CCUS Fund (~DKK 8bn)**: first tender awarded **2023-05-15** to Ørsted for BECCS at Asnæs/Avedøre, **430,000 t/yr from 2026**. - **CCS Fund (DKK 28.7bn)**: first tender concluded **June 2026** with **Aalborg Portland (ACCSION) contracted for up to DKK 16.5bn**, ~1.25 Mt CO₂/yr from 2030. **Note: that is cement, i.e. fossil CCS, not removal** — the fund is pathway-agnostic. Whether the remaining ~DKK 12bn goes to a second tender is unverified in this corpus. ### United States — small purchase, large adjacent subsidy The **DOE FECM Carbon Dioxide Removal Purchase Pilot Prize** is the only federal instrument here that actually buys tonnes: **24 Phase 2 semifinalists, up to 10 winners receiving purchase agreements of up to $3M each**, preserved with a **$45M FY26 appropriation** ([DOE](https://www.energy.gov/fecm/carbon-dioxide-removal-purchase-pilot-prize)). Maximum realistic outlay: ~$30M of purchases. Separately, DOE's **Regional DAC Hubs** funding was **restored around April 2026** after a federal court ruled the October 2025 mass grant terminations unconstitutional — but that is capital grant, not procurement. **California SB 643 is dead**: passed 37–0 in the Senate (2025-06-03) and 78–1 in the Assembly (2025-09-11), **vetoed by Governor Newsom on 2025-10-13** citing budget constraints, **veto sustained 2026-03-02**. Reported scale was ~$50M. ### Germany — first dedicated line, mostly R&D The **2026 federal budget** contains Germany's first dedicated negative-emissions line under BMUKN: **>€111M for 2026**, of which **~€98M is project funding and €11.5M is direct government purchase of CO₂-removal certificates**, plus **€320M in commitment authorisations through 2033**. Passed Bundestag 2025-11-28, Bundesrat 2025-12-19. **Caveat flagged in the record**: the €111M/€98M/€11.5M split was verified against the July 2025 *draft*, not line-by-line against the enacted budget. ### Subnational — the sleeper The **City of Stockholm** agreed to buy **50,000 t/yr for fifteen years** from Stockholm Exergi (announced 2026-05-26), which Stockholm Exergi claims makes the city **the world's fifth-largest buyer of carbon removals**. The corpus deliberately left the total tonnage null — 750,000 t is a derivation, not a published figure. No price disclosed. This is arguably the single most consequential *actual government purchase* on the books. --- ## Committed but not yet spending | Instrument | Status as of Aug 2026 | Scale | |---|---|---| | **Canada federal CDR procurement** (Greening Government Strategy) | Announced Oct 2024, RFI Feb 2025, **award still pending** | ≥CAD $10M by 2030 (~US$7M) | | **UK GGR Business Model** | **Proposed**; full docs published 2025-08-27; **no signed contract found** as of Aug 2026 | 15-yr CfD-style; **no strike price or budget published** | | **UK Power BECCS** | Dual-CfD confirmed Mar 2023; regs consultation closed 2025-06-09, **no response verified** | No strike prices published | | **EU CRCF Buyers' Club** | Operational as a convening body; model published May 2026 | **First permanent-removal purchases targeted by Dec 2026** | | **Switzerland KlV amendment** | Consultation closed 2026-02-12, **no adoption verified** | First roadmaps planned for the 2028–2032 legislature | The UK is the most striking gap between rhetoric and cash: the **2026-07-17 government response** to the Whitehead Independent Review reaffirmed the business model and committed to integrating GGRs into the UK ETS "as soon as practicable" — but Ince BECCS (Evero) is still only a "Priority" negotiation-list project and Climeworks' Silver Birch DACCS is "Standby." --- ## The big proposed thing that isn't money yet The **EU ETS Article 9c permanent removals procurement programme**, published **2026-07-17** as part of the ETS review, would increase the EU-wide allowance quantity by **250 million EUAs**, auction them, and have the Commission centrally purchase an equivalent volume of permanent removals from BioCCS and DACCS in the **2030s**, with a review by end-2034 ([cdr.fyi's explainer](https://www.cdr.fyi/blog/the-eu-ets-proposal-explained)). Three cautions the corpus is explicit about: - The funding mechanism supports an average removal price only about **104% of the average EUA auction price**. If removals cost €100 more than an EUA, buying 250 Mt needs roughly **€25bn of additional funding** that does not exist in the proposal. - The legal text **does not make allowance release conditional on removal contracting or delivery**, nor cancel unmatched allowances if procurement falls short. 250 Mt is a ceiling, not a financed commitment. - **Biochar is excluded** on evidentiary grounds despite CRCF recognition. Modelled portfolios are ~175–330 Mt BioCCS vs ~20–35 Mt DACCS. This is a legislative proposal entering ordinary legislative procedure, not procurement. --- ## How the money compares **Private side, from this corpus (59 deals, 2023-05-01 to 2026-08-04):** - **51,950,967 tonnes contracted** - **Only 21 deals disclose a value, totalling US$825.4M, covering 2,318,205 tonnes — 4.5% of contracted tonnes** That disclosure figure is the single most important caveat in this whole answer. **Frontier publishes a price on every deal; Microsoft publishes none.