Direct Air Capture Economics in 2026: Stripe Frontier's $1B AMC, Climeworks vs Heirloom, Project Bison, and the Path to $100/Ton
Stripe Frontier's billion-dollar Advance Market Commitment, Climeworks vs Heirloom on storage durability, the path from 1000 USD/ton to 100 USD/ton, Project Bison in Wyoming, Battelle, and the MRV verification layer.
The most important number in carbon removal in 2026 is the cost per tonne of CO2 captured and durably stored by direct air capture. The industry is publicly hovering between 500 and 1000 USD per tonne on operating plants, with credible roadmaps to 200-300 USD by 2030 and a target of 100 USD by sometime in the 2030s. Whether that trajectory holds is the question that decides whether DAC scales to climate-relevance — gigatonnes per year by mid-century — or stays a niche compliance tool.
This post is the engineer-friendly walk through the actual economics: where the money goes, why the published numbers look the way they do, and what has to change for the cost curve to bend further.
What 600-1000 USD per tonne actually buys#
The headline cost per tonne for a DAC plant is the sum of several distinct line items, each with its own scaling characteristics.
Capital cost amortization. Building a DAC facility is expensive. Climeworks’s Orca cost roughly 10-15 million USD per thousand-tonne nameplate; Mammoth’s cost was not publicly broken out but is in the hundreds of millions for 36,000-tonne nameplate. Capital amortized over a 20-year plant life is a substantial share of the per-tonne cost. Larger plants amortize fixed engineering and permitting costs over more tonnes, which is one of the cleanest paths to lower costs.
Energy. DAC is energy-intensive. The thermodynamics of pulling CO2 from 420 ppm in air to a high-purity stream cost meaningful energy regardless of the chemistry. Climeworks’s Mammoth uses Iceland’s geothermal heat, which is unusually cheap and clean. Plants in less geologically blessed locations have to either pay for grid electricity (with all the carbon-intensity questions that opens) or co-locate with dedicated renewables. The 1PointFive Stratos plant in Texas is partially powered by solar; Heirloom’s plants in Tracy and Louisiana use grid electricity with renewable energy procurement.
Sorbent cost. The chemicals that bind CO2 — amine sorbents for Climeworks, calcium oxide for Heirloom, hydroxide solutions for Carbon Engineering — have unit costs that look small per kilogram but scale with throughput. Sorbent degradation and replacement is a real operating cost line item. The 2024-2025 research push on longer-lived sorbents (Verdox, Mosaic Materials, several academic groups) targets this.
Storage and injection. Geological sequestration via wells (Climeworks/Carbfix in Iceland, 1PointFive in Texas, Project Bison in Wyoming) has its own capex and operating cost. Mineralization in concrete (Heirloom plus CarbonCure) substitutes the well for a different infrastructure entirely.
MRV and verification. Measurement, reporting, and verification by third parties is non-trivial overhead that scales modestly with project size. The growing standardization through Puro Earth, Isometric, the CDR.fyi index, and the emerging ICVCM core carbon principles is starting to bring this cost down.
The honest summary: roughly half of the published cost is energy and operations, roughly a third is capital amortization, and the rest is sorbent, storage, and overhead. Each of those scales differently with plant size and runtime.
Stripe Frontier and the AMC mechanism#
The Frontier Fund — initially launched in April 2022 with around 925 million USD from Stripe, Alphabet, Shopify, Meta, McKinsey, and the next several years adding more partners — is the most consequential single demand-side intervention in the DAC market. Through 2024-2025 the cumulative committed AMC value crossed 1 billion USD, with offtake agreements with most major DAC operators.
The AMC structure is borrowed from vaccine markets — buyers commit to purchase a future quantity at a known price, which lets operators raise capital and scale operations without depending entirely on grants or speculative voluntary markets. Frontier’s purchases through 2024-2025 have included Climeworks, Heirloom, Vesta, Carbon Engineering / 1PointFive, Lithos Carbon, Charm Industrial, and several smaller players. The published per-tonne prices have varied — from several hundred USD at the high end to meaningfully lower for less expensive storage modalities — and the structure is becoming a template that other corporate offtakers (Microsoft, JPMorgan, Boston Consulting Group, Salesforce) are using.
The economics matter because the alternative — selling into the voluntary carbon market at the prices that have historically prevailed there — does not pay for DAC. Voluntary credits from forestry projects have traded in the 5-30 USD per tonne range. DAC needed a separate demand channel, and Frontier-style AMCs created one.
Climeworks versus Heirloom on storage durability#
The storage durability question is one of the cleaner technology comparisons in the industry.
