Direct Air Capture (DAC) — Pulling CO₂ from the Atmosphere, Technology, Costs & the Path to Gigaton Scale
Direct Air Capture (DAC) extracts CO₂ directly from the ambient atmosphere — at ~420 ppm concentration — using engineered chemical systems. Unlike point-source CCS which captures CO₂ at emission sites (power plants, factories), DAC can be located anywhere and can reverse historical emissions, not just prevent new ones. In every IPCC 1.5°C scenario, some DAC or other Carbon Dioxide Removal (CDR) is required to offset residual emissions from agriculture, aviation, and other hard-to-abate sectors. The two leading technologies are Liquid Solvent DAC (Carbon Engineering / Oxy, now Stratos in Texas) and Solid Sorbent DAC (Climeworks, now operating Mammoth in Iceland). Current cost: $400–1,000/t CO₂. Cost targets: $100–200/t by 2030, $50–100/t by 2040 with learning curves and scale. The Inflation Reduction Act's $180/t 45Q credit and DOE's $3.5B Regional DAC Hubs programme represent the largest government commitment to CDR in history. But the gap between current ~10,000 t/yr and the multi-gigaton scale needed by 2050 is enormous — requiring 100,000–1,000,000× scale-up.
~10,000 t CO₂/yr
Current operational DAC capacity globally (2024); ~0.00003% of needed 10 Gt/yr by 2050 in IPCC scenarios; Mammoth (Iceland, 36,000 t design) + Stratos (Texas, 500,000 t design); IEA DAC tracking 2024
$400–1,000/t CO₂
Current DAC cost range; Climeworks ~$1,000/t; Carbon Engineering pilot ~$400/t (own estimate); wide spread reflects different technologies and maturity; target: $100–200/t by 2030
$180/t 45Q + $3.5B
US IRA 45Q credit for DAC geological storage ($180/t — highest in the credit); plus $3.5B DOE Regional DAC Hubs; largest public CDR investment globally; 4 hubs selected 2023–2024
1,000–2,000 kWh/t CO₂
Energy required per tonne of CO₂ captured in current DAC systems; thermodynamic minimum ~250 kWh/t; current best ~1,000 kWh/t; must use clean energy or energy use offsets climate benefit
~2–10 Gt CO₂/yr
CDR (including DAC) needed in IEA NZE 2050; requires massive learning curve cost reductions and scale-up; 2030 interim target ~0.1–0.4 Gt/yr (IPCC AR6)
Microsoft / Stripe / Shopify
Early corporate buyers forming "advance market commitments" (AMCs) for DAC credits; Frontier Climate ($925M AMC, 2022); essential for early-stage commercialisation before government policy matures
DAC Technology Comparison — Solid vs. Liquid Systems
Source: McQueen et al. 2021 (Progress in Energy — DAC literature review); Fasihi et al. 2019 (Joule — DAC techno-economic assessment); Keith et al. 2018 (Joule — industrial DAC feasibility); Sanz-Pérez et al. 2016 (Chemical Reviews — direct air capture CO₂); Lackner 2003 (Science — CO₂ capture).
How Direct Air Capture Works
Liquid Solvent DAC (Carbon Engineering / Oxy "Stratos"): Air is drawn through a contactor where it meets a potassium hydroxide (KOH) liquid solution. CO₂ reacts with KOH to form potassium carbonate (K₂CO₃). The solution is then processed through a pellet reactor (adding calcium hydroxide) to precipitate calcium carbonate (CaCO₃). The CaCO₃ is heated in a calciner at ~900°C to release pure CO₂ and regenerate calcium oxide (CaO). This is essentially a large-scale industrial chemical plant process.
Solid Sorbent DAC (Climeworks "Mammoth"): Air passes over solid sorbent materials (amines on a porous support) that bind CO₂ from the atmosphere at ambient temperature. When saturated, the sorbent is heated to ~80–120°C (much lower than liquid DAC) to release concentrated CO₂. The sorbent is cooled and regenerated. This is a batch/cyclic process. Lower temperature means it can use waste heat; faster temperature swings allow smaller equipment per tonne of CO₂.
Source: Keith et al. 2018; Fasihi et al. 2019; Climeworks technical disclosures 2022–2024; Carbon Engineering 2023 process description; National Academies of Sciences 2019 (Negative Emissions Technologies report).
Operating & Under-Construction DAC Capacity (t CO₂/yr design, 2024)
Source: IEA 2023 (Tracking Clean Energy Progress — DAC); Global Thermostat 2023; Heirloom Carbon 2024; Climeworks Mammoth project data 2024; Oxy 1PointFive Stratos 2024 commissioning announcements; DOE DAC Hubs programme selections 2023; Carbon180 2024.