** So the $825M is essentially Frontier's cheque book plus Google–Holocene, and any per-tonne average derived from it describes the disclosing minority. Disclosed private prices range from **$100/t** (Google↔Holocene, 100,000 t, 2024-09-10 — lowest disclosed DAC price on record, expected to stack 45Q) and **<$200/t** (JPMorgan↔CO280, 450,000 t, 2025-05-20) up through Frontier's **$214–$989/t** band (CO280 $214; Planetary OAE $272; Phlair DAC $651; Heirloom $989). **Rough comparison of committed capital:** | Source | Committed | Nature | |---|---|---| | Sweden BECCS auction (first award) | ~SEK 20bn (~US$2bn) for ~11 Mt | State subsidy, 15 yrs | | Sweden total envelope | SEK 36bn (~US$3.5bn) to 2046 | State subsidy | | Denmark NECCS | DKK 2.5bn (~US$370M) for ~1.1 Mt | State subsidy | | Disclosed private deal value (this corpus) | US$825.4M for 2.3 Mt | Credit purchase | | Germany certificate purchase line | €11.5M (2026) | Credit purchase | | US DOE purchase prize | ≤US$30M | Credit purchase | | Canada federal | ≥CAD $10M by 2030 | Credit purchase | *USD conversions are mine at approximate 2025–26 rates; the underlying records store native currency only.* **Read it this way:** on **subsidy**, one mid-sized European country already outspends every disclosed private purchase in the corpus combined. On **purchase-and-retire**, all tracked governments together commit well under **$60M** — roughly one Frontier offtake, and a rounding error against Microsoft's undisclosed book (Vaulted Deep alone was up to 4.9 Mt over 12 years; third-party ~$1B estimates are explicitly flagged as speculation). For scale on the private side: Microsoft's durable deals in this corpus total roughly **24 Mt** (AtmosClear 6.75 Mt, Vaulted Deep 4.9 Mt, CO280 3.685 Mt, Stockholm Exergi 3.33 Mt, Ørsted 2.76 Mt + 1 Mt, and others), plus ~20 Mt of nature-based. Sweden's single auction, at ~11 Mt, is therefore roughly **half of Microsoft's entire permanent-removal book** — from one country, one instrument. --- ## Why this matters right now An uncorroborated **CDR.fyi market update (2026-06-12)** in the corpus claims **Microsoft accounted for 78.5% of all disclosed durable CDR tonnes purchased as of April 2026**, that its widely reported purchasing pause is "a meaningful market stress test," and that excluding Microsoft and Frontier, other buyers hold only **17% of contracted tonnes but 90% of delivered tonnes and 94% of retired tonnes** — total delivery rate rising from ~3% to ~13% once the two whales are stripped out. It also names the **July 2026 EU ETS review as the largest near-term watchpoint** and states policy "is not yet broad or fast enough to replace Microsoft-scale demand." Treat all of that as a source claim from a market-data vendor, not a verified fact. That is the structural picture: private demand is deep but dangerously concentrated in one buyer; government demand is broad in intent but, outside Scandinavian subsidy auctions, almost entirely pre-award. --- ## Gaps and uncertainty you should hold onto - **This corpus is a curated floor, not a census.** Non-Frontier deals need ≥50kt or ≥$10M to be included. CDR.fyi's public reports are the fuller transaction scoreboard. - **Policy coverage spans 14 jurisdictions** (US, EU, UK, Canada, Germany, Denmark, Sweden, Japan, Australia, Switzerland, Norway, Kenya, Finland-based registries, international schemes). **Subnational and non-OECD coverage is explicitly thin.** I found **no procurement record for Japan, Norway, South Korea, China, India or the Gulf** — that is absence of a record, not absence of activity. Japan appears only as the **GX-ETS**, whose mandatory phase began **2026-04-01** with a government-run allowance market due autumn 2027 — compliance architecture, not procurement. - **Freshness:** deals held through 2026-08-04, policy events through 2026-07-17. Anything after that is outside this dataset. - **Several statuses rest on absence of evidence**, and the records say so: the UK GGR model is marked "proposed" because no contract award could be found; the Swiss KlV sits at "consultation" because no adoption announcement was located; Denmark's Aalborg Portland award carries month-only precision. - **14 of 59 deal records are pending human review**, including several 2026 entries where the extractor flagged genuine ambiguity (e.g. whether Novo Nordisk is legally the buyer in the re.green agreement; whether ENGIE↔Deep Sky is firm procurement or a partnership). - **I could not verify** whether the DOE purchase prize has actually named winners and disbursed, nor whether Canada's competitive award has landed. Both were pre-award as of the last review.
