The U.S. Has Gotten 3 Direct Air Capture Plants in 13 Months

A new direct air capture facility built by the Alphabet-backed 280 Earth is officially plucking carbon dioxide from the surrounding air along the Columbia River in Oregon, the company announced on Monday. It’s the third-largest “direct air capture” plant operating in the United States and the latest entrant in the race to design the cheapest, most efficient machine to strip the heat-trapping gas from the atmosphere.

The small-scale demonstration project, which neighbors a Google data center in a city called The Dalles, is expected eventually to capture carbon at a rate of 500 tons per year. The two other U.S. facilities — Global Thermostat’s plant in Commerce City, Colorado, and Heirloom’s plant in Tracy, California — are both designed to capture 1,000 tons per year. All three came online in just the past 13 months. (There are also a handful of smaller facilities operating in the U.S. that capture 100 tons per year or less.)

The team that is now 280 Earth first began working on their direct air capture system inside X, the tech incubator at Google also known as the “Moonshot Factory.” They spun out into their own company in 2022, after four-and-a-half years of research and development. The name comes from 280 parts per million, the amount of carbon in the atmosphere before industrialization. Today we’ve reached nearly 420 parts per million. But if the world manages to reduce emissions nearly to zero, it may be possible to pull enough carbon out of the air to restore the atmosphere to levels closer to pre-industrial times.

In general, direct air capture technologies suck in ambient air and pass it through a special material called a sorbent that attracts CO2 molecules. They then use heat to remove the carbon from the sorbent so that it can be transported and safely stored underground.

280 Earth’s approach is unique in a few ways. To begin with, the company is using a “pelletized” sorbent — CEO John Pimental described it to me as a “half a piece of uncooked rice,” or the innards of a bean bag chair. The tiny pellets of sorbent flow through the system almost like water, resulting in some operational efficiency gains.

For the second step, the company plans to use waste heat from other industrial facilities like data centers to remove the captured carbon from the sorbent. Many data centers circulate cold water through their facilities for cooling, then send the hot water to a cooling tower where the heat is released into the atmosphere. 280 Earth can instead take that hot water and run it through a heat exchanger, sending the now-cooled water back to the data center. “It means their cooling tower needs to work less hard, it has less load on it,” said Pimental. “So it's an additional revenue source for our company to provide those cooling services to a neighbor.”

This waste heat can meet up to 80% of 280 Earth’s operational needs, reducing the amount of electricity the company buys. It’s also a win-win for the data center — 280 Earth’s process pulls water from the air in addition to carbon, and can supply that water to the data center, which in turn doesn’t have to rely as much on natural sources.

Direct air capture technology is often called “speculative” and “unproven.” But with an increasing number of deployments in the real world, it’s worth being more specific. These machines have proven to be able to separate carbon out of the air. The question is whether they can do so permanently, economically, and at a scale that will actually make a difference for climate change.

Although more plants are coming online every year, those questions are unlikely to be answered anytime soon. For example, it will be impossible to judge the efficiency claims made by 280 Earth or any other company until there is more public data — or any public data — about the energy these plants consume or what they cost to operate. Even the companies that are farthest along, like Climeworks, which has been operating a 4,000 metric ton per year commercial plant in Iceland since 2021 and just opened a 36,000-ton plant earlier this month, say that they are still testing the technology and therefore are not ready to share any stats that could be misinterpreted.

The potential to scale could also have less to do with the details of any one company’s technology and more to do with the ability to procure clean energy or to find somewhere to store the captured carbon.

Though 280 Earth is officially collecting CO2, the company doesn’t yet have anywhere to put it. Pimental told me the company plans to transport the gas by truck or rail to a carbon dioxide storage well, but it has not yet signed any agreements with well operators, and it’s unclear how long that could take. There are currently only a few operating carbon storage wells in the country, located in Illinois and North Dakota. But additional wells have been permitted in California, Indiana, and Wyoming, and many more are under review by the Environmental Protection Agency. Rather than sequester the carbon underground, the company could also sell it for industrial uses. Heirloom, for example, has an agreement with a company called CarbonCure to take the CO2 it captures and store it in concrete.

Regardless, 280 Earth company is aiming to scale up quickly and plans to build a new unit that can capture 5,000 tons of CO2 per year by 2025. Pimental told me that equipment procurement and permitting for that project are already underway. 280 Earth has not been awarded any government funding to date, but the company plans to compete to be one of the Department of Energy’s next direct air capture “hubs.”

Pimental told me he likes his odds. “I think we'll be in a very strong position because not many people have a 500 ton commercial demonstration facility up and running.”