Key Players & Projects
Climeworks (Switzerland) — Orca & MammothOrca: 4,000 t/yr (2021, Iceland); Mammoth: 36,000 t/yr (2024, Iceland); geothermal powered; CarbFix mineralisation in basalt; most advanced commercial DAC company; ~$1,000/t; Series E raised $650M (2022)
Carbon Engineering / Oxy (1PointFive) — StratosCanadian tech (Keith group); acquired by Oxy (Occidental Petroleum); Stratos plant: 500,000 t/yr design, Permian Basin Texas; commissioning 2024–2025; uses Oxy oilfield CO₂ injection infrastructure; ~$400/t cost estimate
Heirloom Carbon (USA) — solid limestoneNovel approach: heats calcium carbonate into CaO, exposes it to air (CaO spontaneously absorbs CO₂); renewable electricity + solar thermal; ~$300/t target; received $600M DOE DAC Hub (Project Cypress, Louisiana)
Global Thermostat (USA)Amine monolith sorbents; waste heat integration; Georgia Tech founded; pilot plants; slower commercial progress; pivoting to co-location with industrial waste heat sources
DOE DAC Hubs (USA, 2023–2025)4 hubs selected: Project Bison (Wyoming, solid sorbent), Project Cypress (Louisiana, Heirloom), South Texas DAC Hub (Oxy), Pacific Northwest (electroswing); total ~$3.5B; target: >1 Mt CO₂/yr each by 2030
Source: Climeworks press releases 2024; Oxy 1PointFive Stratos 2024; DOE FECM 2023 (DAC Hub awards); IEA DAC 2023; Project Cypress EIA 2023.
DAC Cost Learning Curve — Historical & Projection ($/t CO₂)
Source: Fasihi et al. 2019 (Joule); IPCC AR6 WG3 2022 (Annex III — DAC costs); IEA 2022 (DAC Tracking Clean Energy Progress); Rubin et al. 2015 (Progress in Energy & Combustion Science — learning rates for energy tech); Beuttler et al. 2019 (Frontiers in Climate — role of DAC); McQueen et al. 2021.
Cost Drivers & Learning Rate Analysis
Current Climeworks cost (Orca/Mammoth)~$800–1,000/t CO₂ (2024); primarily heat energy (~70% of opex) + capex amortisation; transparency is commendable vs. industry; learning rate from Orca→Mammoth: ~40% cost reduction per doubling of capacity
Carbon Engineering / Stratos estimate~$300–400/t (Keith et al. 2018 analysis); disputed as optimistic; at commercial scale with heat integration; actual Stratos costs not yet disclosed
Historical energy tech learning ratesSolar PV: 28% cost reduction per doubling of capacity; wind: 23%; lithium batteries: 18–20%; if DAC achieves 15–20% LR: at 1 Gt/yr scale → ~$100–200/t by 2035 (Fasihi 2019)
Thermodynamic minimum cost~250 kWh/t CO₂ + compression/storage; at $0.02/kWh = $5/t electricity alone; total minimum ~$25–50/t including balance of plant; current systems 20–40× above thermodynamic limit
2030 cost target (DOE Earthshot)US DOE Carbon Negative Shot: "DAC at $100/t CO₂e by 2030"; ambitious; requires 10× cost reduction in 6 years; analogous to $1/W solar goal (achieved) but harder
Source: Fasihi et al. 2019; McQueen et al. 2021; IEA 2022; DOE Carbon Negative Earthshot 2021; Rubin et al. 2015.
The "420 ppm problem" — why DAC is inherently more expensive than point-source CCS: The fundamental challenge of DAC is thermodynamic: it must extract CO₂ from a mixture where CO₂ is 420 parts per million of the air — compared to 10–15% (100,000–150,000 ppm) in a power plant flue gas, or up to 95%+ in a natural gas sweetening stream. The energy required to concentrate CO₂ scales logarithmically with the starting concentration: capturing from 420 ppm requires ~3× more theoretical separation work than from 10% concentration. This is why DAC will always cost substantially more than point-source CCS, and why the technology needs to be reserved for cases where point-source capture is impossible (distributed transport emissions, legacy atmospheric carbon, residual emissions from agriculture). DAC is a "luxury" removal option for the last ~10–20% of emissions — not a substitute for eliminating emissions at source.
Energy Required for 1 Gt CO₂/yr DAC — Comparison to Other Uses (EJ/yr)
Source: McQueen et al. 2021 (Progress in Energy); National Academies 2019; IEA 2022; Fasihi et al. 2019; Davis et al. 2018 (Nature Energy — energy systems and CDR); IRENA 2022 (Renewable Energy Statistics 2022).
Energy, Land & Water Footprint
Energy per tonne CO₂ (current liquid DAC)~8.1 GJ/t (2.25 MWh thermal + 0.37 MWh electric); Carbon Engineering figures; dominated by calciner heat; must come from clean sources
Energy per tonne CO₂ (solid sorbent DAC)~5–6 GJ/t (1.1–1.4 MWh thermal at <150°C + 0.15–0.3 MWh electric); lower temperature is key advantage; can use low-grade waste heat
Energy to capture 1 Gt CO₂/yr (solid DAC)~5–6 EJ/yr heat + 0.2–0.3 EJ/yr electricity; global electricity generation 2022 was ~89 EJ; so 1 Gt/yr DAC ≈ 0.6% of current global electricity generation
Land use for DAC vs. forests1 Gt/yr at current fan capacity: ~700–1,200 km² (engineered system); natural reforestation for 1 Gt/yr: ~300,000–700,000 km² (varies by productivity); DAC uses ~300–1,000× less land than reforestation
Water useSolid DAC: ~2,000 t H₂O / t CO₂ in humid conditions (releases on sorbent heating — water capture); liquid DAC: ~1,500 L/t CO₂; less water than agriculture-based CDR
Source: McQueen et al. 2021; Keith et al. 2018; National Academies 2019; Fasihi et al. 2019; Davis et al. 2018.
