Climeworks Cuts Mammoth DAC Operating Costs by Over 50%

Climeworks has more than doubled CO2 capture performance in upgraded collector containers at Mammoth, its direct air capture plant in Iceland. Operating cost per tonne fell by more than 50%. The gains came from 18 months of running a large plant in real conditions. Climeworks announced the results on September 10, 2026.

>> In Other News: [x](x)

What Did Climeworks Report at Mammoth?

Climeworks, the Zurich-based carbon removal company founded in 2009, reported two linked results at Mammoth. Upgrades to its sorbent material and plant design more than doubled CO2 capture performance in upgraded collector containers. Mammoth is the company's second commercial plant in Iceland and began operating in 2024.

Precision matters with these figures. The doubling applies to upgraded collector containers, not to Mammoth's total plant output. Climeworks ran two upgraded containers for more than six months before reporting the result.

The cost figure is equally specific. Operating cost per tonne fell by more than 50% for the upgraded systems. Climeworks did not disclose the underlying dollar amount.

How Do Sorbents Drive Direct Air Capture Costs?

Each collector container holds a sorbent, the filter material that chemically binds CO2 from passing air. Once the sorbent saturates, the container closes and heats up to release concentrated CO2 for permanent storage. The cycle then repeats.

Sorbent behaviour drives DAC economics more than any other single factor. Capture capacity per cycle, durability across thousands of cycles, and regeneration energy all feed straight into cost per tonne. Faster sorbent testing has been a priority for Climeworks for several years.

The company refined both the sorbent formulation and its production process. It also upgraded equipment and changed how the plant is operated. Both streams of work contributed, so this was not a case of operators simply getting better with practice.

How Much Carbon Is Mammoth Actually Removing?

Mammoth delivered 675 tonnes of net carbon removal in the first half of 2026. That compares with approximately 119 tonnes in the same period of 2025, a near sixfold increase (Climeworks via Bloomberg, September 2026).

Mammoth's design capacity is up to 36,000 tonnes of CO2 per year. At the first-half 2026 pace, annualised output lands near 1,350 tonnes. The plant is still early in its ramp-up, which is exactly why its operating data carries value.

Four different measurements appear in DAC coverage, and they are easy to confuse:

• Design capacity: what the plant was built to capture at full operation, up to 36,000 tonnes per year at Mammoth.
• Gross capture: total CO2 pulled from the air before energy and process emissions are subtracted.
• Collector performance: CO2 captured per upgraded container under real operating conditions.
• Net carbon removal: verified removals delivered to buyers after all associated emissions are deducted.

The September announcement covers collector performance and operating cost. Net carbon removal, at 675 tonnes for the first half of 2026, is the figure buyers contract against.

Metric	H1 2025	H1 2026	Design or Target
Net CO2 removal (tonnes)	Approximately 119	675	Up to 36,000 per year
Operating cost per tonne	Baseline	Down more than 50%	$400 to $600 by 2030 (Gen 3)
Upgraded collector performance	Below design run rate	More than doubled, at design run rate	72 containers installed at Mammoth

Carbon removal agreements often run a decade or longer. Buyers weigh a developer's operating record as heavily as its headline capacity, which is why demonstrated performance gains carry commercial weight.

>> RELATED: Project Cypress Gets Green Light: DOE Clears $600M Louisiana DAC Hub

Why Operating a Large Plant Speeds Up Improvement

Large plants surface problems that pilots never show. Equipment wear rates, sorbent degradation across dozens of containers, energy losses and airflow behaviour all shift at commercial scale.

Climeworks feeds that field data straight into development. Its engineers process close to 200 million data points daily across operating facilities (Climeworks, 2024). An innovation centre runs alongside the commercial plants, so research and operations inform each other quickly.

The loop itself is straightforward. Build, operate, measure, redesign the sorbent and hardware, then deploy the improved version. Scaling DAC from pilot to commercial operation is where most of the learning happens.

Climeworks plans to fit the upgraded sorbent and mechanical changes across all 12 collector containers in one Mammoth module by the end of 2026. Next-generation technology testing at Mammoth is planned for early 2027. In lab testing, that technology has shown a tenfold improvement in sorbent lifetime and a fourfold improvement in process densification (Climeworks, September 2026).

How Does Climeworks Compare With Other DAC Developers?

1PointFive, a subsidiary of Occidental Petroleum, is building Stratos in Ector County, Texas. The facility is designed to capture up to 500,000 metric tonnes of CO2 per year using a liquid-solvent process. Phase 1 commissioning hit a delay in 2026 tied to non-process components.

Heirloom Carbon Technologies takes a mineral-based approach and is a partner in the DOE-backed Project Cypress hub in Louisiana. Its first phase there targets 100,000 metric tonnes per year. Several developers are now pursuing different technical routes to the same outcome.

Cost is the shared constraint across every approach. Climeworks has confirmed that Mammoth runs closer to $1,000 per tonne than to the $100 per tonne long-term benchmark. Its Generation 3 technology targets $400 to $600 per tonne by 2030.

What to Watch Next at Mammoth

Three checkpoints will show whether the September results hold. The first is whether upgraded containers sustain output well past the six-month run already logged.

The second is the full module upgrade due by the end of 2026. Twelve containers running in parallel is a different engineering problem than two.

The third is Project Cypress in Louisiana, targeted for operations by the end of 2027 using Generation 3 technology. Whether net removal there starts closer to design capacity than Mammoth did will say a great deal.

Mammoth's numbers carry both halves of the story. Net removal climbed from approximately 119 tonnes to 675 tonnes year over year, operating cost per tonne dropped by more than 50%, and upgraded collectors more than doubled their capture performance.

For ongoing coverage of carbon removal, BECCS, and corporate CDR procurement, subscribe to Decarbonfuse.com.