What carbon capture at scale means

Two different things share the name. Point-source capture takes CO2 from a flue gas where it is concentrated, which is thermodynamically easy and avoids emissions rather than removing them. Direct air capture takes CO2 from ambient air at roughly 420 parts per million, which is thermodynamically expensive and genuinely removes it. This part concerns the second, because only removal addresses carbon already emitted — and the dilution is precisely what makes it hard.

Where the science stands

Established DAC works and operates commercially. The largest operating plant, in Iceland, switched on in 2024 with a design capacity of 36,000 tonnes a year, using solid sorbents and geothermal power with geological storage. Against roughly 37 gigatonnes of annual emissions, that design capacity is about one millionth.

Established Costs are far above the viability threshold. The US Department of Energy's target for climate-scale deployment is $100 per tonne. The leading operator's co-chief executive described its own operating cost as closer to $1,000 per tonne than $100; a major bank paid roughly $800 per tonne for removal credits in 2023. An independent analysis projects $230–540 per tonne by 2050 — well above the $100–300 figure often quoted.

Frontier Scale is nonetheless arriving. A 500,000-tonne-per-year facility in West Texas, using liquid solvents, would at design capacity increase global DAC capacity by several hundred per cent from one plant. Another developer using limestone-based capture is building two facilities totalling 320,000 tonnes per year and targeting costs near $100. The US allocated $3.5 billion to DAC hubs, and the 45Q tax credit underwrites much of the economics.

Established Even the largest planned facility is roughly one seventy-thousandth of annual emissions, and a fraction of the IEA's 2030 target — which is itself a fraction of what mid-century pathways assume. Frontier The sector's difficulties are also real: the largest solid-sorbent operator laid off at least 10% of its roughly 500 staff in 2025 despite having raised over $800 million, and has faced criticism over delivery timelines against subscriptions already sold.

Frontier A structural critique deserves stating: because DAC provides a mechanism for continued emissions to be nominally offset, it can function as a reason to delay abatement. That is an argument about incentives rather than about the technology, and it is made seriously by people who understand the technology. Handwave Claims that DAC can substitute for emissions reduction are not supported by any credible pathway.

The binding bottleneck

Established The constraint is energy per tonne, and it is thermodynamic before it is economic: separating a dilute gas costs work, and no engineering removes that floor. Frontier Second is cost, which follows from the first and from capital intensity. Frontier Third is clean energy supply — DAC powered by fossil electricity can approach or exceed break-even in emissions terms, so removal competes for the same clean power everything else in this programme needs. Frontier Fourth is storage: permanent geological sequestration must be verified over centuries.

Ethics and governance

Established Verification and permanence are the core governance problems, and the credit market currently prices them imperfectly. Frontier Moral hazard is the substantive ethical objection and should not be dismissed as rhetoric. Established Consumer-facing subscriptions that sell removal delivered years later raise a straightforward disclosure question — buyers have discovered that CO2 they paid to remove had not been removed. Frontier That much of the capital comes from oil and gas subsidiaries is a fact readers should weigh for themselves rather than a disqualification.

Timelines

  • 10 yr: Established megatonne-scale capacity plausible; costs falling but likely still several hundred dollars per tonne.
  • 25 yr: Frontier tens of megatonnes annually if policy support persists — still under a thousandth of current emissions.
  • 50 yr: Speculative gigatonne-scale removal, which every mid-century pathway assumes and no current trajectory reaches.
  • 100 / 250+ yr: Speculative drawdown below present concentrations, requiring energy at a scale that returns this to Part 1.