The researchers tested a lab-scale device made of nine cells stacked together, each with an area of about half a letter-sized sheet of paper. A blower pushed air through sinuous air channels in the plates sandwiching each cell together (using components borrowed from fuel cells).

This small device used electricity at a rate of about 0.8 megawatt-hours per ton of captured CO₂. Current facilities capturing carbon from ambient air are in the neighborhood of 1.5 to 3 megawatt-hours per ton of CO₂.

Piloting the process

Several of the researchers on the team are part of a startup called RepAir Carbon that is based on this technology, and part of the paper is spent on sketching out financial feasibility at scale.

Estimating the cost for an initial small pilot plant (about $566 per ton of captured CO₂), they use the learning rate from the lithium-ion battery industry and some common cost scaling for bigger plants to project forward a couple of pilot generations. For a plant with a thousand times the capacity of their pilot, they estimate they could get to $92 per ton of captured CO₂.

That would be much cheaper than has so far been achieved by companies like Climeworks, which is aiming to get down to $250–$350 per ton by 2030—and has demonstrated how difficult it can be to meet optimistic scaling projections.

Targeting $100 per ton has long been the goal in the carbon-capture world, as costs that low could spur much wider adoption. Adding to the number of technologies being pursued may improve the odds that one of them reaches that goal.

Nature Energy, 2026. DOI: 10.1038/s41560-026-02129-z (About DOIs).