| question | setup | facts | beyond | cites | honesty | overall |
|---|---|---|---|---|---|---|
| Who has bought bio-oil carbon removal, and a | Web search only | 3/4 | 7 | 4 | 5 | 7 |
| Who has bought bio-oil carbon removal, and a | Plugin connected | 4/4 | 7 | 4 | 5 | 9 |
| Who has bought bio-oil carbon removal, and a | Plugin + skill file | 4/4 | 10 | 4 | 5 | 9 |
| Who has bought bio-oil carbon removal, and a | Database only | 4/4 | 5 | 5 | 5 | 9 |
| How much money has Frontier actually committ | Web search only | 5/5 | 6 | 3 | 5 | 8 |
| How much money has Frontier actually committ | Plugin connected | 5/5 | 8 | 4 | 5 | 9 |
| How much money has Frontier actually committ | Plugin + skill file | 5/5 | 6 | 5 | 5 | 9 |
| How much money has Frontier actually committ | Database only | 2/5 | 2 | 4 | 5 | 5 |
| Where does EU carbon removal policy stand? | Web search only | 2/3 | 0 | 4 | 5 | 8 |
| Where does EU carbon removal policy stand? | Plugin connected | 3/3 | 6 | 5 | 5 | 9 |
| Where does EU carbon removal policy stand? | Plugin + skill file | 3/3 | 5 | 5 | 5 | 8 |
| Where does EU carbon removal policy stand? | Database only | 3/3 | 0 | 5 | 5 | 8 |
| What changed in CDR policy since March? | Web search only | 1/3 | 0 | 4 | 5 | 5 |
| What changed in CDR policy since March? | Plugin connected | 3/3 | 0 | 4 | 5 | 7 |
| What changed in CDR policy since March? | Plugin + skill file | 3/3 | 3 | 5 | 5 | 9 |
| What changed in CDR policy since March? | Database only | 3/3 | 0 | 4 | 5 | 7 |
| How much ocean alkalinity science is real-wo | Web search only | 0/3 | 8 | 4 | 5 | 6 |
| How much ocean alkalinity science is real-wo | Plugin connected | 3/3 | 3 | 5 | 5 | 9 |
| How much ocean alkalinity science is real-wo | Plugin + skill file | 3/3 | 3 | 5 | 5 | 9 |
| How much ocean alkalinity science is real-wo | Database only | 3/3 | 1 | 5 | 5 | 9 |
| Can enhanced rock weathering be measured — a | Web search only | 2/3 | 11 | 3 | 5 | 7 |
| Can enhanced rock weathering be measured — a | Plugin connected | 2/3 | 14 | 5 | 5 | 9 |
| Can enhanced rock weathering be measured — a | Plugin + skill file | 3/3 | 10 | 5 | 5 | 9 |
| Can enhanced rock weathering be measured — a | Database only | 3/3 | 8 | 5 | 5 | 8 |
| Due diligence on Vaulted Deep | Web search only | 3/3 | 10 | 4 | 5 | 9 |
| Due diligence on Vaulted Deep | Plugin connected | 3/3 | 8 | 5 | 5 | 9 |
| Due diligence on Vaulted Deep | Plugin + skill file | 3/3 | 8 | 4 | 5 | 9 |
| Due diligence on Vaulted Deep | Database only | 3/3 | 1 | 4 | 5 | 8 |
| Government buying vs private buying — where | Web search only | 3/3 | 18 | 5 | 5 | 8 |
| Government buying vs private buying — where | Plugin connected | 3/3 | 14 | 5 | 5 | 9 |
| Government buying vs private buying — where | Plugin + skill file | 3/3 | 8 | 4 | 5 | 8 |
| Government buying vs private buying — where | Database only | 3/3 | 3 | 4 | 5 | 8 |
You are a careful analyst covering the carbon dioxide removal (CDR) industry. Answer thoroughly with specific figures, dates and named sources. Cite sources for factual claims and be explicit about uncertainty and gaps.
