The Issue

One of the newest and most heavily subsidized approaches to atmospheric carbon dioxide (CO₂) reduction is direct air capture (DAC). DAC systems are often described as giant vacuum cleaners; these systems extract CO₂ directly from ambient air using chemical processes that pull air over sorbent materials—substances used to collect liquids or gases. The captured CO₂ is then compressed and either transported for permanent underground storage or used elsewhere (for example, in enhanced oil recovery.

Several dozen DAC systems are in operation today, primarily in Europe, but most are small demonstration projects. One exception is Occidental Petroleum’s Stratos Plant, located in the Permian Basin of Texas. Once fully operational, it will be capable of extracting over 500,000 tons¹ of CO₂ annually, which would make it the largest DAC plant in the world—far ahead of the second largest, Iceland’s Climeworks Mammoth, which captures nearly 40,000 tons of CO₂ annually). For comparison, current anthropogenic annual global CO₂-equivalent emissions exceed 70 billion tons without adjusting for natural CO₂ uptake.²

Proponents of DAC and other carbon dioxide removal (CDR) methods—often grouped under the broader umbrella of carbon capture and storage (CCS; or, if CO₂ utilization is included, CCUS)—argue that removing billions of tons of CO₂ from the atmosphere will be necessary, in addition to aggressive net-zero emissions-reduction policies, to limit global temperature increases. The Inflation Reduction Act of 2022 in the U.S. and other mostly European legislation significantly boosted incentives for DAC. In the U.S., the incentive is through enhancements of the section 45Q tax credit, raising the maximum to $162 per ton³ for CO₂ that is captured and permanently stored underground (with lower amounts—up to $118—for CO₂ used in enhanced oil recovery).⁴ By comparison, market prices for CO₂ emissions allowances in compliance trading systems are substantially lower: about $23 per ton in the most recent auction;⁵ about $30 per ton in the May 2026 California-Quebec cap-and-trade auction;⁶ about $60 per ton in the March 2026 Washington State auction;⁷ and about €70 per ton in Europe during May 2026.⁸

The substantial subsidies for DAC raise several important policy questions. Is large-scale DAC physically feasible, given its high energy requirements and current and future cost? If it proves both physically and economically viable, what would its actual impact be on the climate or on the environment more broadly? Does DAC pass a basic cost-benefit test—that is, do the claimed benefits of reduced atmospheric CO₂ concentrations outweigh the full costs of deployment?

Energy Requirements

A key factor in evaluating the cost-effectiveness and practicality of large-scale DAC is its energy demand—both for operations and for the equipment and raw materials required at scale.

Although a variety of DAC technologies have been demonstrated, all are bound by the laws of thermodynamics. Every system must move vast quantities of air over or through CO₂-absorbing materials. This is typically done with large fans that create a pressure drop to pull air across sorbent beds or contactors. Given that CO₂ makes up only about 0.042% of the atmosphere by volume (or 425 parts per million [ppm]), capturing 1 billion tons of CO₂—even at an idealized 100% capture efficiency—would require moving about nearly 2.8 trillion tons of air, or about a cube of air nearly 80 miles on each side.⁹ In practice, real-world systems achieve far lower single-pass efficiencies, so the actual air volume (and energy) required is significantly higher.

Increasing the CO₂ capture rate requires higher airflow rates. Higher airflow, in turn, requires increasing the pressure drop—and therefore greater energy input—regardless of how efficient a fan’s motor and design are.¹⁰ That minimum energy requirement stems from fundamental fluid dynamics, including Bernoulli’s principle, which relates pressure, velocity, and elevation in moving fluids; this principle explains how airplane wings create lift.¹¹ The electricity requirement for real-world fans—ranging from 300 to 900 kilowatt-hours (kWh) per ton of CO₂ captured—is much higher than that of a theoretical 100%-efficient fan.¹²

There is also a minimum energy requirement to separate CO₂ from ambient air, again based on thermodynamics. At the current atmospheric CO₂ concentration of approximately 425 ppm, this minimum work of separation—at sea level and 32 degrees Fahrenheit—is just under 130 kWh per ton. This value increases at higher elevations or temperatures because air density decreases, requiring even more energy to process the same mass of CO₂.

Finally, the captured CO₂ must be purified, compressed, and prepared for transportation—whether for underground storage or for other purposes, such as enhanced oil recovery. Thermodynamics again determines the minimum amount of energy required for compression: approximately 60 kWh per ton. Adding these energy requirements together—fan energy (~90 kWh per ton), separation (~130 kWh per ton), and compression (~60 kWh per ton)—yields a theoretical minimum energy requirement of about 280 kWh per ton of CO₂ captured.¹³ No real-world DAC process can ever go below this floor, as it assumes 100% efficiency—which is impossible.

A single commercial-scale facility capturing 1 million tons of CO₂ per year would have a minimum theoretical energy requirement of over 280 gigawatt-hours (GWh) of electricity (0.28 megawatt-hours [MWh] per ton of CO₂). Real-life facilities consume a multiple of this theoretical floor.

Real DAC systems logically require far more energy than the theoretical minimum. No mechanical or chemical process operates at 100% efficiency. The electric motors that drive the fans lose some energy to friction and heat. The fan blades must overcome aerodynamic drag and friction within the system. The chemical processes and filters do not remove 100% of the CO₂ in a single pass. Energy is consumed by pumping chemicals to regenerate the liquid hydroxide solutions. Compressors are not 100% efficient. And so on.

Academic studies since 2011 have shown wide variation in projected DAC energy requirements, depending on the technology employed (see figure 1).¹⁴ The dashed line in figure 1 illustrates the estimated average energy requirement for DAC facilities, which is about 1,700 kWh per ton of CO₂ removed. One optimistic projection assumed more than a 50% reduction in energy use by 2030 through technological improvements—an assumption that many view as unrealistic.¹⁵

Climeworks, a developer of commercial DAC facilities, has claimed that its Generation 3 (Gen 3) system will achieve significantly higher energy efficiency, potentially requiring less than 1 MWh per ton of CO₂ removed.¹⁶ However, such claims are unlikely to reflect the full energy footprint when the entire DAC process is considered—including raw material extraction and processing, equipment manufacturing, transportation, construction, and ongoing facility operation. Although comprehensive data for full-system energy use are not publicly available, a more realistic total energy requirement for DAC—including all upstream and infrastructure inputs—is likely in the range of 1.8 to 3.6 MWh per ton of CO₂ removed.¹⁷

Figure 1. Published Estimates of DAC Energy Requirements

Source: Adapted from Lucas Desport et al., “Deploying Direct Air Capture at Scale: How Close to Reality?” Energy Economics 129 (January 2024): 107244, table 1. Values in the original studies were given in metric units. 

A recent U.S. Department of Energy report indicated that the U.S. may need to remove between about 100 million and 2 billion tons of CO₂ per year by 2050.¹⁸ Using the conservative estimate of 1 MWh per ton of CO₂, capturing 100 million tons of CO₂ in one year would require about 100 terawatt-hours (TWh) of electricity (about 25% of Germany’s annual electricity consumption).¹⁹ Capturing 2 billion tons would require over 2,000 TWh—almost half the total U.S. electricity consumption in 2025.²⁰

Producing this additional electricity would require vast quantities of new, reliable, uninterrupted generating capacity, since industrial CO₂ capture requires on-demand power independent of weather conditions. Modern nuclear plants in the U.S. operate at a net load factor above 90%.²¹ A typical 1,000-MW plant therefore produces 7.9 TWh per year.²² Thus, capturing 100 million tons of CO₂ at an average of just over 1 MWh per ton with DAC would require 100 TWh of electricity and about 13 GW of new nuclear capacity. Capturing 2 billion tons of CO₂ annually with DAC would require building at least 260 GW of new nuclear capacity—nearly three times the size of the existing U.S. nuclear plant fleet, which is just under 100 GW.²³

The U.S. Energy Information Administration (EIA) estimated the overnight capital cost (excluding financing) of building an advanced nuclear plant at about $8,000 per kW (in 2023 dollars) for twin 1-GW reactors.²⁴ Building the 260 GW of new nuclear capacity needed for 2 billion tons of annual DAC removal would therefore cost over $2 trillion. Financing costs, inflation during the multiyear construction period, and project overruns would likely double this figure. This estimate also excluded the additional high-voltage transmission infrastructure (e.g., power lines, substations, transformers) that is required to deliver this new capacity, which would add hundreds of billions of dollars more. When annualized over the life of the plants, the electricity cost alone for DAC would exceed $90 per ton of CO₂ removed.²⁵ Using wind and solar with battery and other storage and backup systems would face even greater challenges and even higher costs. The intermittency of wind and solar, the need to overbuild capacity, and the massive land and material requirements would make this approach prohibitively expensive and impractical at the required scale.²⁶

There is also the substantial capital cost of the DAC facilities themselves. The most optimistic projections claimed that total DAC costs could fall by roughly 80% to as low as under $100 per ton of CO₂ removed by 2050.²⁷ However, as explained previously, electricity costs alone are likely to exceed $90 per ton, which leaves little room for capital recovery, maintenance, and other operating expenses.

A 2019 National Academies of Sciences, Engineering, and Medicine study estimated the overnight capital costs for a generic, approximately 1-million-ton-per-year liquid-solvent DAC plant at between about $615 million and $1.15 billion.²⁸ Assuming a 20-year plant life and an 8% cost of capital, this translates into an annualized capital cost of approximately $60–$120 per ton of CO₂ captured.²⁹

More recent analyses indicate that earlier cost estimates for DAC have been overly optimistic.³⁰ For example, Occidental Petroleum’s Stratos Plant in the Permian Basin, which uses Carbon Engineering’s liquid-solvent DAC technology,³¹ has a current estimated capital cost of approximately $1.3 billion for its full 550,000 ton capacity.³² Assuming a 20-year plant life and an 8% cost of capital, this equates to an annualized capital cost of just over $130 million per year, or roughly $240 per ton of CO₂ captured—before accounting for energy, operations, and maintenance expenses.³³

A 2021 study examined different configurations of liquid DAC systems, including those using natural gas for calcination and fully electric systems powered by wind, solar, or nuclear.³⁴ For wind- and solar-powered configurations, the researchers estimated total costs (DAC facilities plus energy) for a 1.1-million-ton-per-year plant at approximately $330–$630 per ton. For nuclear-powered systems, the range was about $360–$560 per ton.³⁵

These cost estimates—which are likely too low, given rising costs for new capacity over the past five years—far exceed the most recent official estimates of the social cost of carbon (SCC), a modeling exercise that attempts to quantify the future damages caused by emitting an additional 1.1 tons of CO₂. In 2023, the U.S. Environmental Protection Agency (EPA) published updated SCC values of approximately $190 per ton in 2025, with a projected rise to about $210 per ton by 2030.³⁶ These figures are more than four times higher than the $46-per-ton estimate published in 2021 by the Interagency Working Group.³⁷ Although SCC estimates are controversial—relying on assumptions and predictions that extend centuries into the future³⁸—even these elevated values are lower than the full costs of CDR using DAC.

Climate Impacts

If large-scale DAC capability could be developed and deployed successfully, how would it affect the climate? In 2024, U.S. energy-related CO₂ emissions were just over 5 billion tons.³⁹ Another 330 million tons were estimated to come from non–energy-related sources.⁴⁰ Overall, world energy-related CO₂ emissions were more than 40 billion tons in 2024; when other greenhouse gases such as methane were included, total emissions approached 80 billion tons —without adjusting for natural CO₂ uptake by the biosphere or oceans.⁴¹ Those emissions have increased by an average of nearly 300 million tons per year over the past decade. Removing a mere 100 million tons of CO₂ in one year would be equivalent to just over one week of 2024 U.S. emissions and less than one day of 2024 global emissions, and it would offset less than five months of average annual energy-related emissions growth.

The potential effects on global temperature from different levels of DAC deployment can be estimated using the Intergovernmental Panel on Climate Change estimates of climate sensitivity. At the current atmospheric CO₂ concentration of about 425 ppm, removing 110 million tons of CO₂ with DAC would reduce atmospheric CO₂ by about 0.0116 ppm and lower the average global temperature by about 0.0002 degrees Fahrenheit. Removing 2 billion tons would reduce the atmospheric concentration by 0.232 ppm and lower the average global temperature by 0.004 degrees Fahrenheit.⁴²

To put these DAC-driven temperature “improvements” in context, global mean temperatures have uncertainty ranges of about ±0.09 degrees Fahrenheit to ±0.234 degrees Fahrenheit.⁴³ In other words, the estimated temperature reductions would be far smaller than the inherent noise—or measurement uncertainty—in global temperature records.

Using MAGICC, a climate carbon cycle model,⁴⁴ it is possible to project the potential temperature impacts from sustained DAC removal over several decades. For example, if 2.2 billion tons of CO₂ were removed annually starting in 2030 and continuing through 2100, a cumulative total of 154 billion tons of CO₂ would be removed. Under the median shared socioeconomic pathway 4.5 scenario,⁴⁵ this sustained removal would reduce average global temperature by 0.108 degrees Fahrenheit by 2100, lowering projected warming from 4.89 degrees Fahrenheit to 4.78 degrees Fahrenheit above preindustrial levels (see figure 2). The temperature reduction is less than one-fourth of the typical ±0.234 degrees Fahrenheit uncertainty band in global mean temperature—that is, indistinguishable from natural variability and measurement noise.

Figure 2. Estimated Impact of DAC Removal, 2030–2100

Note: SSP 4.5 refers to shared socioeconomic pathway 4.5, a framework used by climate scientists to model future greenhouse gas emissions scenarios.

Source: Authors, using “MAGICC (Model for the Assessment of Greenhouse Gas Induced Climate Change),” Integrated Assessment Modeling Consortium, 2022, accessed June 5, 2026, https://magicc.org.

Perspectives

This issue brief demonstrates that DAC is neither economical nor practical for addressing climate change. Even if deployed on an unrealistically massive scale, DAC would have virtually no measurable impact on global temperatures. Despite its exorbitant cost and negligible impact on global climate, DAC is heavily subsidized through the section 45Q tax credit. If DAC facilities capable of capturing about 100 million tons of CO₂ are built by 2033, the credit would cost taxpayers approximately $18 billion per year—or $180 billion over the 10-year eligibility period.

With high costs far exceeding any realistic climate benefit, DAC fails any reasonable cost-benefit test. The current tax credits therefore represent a significant and unjustifiable waste of taxpayer dollars. A more productive use of these funds would be to redirect them toward research and development of lower-cost, advanced nuclear power technologies, which offer far greater potential for meaningful emissions reductions.