Technology that captures carbon emissions from power plants may get a breakthrough as deep-pocketed tech companies try to meet climate goals while powering the AI race, according to Axios. The push is not coming from communities deciding how to live with less extraction and less waste; it is coming from corporations trying to keep the AI machine fed while dressing the expansion up as climate responsibility.
Who Is Driving the Buildout
The article said at least five projects under consideration across the U.S. would capture carbon dioxide emissions from natural gas plants connected to data centers, based on company statements, news coverage and independent reports. Those projects include Google's publicly touted project in Illinois, another Google is reportedly behind in Nebraska, projects by ExxonMobil and Chevron, and another by Meta that has the option to add the technology. Axios said multiple interviews with top executives in both the energy and AI sectors indicate the technology is under active discussion.
KR Sridhar, co-founder and CEO of Bloom Energy, said big tech companies "will be the leaders in demonstrating carbon capture." He said that leadership would help the technology proliferate around the world and added, "I strongly believe carbon capture use and storage will be the only way we will decarbonize the planet in a big way over the next two decades." Sridhar said Bloom is in early talks with "many of the hyperscalers" about deploying carbon capture, but said the company will not talk about it publicly until next year.
Axios said the AI-driven surge in electricity demand is lifting nearly every type of energy technology, especially natural gas, but also more novel tech like fusion and long-term storage. That demand is the real engine here: the AI race is forcing more power generation, and the corporate answer is to bolt on another layer of industrial complexity rather than slow the machine down.
What the Companies Call a Solution
The article said oil and gas companies have led much of the early development of carbon capture, but high costs have limited its deployment. It said big tech companies are largely sticking to their climate goals even as AI drives a surge in power demand and political support for climate action weakens. Axios said that tension, combined with deep balance sheets, could finally push carbon capture toward broader use.
Alex Dewar, managing director at consulting firm BCG, said, "Unabated natural gas is being pursued by all the hyperscalers no matter how stringent their climate goals." He said companies are trying to square those goals with rising gas use and added, "That's where the solution does come back to carbon capture."
Cully Cavness, co-founder and president of data center developer Crusoe, said, "We have not yet announced a project ... but we're pursuing a bunch of opportunities." He added, "Ultimately, it's an economics question." Cavness also said, "Soon we'll be able to share something." Axios noted that Cavness has been making comments like this for the last couple of years without much to show for it publicly.
The Price Tag and the State's Role
The article said no natural gas plants operating in the U.S. have carbon capture equipment right now, though progress is slowly being made on a global level and domestically capturing carbon from different types of facilities. Michael Terrell, head of advanced energy at Google, said, "To be sure, it's still a technology that has a long way to go before it can be commercialized at scale." He added, "And we're committed to helping get it there."
Axios said gas plants paired with carbon capture could be especially attractive for data centers, which need reliable, around-the-clock power. It said BCG's recent analysis found they were the only option that scored consistently well across cost, speed, scalability and emissions among eight power sources. A separate report from the Great Plains Institute released last week identified the Gulf Coast, West Texas and Oklahoma as prime regions for this kind of buildout.
BCG estimates up to $80 billion could be spent on new gas plants and retrofits with carbon capture. The article said a federal tax credit, currently set to expire in 2033, is key to making projects viable, and that upfront capital costs remain high even with subsidies covering some operating expenses. Dewar said, "The biggest hurdle is the scale of the capital required." He added, "There's a lot more in the pipeline than what's publicly known. We're seeing a shift to smaller-scale projects to demonstrate the technology and scale up from there."
What emerges is a familiar arrangement: corporate expansion, public subsidy, and a technology still far from commercial scale, all moving forward because the power-hungry AI economy demands it. The people living near the plants, the data centers, and the infrastructure are left to absorb the consequences while the firms and consultants talk about balance sheets, scalability, and the next phase of the buildout.