Direct Air Capture's supposed energy problem is overblown

DAC companies will be responsible stewards of American grids and increase renewable penetration in the long run

You’re reading Terraform Now, a newsletter on the business of carbon removal. To support my work, you can subscribe or share this article with interested folks:

The New York Times recently highlighted a thorny issue for Direct Air Capture (DAC) companies: procuring low-emissions or zero-carbon power. As they put it:

Thanks in large part to booming data center construction and the surge in artificial intelligence, electricity demand in the U.S. is rising for the first time in decades.

But the timing of this rise in demand, as far as the race to fight global warming goes, is awkward.

To help fight climate change, the nascent direct air capture industry is racing to develop technology to remove carbon dioxide directly from the sky. But in some areas of the country, the DAC industry is facing a shortage of available renewable energy to run its operations, particularly wind and solar power.

It’s a fair concern. To remove carbon effectively and credibly, DAC must rely on clean power. But a closer look at the numbers should put to rest any notion that these startups will overload America’s grids. Like the big data center builders who came before them, DAC developers will ultimately add more clean power to the system than they take out.

A near-term challenge, but not a crisis

Yes, finding renewable energy (and the transmission capacity to move it around) is challenging, especially in Louisiana where many of these plants will be built. But in the U.S., we have a menu of workarounds:

• Purchasing green tariffs on “dirtier” grids (like Louisiana’s)

• Investing in renewable energy credits (RECs) or power purchase agreements (PPAs) on other grids

• Relying on other incentive programs and the broader drive to decarbonize the grid

Each of these approaches directly or indirectly finance new clean energy capacity.

In the short run, it’s easy to assume that DAC plants will become voracious electricity hogs. But the scale of electricity usage through the early 2030s will be modest. Take the three large-scale plants being discussed in Louisiana, one from Climeworks, Heirloom, and CarbonCapture. Each will likely have a nameplate capacity in the range of 100,000 to 250,000 tons of CO₂ per year.

Let’s assume these plants will require 1.5 MWh per ton of CO₂ captured, which is on the higher end of public estimates by DAC companies. That means each plant might draw about 150–375 gigawatt hours annually. If all three of these plants (again, from Climeworks, Heirloom, and CarbonCapture) come online, that’s somewhere between 0.5 and 1 terrawatt hours (TWh) of total annual demand.

1 TWh seems like a lot but it’s small in the grand scheme of things. Meta, for example, is building a data center complex in Louisiana in the 6 - 12TWh range. In all, AI data centers are going to add hundreds, if not thousands, of TWh in the coming decade.

And as a generation load, 1 TWh is not that large. If Louisiana approved a new large nuclear reactor, it could power all three next-gen DAC plants 10 times over.

1 TWh is a worthy price to find out if DAC scales well

DAC providers are on a mission to drive down costs from today’s $500–$1,000 per ton of CO₂ to a target of $100–$200 per ton by the late 2030s. At that cost, DAC could potentially scale to mitigate ~50% of human CO₂ emissions in 2050 and beyond. It’s the only carbon removal technology that can get big enough to have a truly lasting impact on our atmosphere.

But to get big, DAC must get cheap. To get cheap, it needs next-gen plants to take advantage of:

• Learning: Larger plants allow engineers to learn and refine processes in real-world operating conditions—R&D that can’t happen purely on lab benches

• Economies of scale: As we’ve seen in solar, once manufacturers and plant builders start repeating the same process at scale, costs can plummet

• Supply chain optimization: Sourcing cheaper materials and standardizing manufacturing steps become easier at higher volumes

These new plants could demonstrate 25 - 50% drops in cost per ton, and unlock improvements in energy efficiency. If they can deliver on projected cost declines, that opens up the opportunity for more DAC plants. We should give these companies a chance to execute — 1TWh is a small price to pay to potentially unlock a powerful new emissions-reduction tool.

DAC companies are following the data-center playbook

If there’s still doubt that DAC companies will do right by the grid, consider the way data-center builders have reshaped the energy market. For the last decade, hyperscalers like Microsoft, Meta, and Google have purchased green tariffs, secured PPAs, and championed new transmission lines. Put differently, they have actively funded the renewable energy transition.

At a conference last week, Microsoft’s Charlie Sellars noted that the company has added around 34 GW of clean energy to grids worldwide — that’s close to 100 TWh on an annual basis.

DAC companies are poised to do the same. Climeworks, Heirloom, and CarbonCapture surely recognize that their reputation, compliance requirements, and environmental outcomes hinge on powering facilities with as much clean electricity as possible. They will all likely be willing to pay premium rates for the renewable energy incentives I mentioned earlier, knowing that these efforts eventually help bring more wind, solar, nuclear, or geothermal projects online.

Closing thoughts

In the short run, the anxiety over DAC’s electricity demand is understandable. We are living through a transition in our energy systems—with new demands and new decarbonization mandates piling up all at once. But DAC’s power requirements, at least in these early plants, are miniscule compared to massive data-center complexes and other large industrial users.

And investigating whether a DAC scale-up is possible is worth it. We need a number of cheap, scalable carbon removal solutions to manage with the atmosphere in the second half of this century.

The only way to figure out whether $100–$200 per ton is genuinely attainable is by building the next generation of plants, letting engineers iterate, and (yes) buying up enough clean power and credits to ensure that net CO₂ removal remains credible. Call me an optimist, but I genuinely believe that DAC, like the data centers before it, will be a catalyst for more green power—not a burden.

Disclosure: My employer, Climeworks, is one of the companies building a DAC plant in Louisiana. This article was compiled entirely from public information, and does not reflect the views of Climeworks.