You are a careful analyst covering the carbon dioxide removal (CDR) industry. Answer thoroughly with specific figures, dates and named sources. Cite sources for factual claims and be explicit about uncertainty and gaps. You have the Carbon Curve Intelligence tool connected. Follow this procedure: # Carbon Curve research Carbon Curve Intelligence is a curated database of CDR deals, policy instruments with event timelines, research papers and publications. Every record is cited and carries its own provenance. It is a **floor, not a ceiling.** It is deep on 14 jurisdictions and thin elsewhere, and it holds no record of anything its weekly review has not yet reached. Answering from it alone produces a confident answer with holes in it. ## The procedure Follow all four steps. Step 3 is the one that gets skipped, and skipping it is the failure this skill exists to prevent. 1. **Call `describe_coverage` once** at the start of a CDR session, unless you already did. It tells you the pathway taxonomy, which jurisdictions are tracked, and how much of the corpus is price-disclosed. 2. **Call the relevant tool** for the question: `search_records` for deals, entities, papers and publications; `get_policy_landscape` for what exists in a jurisdiction; `get_recent_changes` for what changed in a window; `get_evidence_profile` for how strong a literature is; `get_entity` to join everything about one organisation. 3. **Then search the web for what the response says it does not cover.** Every response, from every tool, carries a `boundary` field with three parts: the freshness of each corpus layer (deals, policy events, papers, media), `not_covered` naming what the corpus structurally cannot contain for this query, and `suggested_searches` with ready-formed queries. Run the searches. This applies to every kind of question, not just policy: a science question must check for literature published after the paper-fetch date, a deal question for announcements since the last ingest, an entity question for the organisation's activity outside the corpus. If a response says Canada and the United Kingdom returned no events for your window, that means no record, not no activity. Go and find out. Do this even when the database answer already looks complete. It will look complete: that is the trap. A tool result that answers the question makes searching feel unnecessary, which is exactly when the gaps are invisible. 4. **Answer from both, and say which is which.** Give the database records with their citations, give what you found by searching, and note where the two disagree. Close with what neither covers. ## What each source is good for **The database** is authoritative on: enumeration (all four offtakes in a pathway, not the two that made the news), computation across a set (derived prices, study counts, trends), state over time (dated policy events), and knowing its own gaps. **Web search** is better for: very recent events, jurisdictions outside the tracked set, the analytical framing and mechanism detail behind a policy, and anything the corpus has no category for. Neither is complete. The answer is both. ## By question type - **Deals and market**: `search_records` with `record_type: "deal"` plus pathway, jurisdiction or date_range. Read the `deal-aggregate` record for totals across the whole filter; the list itself may be truncated and the coverage note will say so. Never sum the visible rows yourself. - **Policy**: `get_policy_landscape` for current state, `get_recent_changes` for deltas, and the `upcoming-events` record for consultations and deadlines still open — the actionable part. - **Science**: `get_evidence_profile` first; it now returns the field-trial list, the deals, and the policy instruments for the pathway in one call. Then `search_records` with `record_type: "paper"` and `study_type` to pull specific studies. Check `boundary.freshness` for the paper-fetch date. - **Diligence on one organisation**: `get_entity`, then run the suggested searches — funding, hiring and delivery track record live outside this corpus, and the database will say so rather than pretend otherwise. ## Rules - Never present database records as a complete picture of a topic. Say what was searched and what was covered. - Prices are derived by division from public totals unless a source quotes one. Only 21 of 59 deals disclose a value, and those are 4.5% of contracted tonnes, so any price trend describes the disclosing minority rather than the market. - Records marked `review_status: pending` have not been through human review. Say so when you cite them. - Signals are uncorroborated claims kept deliberately apart from records. Never present a signal as a verified fact. - The database offers no ratings, rankings or recommendations by design. Do not invent them on its behalf. - An empty result means no record, never no activity.
{
"type": "object",
"additionalProperties": false,
"required": [
"gold_findings",
"beyond_corpus_count",
"citation_quality",
"coverage_honesty",
"clarity",
"overall",
"rationale"
],
"properties": {
"gold_findings": {
"type": "array",
"items": {
"type": "object",
"additionalProperties": false,
"required": [
"item",
"found"
],
"properties": {
"item": {
"type": "string"
},
"found": {
"type": "boolean"
}
}
}
},
"beyond_corpus_count": {
"type": "integer",
"description": "Distinct substantive findings sourced from live web/search beyond the described database corpus (post-2026-08-14 items, delivery data, funding rounds, etc.)"
},
"citation_quality": {
"type": "integer",
"description": "0-5: are claims traceable to named, checkable sources?"
},
"coverage_honesty": {
"type": "integer",
"description": "0-5: does the answer state what it does NOT know / where data is missing?"
},
"clarity": {
"type": "integer",
"description": "0-5: structure, directness, decision-usefulness for an expert reader"
},
"overall": {
"type": "integer",
"description": "0-10 holistic quality for an expert CDR practitioner"
},
"rationale": {
"type": "string"
}
}
}