Climeworks plus Carbfix. CO2 is dissolved in water and injected into porous basalt rock formations in Iceland. The basalt minerals react with the dissolved CO2 to form stable carbonate minerals — calcite, magnesite, dolomite — over months to a few years. The result is geologically stable on million-year timescales. The Carbfix monitoring program has measured the mineralization fraction directly via isotope tracking and is the cleanest verifiable storage story in the industry.
Heirloom plus concrete. CO2 is mineralized into the calcium carbonate matrix of concrete during mixing. The result is also durable — the embedded CO2 stays in the concrete for the structural life of the building, which is decades to centuries, and even after demolition the carbonate continues to mineralize stably. The independent verification is less mature than the basalt mineralization story but is straightforward chemistry.
The trade-off is not technical durability but operational scalability. Basalt formations are geographically limited; concrete pours happen everywhere. Climeworks’s expansion outside Iceland needs different storage partners (the Project Cypress hub in Louisiana, etc.); Heirloom’s expansion follows the concrete supply chain.
Project Bison and the Wyoming geology#
Project Bison — the CarbonCapture Inc. plant planned for Wyoming, with Frontier and other AMC offtake — is the largest single planned US DAC project after Stratos. Initial phases target tens of thousands of tonnes per year, with a long-term plan targeting megatonne-scale by 2030. The Wyoming siting takes advantage of the geological storage potential of the state’s saline aquifer formations and the favourable regulatory environment for Class VI injection wells.
The 2024-2025 update on Bison has been mixed. The original CarbonCapture Inc. modular DAC design — many small standardized units rather than one large plant — has gone through redesign. Some published timeline slippages occurred. The basic geological case is intact and the project remains in the active development pipeline; the exact size and commissioning date are softer than they were at announcement.
Battelle and the MRV verification layer#
Battelle is the operator of the Project Cypress DOE DAC Hub in Louisiana and the broader Battelle work on geological storage verification. The MRV layer — what is actually injected, where it is, what fraction is permanently mineralized or trapped — is the part of the DAC story that most outsiders underestimate. Without rigorous MRV, the credits being sold are not credibly different from older voluntary-market offsets that have lost regulatory and corporate credibility.
The 2024-2025 work on MRV standardization has been one of the more important industry developments. Puro Earth’s CO2 Removal Credits, the Isometric registry, the CDR.fyi public database, and the ICVCM Core Carbon Principles are converging on a standard for what data has to be published, how injections have to be monitored, and what verification a third party has to perform. The 2026 expectation is that DAC credits will be the most rigorously MRV’d carbon credits available and will command price premiums in any corporate procurement that takes net-zero claims seriously.
The path from 1000 USD to 100 USD per tonne#
The roadmap that DAC operators publish for the cost trajectory generally rests on the following levers, in rough order of impact.
Plant scale. Going from 1000-tonne plants to 100,000-tonne plants amortizes capital and engineering costs over much more output. Most credible operator roadmaps put the bulk of the cost reduction here.
Sorbent improvement. Longer-lived sorbents with lower regeneration energy reduce both operating cost and capex (smaller equipment for the same throughput). The active research is large; the commercial readiness is partial.
Energy cost reduction. Co-located cheap renewables — solar in West Texas, geothermal in Iceland, offshore wind in the North Sea, hydro in Quebec — make a meaningful difference. The capacity to procure renewable energy at scale is the rate-limit on some growth.
Learning curve. Wright’s Law applies — every doubling of cumulative installed capacity tends to drop costs by 10-25 percent for industrial processes that go through systematic learning. The DAC industry is at the early stage of the learning curve; the doublings ahead are large.
Integration. Co-locating capture with the storage well, integrating with low-cost renewables on-site, integrating with industrial heat sources — all of these compound modestly into the overall trajectory.
The honest summary: 100 USD per tonne by 2030 is aggressive but plausible for the most favourable sites; mid-century gigatonne-scale at meaningfully lower cost requires the full learning-curve playout and the policy support to underwrite the early-stage capacity build.
What this means for technology teams#
For data and platform teams supporting DAC operators or corporate buyers in 2026, the durable investment is the MRV and reporting infrastructure — telemetry from capture plants, structured injection and mineralization data, audit trails for third-party verifiers, integration with corporate sustainability and finance reporting. We have helped operators stand up the data engineering foundations for this kind of platform; the pattern is recognizable to anyone who has built industrial telemetry and reporting pipelines.
The financial commitments to DAC through Frontier-style AMCs are large and growing. The technical and operational infrastructure that makes them auditable is what turns those commitments into actual durable carbon removal.
Related reading#
- Carbon capture tech in 2026
- Green hydrogen production in 2026
- Climate tech carbon accounting in 2026
DAC economics are bending the right way but the path needs both capital and rigorous MRV. If your organization is sizing the data infrastructure for a DAC project or corporate offtake program, our data engineering team has built this kind of pipeline. Tell us about the program.