Climeworks re-engineered its solid-sorbent chemistry so filters absorb CO2 roughly twice as fast, claiming its Gen-3 technology can halve energy use and cost toward $250-350/ton captured by 2030, first at the Project Cypress hub. It attacks the biggest cost driver — the heat and electricity to regenerate sorbent — but the company has already walked its target down from an earlier ~$100/ton ambition, and lab gains have not yet shown up as delivered tonnes.

1PointFive's STRATOS in Texas uses Carbon Engineering's aqueous-solvent process at up to 500,000 tonnes/yr — an order of magnitude above Mammoth — betting that sheer scale and oilfield-style engineering drive down unit cost. Scale helps, but the $1.3B price tag and a phased ramp show how much capital must be sunk before the cost curve bends, and liquid-solvent systems are energy-hungry in their own way.

Heirloom accelerates natural limestone weathering to cut thermal energy, while enhanced-weathering and ocean-alkalinity players (Mati, Lithos, Equatic, Ebb Carbon) sidestep DAC's energy problem entirely by using rock and seawater chemistry that is far cheaper per tonne on paper. The trade is measurement: these low-cost pathways are hard to verify to the same standard, moving the bottleneck from cost to MRV credibility.

Isometric lists only durable credits and publishes the data, methodologies, and verification behind each one; Puro.earth and others certify engineered pathways. Radical transparency lets buyers inspect the evidence trail, but a registry's authority rests on never listing a bad project, and standardizing MRV across novel pathways that each behave differently in the field remains unfinished.

Sylvera and BeZero rate credit quality using satellite, lidar, radar, and field data to separate high- from low-integrity credits. Ratings give buyers a defensible screen, but raters disagree with one another, buyers resent paying to be told their inventory is junk, and a rating is a probability, not a guarantee — the credit can still reverse.

The EU's Carbon Removal Certification Framework adopted its first methodologies for DAC, BioCCS, and biochar in February 2026, with a unified EU registry due by 2028, creating a government-backed integrity standard. A public methodology raises the floor and gives buyers legal cover, but rulemaking is slow, covers only a few pathways so far, and a certified-but-flawed method can enshrine bad credits with a stamp of approval.

Frontier's ~$1B AMC (2022-2030) pools committed demand from Stripe, Alphabet, Shopify, McKinsey, and others, and has facilitated over $700M in offtakes to pull first-of-a-kind plants into existence. It is the clearest tool for de-risking supply, but it is still a fixed pot of voluntary money aimed at a trajectory, not a self-sustaining market, and it expires in 2030.

The EU's CRCF and discussions of integrating removals into compliance regimes could eventually force regulated buyers into the market, turning voluntary demand into mandatory demand. Compliance demand would dwarf today's voluntary market, but it is years away, politically contested, and the U.S. is moving the other direction — eroding 45Q and cutting industrial subsidies rather than mandating purchases.

The long-run fix is making removal cheap enough that many buyers participate without philanthropy — enhanced weathering and biochar already clear in the low hundreds of dollars per tonne, far below DAC. Cheaper credits broaden the buyer base, but the cheapest pathways are also the hardest to verify, so cost-down and MRV credibility are locked together: solve one and you strain the other.

LanzaJet runs the world's first ethanol-to-jet plant in Georgia and licenses the process globally, converting low-carbon and waste ethanol into drop-in SAF at commercial scale today. It is one of the few pathways actually producing, but ethanol supply is finite and land-linked, and the lifecycle carbon benefit depends heavily on how the ethanol was made.

Twelve and Infinium make eSAF from captured CO2, water, and renewable power — a feedstock that is effectively unlimited if clean electricity is. The abatement is deep (up to ~90%), but it is the most expensive route, gated by the availability of cheap renewables and green hydrogen, and scaling from a first plant to tens of millions of gallons is unproven and capital-hungry.

EU ReFuelEU and the UK mandate force escalating SAF blending, and airlines like Alaska and United sign multi-year offtakes and take equity to underwrite first plants. Guaranteed demand is what makes plants financeable, but mandates create demand without economics — a delay or a credit cut (as with the trimmed U.S. SAF blender's credit) can strand a plant that took years to build.

Sublime makes cement electrochemically at room temperature, Brimstone makes standard portland cement from carbonate-free calcium silicate rock, and Boston Metal reduces iron ore with molten-oxide electrolysis instead of coal. Redesigning the reaction eliminates process emissions at the source, but scaling a demo cell to a full mill is where the capital dies — Sublime paused its first plant after losing a DOE grant and Boston Metal cut jobs after an equipment failure.

CarbonCure and Fortera inject or mineralize CO2 into concrete and cement, locking carbon into the material while displacing a share of portland cement. Because it bolts onto existing plants and codes, it deploys faster and cheaper than clean-sheet routes, but it abates a fraction of emissions rather than eliminating them, so it is a bridge, not a solution.

Stegra is starting up the first large-scale green-steel plant in Sweden using green hydrogen DRI, financed by ~€6.5B and long-dated offtakes, while hyperscaler and building-materials offtakes (Microsoft's Sublime deal; CRH and Holcim's Sublime stakes) underwrite demand. Anchor buyers make first plants financeable, but hydrogen steel needs vast cheap green hydrogen, and without regulation most commodity buyers will not pay a durable green premium.

The durable fix is being cheap enough not to need the subsidy — the explicit goal behind Climeworks' cost-down, Brimstone's parity-with-incumbents pitch, and enhanced weathering's low per-tonne cost. Parity insulates a business from policy, but almost no hard-tech decarbonization is there yet, and the plants that would prove out the cost curve are exactly the ones that need subsidy to get built.

Companies are leaning on the EU's Carbon Removal Certification Framework and European mandates, and on non-U.S. markets, as Washington pulls back. Geographic diversification hedges U.S. policy risk, but Europe's frameworks are newer and slower, and moving projects abroad forfeits the domestic-content and energy-cost advantages that made some U.S. plants pencil in the first place.

Binding corporate offtakes — Microsoft's Sublime and Vaulted deals, airline SAF contracts, Frontier's AMC — substitute private demand for public support, financing plants on purchase commitments rather than tax credits. Private offtakes are more durable than a legislative line item, but they concentrate the whole industry's fate on a few voluntary buyers, converting policy risk into demand-concentration risk.