President Donald Trump said Monday that his administration is now approving energy facilities to power AI data centers in "a matter of weeks," responding to what he called shocking demands from artificial intelligence development. The announcement comes as a new report tallies plans for 74 new or expanded methane gas plants across the U.S. dedicated to data centers, together capable of generating 143 gigawatts of electricity and emitting nearly 662 million tons of greenhouse gases each year.
Fast-track approvals and industry reaction
Trump said Big Tech leaders racing to advance AI told him they need double the country's existing energy capacity to outpace foreign competitors. He added that EPA Administrator Lee Zeldin informed him the companies were not using the administration's promise of expedited approvals for private power plants. Trump then called Mark Zuckerberg of Meta, Amazon's Jeff Bezos, OpenAI's Sam Altman and SpaceX's Elon Musk to ask why they had not submitted plans alongside data center developments.
"They thought we were kidding," Trump told reporters. "They can't believe it, that they're approved in a period of a matter of weeks." The White House did not respond to questions about how the administration is approving power plant plans in weeks given that state and local requirements routinely take months even under the most favorable permitting conditions. The president said tech companies may use any energy source except wind, which he described as "terrible, it just doesn't work." Wind generates roughly a tenth of U.S. electricity.
Surge in gas power plants for data centers
The Environmental Integrity Project, a nonprofit founded by a former EPA enforcement director, published the findings on the planned gas-fired plants. Of the 74 projects, 32 are in Texas, 10 in Ohio and seven in Pennsylvania. The plants would produce 143 gigawatts-enough to power California nearly three times over-and release air pollutants that contribute to smog and lung damage along with the equivalent of Australia's annual greenhouse emissions.
Jen Duggan, the group's executive director, said in a statement that "an industry of the future should not be chained to dirty fuels of the past." She added that while data centers may be needed, "the public has a right to transparency and accountability, clean air, and common sense controls to protect water supplies, especially in areas already struggling with water shortages." Data centers can use large volumes of water for cooling.
Local pushback and state moratoriums
Rural communities have become the main target for data center developments and their accompanying fossil fuel power plants, and the facilities have quickly turned unpopular. U.S. Sen. Bernie Sanders and Rep. Alexandria Ocasio-Cortez introduced legislation in March proposing a halt to new data center construction until worker and environmental safeguards are in place. Texas Gov. Greg Abbott, a Republican facing midterm voters, went a step further last week, calling for blocking new data center projects in rural parts of the state. The New York State Legislature passed a one-year moratorium in June, though Gov. Kathy Hochul has said she believes municipalities should decide locally. Monterey Park, California, and Ashville, Ohio, have already enacted temporary bans or pauses.
The Trump administration said last month it will not set nationwide environmental requirements for the data center industry. At a Politico energy summit, Zeldin argued that states and communities know what works best for them. Clara Vondrich, senior policy counsel with Public Citizen's Climate Program, countered that "Big Tech executives have lobbied hard to ingratiate themselves into the Trump administration's orbit. … Zeldin made clear that their investment was money well spent."
Why this matters for government, IT and development professionals
For those in AI for Government, the patchwork of state moratoriums and the absence of federal standards creates a fragmented regulatory environment that challenges long-term infrastructure and compliance planning. The fast-tracking of energy permits may accelerate deployment, but it also heightens legal and environmental risks at the local level. At the same time, the push to build dedicated power plants directly affects AI for IT & Development: data center siting, energy availability and cooling constraints increasingly shape how and where AI workloads can run, pressuring technical teams to factor power infrastructure into architecture decisions from the start.
Your membership also unlocks:
