The U.S. Spent $3.5 Billion on Machines That Suck Carbon From the Air. A Nature Study Says Solar Panels Remove 10 Times More Carbon per Dollar.

Five thousand five hundred kilowatt-hours. That is how much energy today's commercial direct air capture systems consume to pull a single ton of carbon dioxide from the atmosphere, according to a study published this month in Communications Sustainability by researchers at PSE Healthy Energy, Boston University, and the Harvard T.H. Chan School of Public Health. For context, 5,500 kWh would power an average American home for roughly six months.

Kashtan and colleagues asked a question the DAC industry would rather not hear: what if you spent that money on solar panels instead?

Across nearly every combination of grid region, technology trajectory, and timeline the researchers modeled, renewables deliver more combined climate and public health benefit per dollar. Not marginally more. Dramatically more. And in the scenario that reflects where commercial DAC technology actually sits today, the machines do not just underperform solar and wind. They make things worse.

The $100 Million Experiment

Lead author Yannai Kashtan and colleagues designed what amounts to a controlled fiscal experiment. They took $100 million per year in 2024 dollars, a figure corresponding to the upper end of existing U.S. renewable installations, and modeled two ways to spend it: build a direct air capture facility, or build utility-scale wind or solar. They ran this comparison across 22 U.S. electrical grid regions from 2020 through 2050, using eight grid evolution scenarios from the Energy Information Administration's Annual Energy Outlook and the CoBE Projection tool to estimate both climate benefits (CO2 removed or displaced) and health benefits (reduced particulate matter, nitrogen oxides, and sulfur dioxide from fossil fuel combustion).

They tested four DAC technology trajectories, ranging from where commercial systems sit now to where optimists hope they will land:

Scenario	Energy Use (kWh/ton)	Cost ($/ton)	Result vs. Renewables
Stagnation	5,500	$1,000	Net negative impact
Efficiency Improvement	2,500	$750	Near-zero benefit
Advanced Efficiency	1,500	$500	Modest benefit, less than renewables
Breakthrough	800	$100	Modestly outperforms renewables

Read that first row again. Net negative impact. Under the Stagnation scenario, which maps closely to Climeworks' current operating parameters, grid-connected DAC pulls so much electricity from a still-partly-fossil grid that it generates more greenhouse gases and air pollution damage than the carbon dioxide it captures can offset. You would do less harm to the climate by burying the $100 million in a hole.

Why Renewables Win: The Health Dividend

The mechanism is straightforward once you see it, but it has been almost entirely absent from the DAC policy conversation. When you build a solar farm that feeds the grid, it displaces fossil generation. That displacement does not just reduce CO2. It also cuts the particulate matter, sulfur dioxide, and nitrogen oxides that coal and gas plants emit, pollutants that drive cardiovascular disease, respiratory illness, and premature death in communities near those plants. Under EPA methodology, these avoided health damages are worth $100 to $400 per ton of CO2 displaced, depending on the regional fuel mix.

DAC has no equivalent mechanism. It captures CO2. Period. When grid-connected DAC draws power from a regional grid that still burns fossil fuels, it increases local air pollution emissions to run its fans and heat its sorbents, producing health harms in the same communities the climate spending is supposed to protect.

This is the calculation that changes the policy math. If you evaluate DAC on carbon removal alone, you can construct scenarios where it looks reasonable. Once you add the health bill for the electricity it consumes, the comparison collapses in 21 of 22 grid regions.

What $3.5 Billion Could Have Bought

The U.S. Department of Energy has allocated $3.5 billion for DAC hubs under the Bipartisan Infrastructure Law, including $500 million to 1PointFive's South Texas hub (an Occidental Petroleum subsidiary) and funding for Project Cypress in Louisiana (a Battelle and Heirloom Carbon partnership). A separate incentive, the 45Q tax credit, adds another $180 per ton for DAC facilities that store carbon permanently underground. These are real commitments backed by real appropriations.

Applying the Kashtan study's framework to actual federal allocations produces numbers the DOE has not published. Here is the math.

At the Advanced Efficiency scenario ($500/ton, 1,500 kWh), which is more generous than where any commercial DAC facility operates today, $3.5 billion buys roughly 7 million tons of lifetime capture capacity. Kashtan's analysis shows this delivers modest climate benefits but still less than renewables in nearly every grid region through 2050.

Investing the same $3.5 billion in utility-scale solar at current installed costs of approximately $1.00 per watt would build roughly 3.5 gigawatts of capacity. At the U.S. average solar capacity factor of 25%, that generates about 7.67 terawatt-hours per year. Displacing grid generation at the national average emissions intensity of 0.39 tons CO2 per megawatt-hour, that is approximately 2.99 million tons of CO2 displaced every year. Over a 25-year solar panel lifespan, those panels displace roughly 75 million tons, with no degradation of the climate benefit as long as the panels operate. Health co-benefits, using the EPA's $100-to-$400 range per ton of CO2 displaced, add between $299 million and $1.2 billion annually in avoided premature deaths, hospitalizations, and lost workdays.

By the numbers: solar displaces roughly 10 to 11 times more CO2 per dollar over 25 years than DAC at the Advanced Efficiency cost tier, and the multiple grows wider when you add health benefits. At the Stagnation tier where commercial DAC actually operates, the comparison is not a ratio. It is a sign reversal.

The Strongest Case for Spending Anyway

The DAC industry's best counterargument is structural, not incremental, and it deserves its full weight.

Renewables and DAC address structurally different parts of the carbon problem. Solar and wind prevent new emissions by displacing fossil generation. They cannot remove the trillion-plus tons of legacy CO2 already in the atmosphere. DAC can. After the world reaches net-zero emissions, whenever that happens, atmospheric restoration requires carbon removal at scale, and DAC is one of the few technologies that can deliver permanence (geological storage, not temporary sinks like forests that can burn). Kashtan explicitly acknowledges this: DAC's role is "atmospheric restoration" after anthropogenic emissions reach zero.

Investing $3.5 billion in DAC now, on this argument, is not buying carbon removal at today's prices. It is buying industrial learning, supply chain development, and manufacturing scale-up so that DAC is ready when the grid is clean enough to power it without generating offsetting pollution. You would not compare a 1970s solar panel's cost-per-watt to coal and declare solar a failure. You would recognize the investment as building toward a future cost curve.

This argument has real force, but it also has a quantitative problem the Kashtan study exposes. Reaching cost-competitiveness requires 800 kWh per ton at $100 per ton. Commercial systems today sit at roughly 5,500 kWh and $600 to $1,000 per ton. That is not a learning curve. It is an order of magnitude in energy efficiency and a ten-fold cost reduction simultaneously, a performance improvement that took solar panels approximately 40 years to achieve (from over $70 per watt in 1977 to under $0.30 per watt today). DAC proponents are betting they can compress that timeline. They might be right. But the study shows what we are paying while we wait.

The Grid Timing Problem

There is a deeper structural finding buried in the regional data that the summary statistics obscure. Grid-connected DAC's climate performance improves as the grid itself decarbonizes, because cleaner grid electricity means less fossil pollution per ton of CO2 captured. By 2045, under aggressive decarbonization scenarios, the health penalty for grid-connected DAC shrinks substantially, and the opportunity cost gap narrows.

This creates a paradox for DAC policymakers. The technology makes the most fiscal sense precisely when the grid is clean enough that the emergency justification for deploying it has diminished. A 2045 grid powered predominantly by renewables, nuclear, and storage is a grid where the remaining hard-to-abate emissions come from cement, aviation, steel, and agriculture, not from electricity generation. At that point, the $100 million comparison shifts: DAC versus decarbonizing a cement kiln, not DAC versus solar panels. The Kashtan study does not model that comparison and its authors explicitly flag it as the next research question.

For federal policymakers deciding how to spend limited climate dollars today, though, the 2025-2035 window is the relevant one. And in that window, the numbers are not ambiguous. Every dollar routed to DAC instead of renewables buys less climate protection and produces zero health benefits in communities breathing fossil plant emissions.

What We Did Not Prove

Our opportunity cost calculation for the $3.5 billion DOE allocation uses national averages for solar capacity factor (25%), grid emissions intensity (0.39 tCO2/MWh), and installed solar cost ($1.00/W), which vary significantly by region. Texas grid regions, where major DAC hubs are sited, have above-average solar resources but also different emissions profiles. Our 25-year solar lifespan assumption is standard but panels degrade approximately 0.5% per year, which we did not model; cumulative degradation reduces our 75-million-ton displacement estimate by roughly 6%. We used the Advanced Efficiency tier ($500/ton) for the DAC comparison, which is more generous than current commercial performance, making our ratio conservative. The Kashtan study itself does not model learning-by-doing cost reductions that early DAC deployment may generate, nor does it compare DAC against decarbonizing hard-to-abate industrial sectors where renewables cannot directly substitute. The 45Q tax credit effectively reduces DAC's private cost to $320-$820 per ton depending on the base cost tier, but the opportunity cost to the public fisc remains the same regardless of which pocket the money comes from.

The Bottom Line

The Kashtan study reframes the DAC conversation from "can it work?" to "compared to what?" The answer, through 2050, is that renewables deliver more climate benefit and substantially more health benefit per dollar in nearly every U.S. grid region. DAC only outperforms renewables under a breakthrough scenario that requires a simultaneous 7x improvement in energy efficiency and 6-10x cost reduction, performance targets that no commercial system has approached.

If you are a climate policy staffer deciding how to allocate limited federal funds between 2026 and 2030: the peer-reviewed evidence now says each dollar spent on utility-scale solar or wind delivers more combined climate and health benefit than the same dollar spent on direct air capture in nearly every region of the country, and you should weight that finding against the speculative industrial-learning argument for early DAC deployment. If you run a DAC company: your target is sharp and the study quantifies it precisely, meaning 800 kWh per ton and $100 per ton, because anything above those thresholds loses to renewables on every metric that matters. If you are a corporate buyer purchasing carbon removal credits to offset emissions: ask your DAC provider what grid region powers their facility and what the grid's current fossil share is, because in coal-heavy regions, your credit may be producing net positive emissions. If you are a taxpayer: $3.5 billion in DAC hub funding represents a bet that the technology will achieve a cost reduction that took solar four decades, compressed into roughly one, and the opportunity cost is approximately 75 million tons of CO2 displacement that those dollars would have delivered as solar panels generating clean electricity for a quarter century.