DAC Credit Market — Price & Volume (2019–2024)
Source: Ecosystem Marketplace 2023 (Voluntary Carbon Market Insights); Patch 2023; Frontier Climate 2022 (advance market commitment report); Stripe Climate 2023 (annual CDR purchases report); Shopify Sustainability Fund 2023; Carbon Direct 2023 (CDR portfolio analysis); Puro.earth 2023 (CDR marketplace data).
Advance Market Commitments — Corporate Buyers
Frontier Climate ($925M AMC, 2022)Stripe, Alphabet, McKinsey, JPMorgan, Shopify; committed to buy $925M of CDR by 2030; catalytic signal; 9 CDR suppliers in 2023 portfolio; ~$300–1,000/t price; focus on durable removal
Stripe Climate (2020–present)First major corporate CDR buyer; $50M+ in CDR purchases; Climeworks, Charm Industrial, Heirloom, CarbonCure; pioneered "durability" as credit quality criterion; most rigorous buyer
Microsoft — carbon negative by 2030Committed to be carbon negative by 2030 and remove all historical emissions by 2050; purchasing DAC + bioenergy CCS; 2023 CDR portfolio: 3M t CO₂ contracted; ~$200–600/t range
Airlines — SAF + CDRDelta, United, Alaska purchasing DAC credits as part of net-zero offset strategy; aviation residual emissions (hard to eliminate completely) are primary long-term DAC market
EU Innovation Fund€10B+ fund supporting CCS and CDR in Europe; €3B+ allocated to DAC projects including Climeworks expansion and Point Hummingbird (Statkraft Norway DAC)
Source: Frontier Climate 2022; Stripe Climate Annual Report 2023; Microsoft Environmental Sustainability Report 2023; EU Innovation Fund 2023 award announcements.
DAC Scale-Up Trajectory vs. IPCC Requirements (Mt CO₂/yr)
Source: IPCC AR6 WG3 2022 (CDR pathways in 1.5°C scenarios); IEA NZE 2050; Fasihi et al. 2019; McQueen et al. 2021; Lawrence et al. 2018 (Nature Climate Change — CDR portfolio); National Academies 2019; DOE Carbon Negative Earthshot 2021.
Scaling Challenges & Enablers
Supply chain for sorbentsSolid amine sorbents degrade over time; replacement frequency critical to cost; supply chains for amines, alumina support, MOF materials need to scale 1,000–10,000× for Gt/yr scale
CO₂ storage infrastructureDAC must be co-located with storage or CO₂ transport infrastructure; geological surveys needed; CarbFix (Iceland basalt) shows fast mineralisation but site-specific
Renewable energy availability1 Gt/yr DAC requires ~5–8 EJ renewable energy — ~0.6% of current global power; doable but requires purposeful siting near cheap renewables; geothermal (Iceland), solar (Sahara/Atacama), wind (Patagonia)
Manufacturing scale-up (modular DAC)Solid sorbent DAC units are modular (like shipping containers); amenable to factory production; Climeworks manufacturing scale-up is key bottleneck; target 10,000× unit production by 2030
Workforce & permitting1 Gt/yr DAC ≈ 1 million workers in operations + supply chain; permitting for CO₂ injection at scale requires Class VI UIC wells (EPA) — limited approvals historically; DOE streamlining
Source: IPCC AR6 2022; IEA 2022; Fasihi et al. 2019; DOE DAC Hubs 2023; McQueen et al. 2021.
Mammoth — Climeworks' 36,000 t/yr Iceland plant and what it teaches us about scale: Climeworks' Mammoth plant at Hellisheiði, Iceland (commissioned 2024) is 9× larger than its predecessor Orca, demonstrating the capacity to scale by near-order-of-magnitude in a single project cycle. It uses geothermal energy — making the energy footprint near-zero — and injects CO₂ into basalt via CarbFix, where it mineralises into rock within 2 years. But Mammoth's 36,000 t/yr design capacity, even if fully achieved, is 0.0000036 of the 10 Gt/yr CDR needed by 2050 in optimistic scenarios. The world needs ~280,000 plants of Mammoth's size, or alternatively ~280 plants of 10 Mt/yr size (Stratos-class). The scale of the manufacturing, energy, and geological infrastructure challenge is without precedent. That said, the same numbers applied to solar PV in 2000 would have seemed equally impossible. The question is whether policy signal, corporate demand, and manufacturing cost reduction can follow a similarly steep trajectory.