The U.S. Department of Energy in late July announced that it has selected four department-owned sites—including the Idaho National Laboratory, east of Arco—to move forward with developing large new artificial intelligence data centers.
The other federal sites nominated for development are the Oak Ridge Reservation in Tennessee, the Paducah Gaseous Diffusion Plant in Kentucky and the Savannah River Site in South Carolina.
“DOE looks forward to working with data center developers, energy companies, and the broader public in consultation with states, local governments, and federally recognized tribes that these projects will serve to further advance this important initiative,” the DOE stated.
U.S. Secretary of Energy Chris Wright referred to the plan as the “next Manhattan Project.”
Each federal site is “uniquely positioned to host data centers as well as power generation to bolster grid reliability, strengthen our national security, and reduce energy costs,” he said.
Earlier this year, on April 7, the Department of Energy published a formal request for information, or RFI, to gauge market interest in building data centers on federal land.
The RFI, issued in response to Trump’s Jan. 23 executive order to expedite AI data center development, listed 16 possible sites, including the INL, and asked for input on design, power, cooling and economic feasibility.
According to the solicitation document, the INL site stood out because “Idaho regulatory and tax structures also are favorable towards ambitious projects that seek to advance U.S. global leadership.”
The DOE said in the RFI that it aimed to break ground on select sites “by the end of 2025, with a target of commencing operation by the end of 2027.”
OpenAI responded with interest. In a May 7 letter to the DOE, the prominent AI developer outlined its site needs—at least 300 acres, low seismic risk, adequate water supply, highway and rail links, and skilled labor—while asking for further “topographic and geotechnical reports, as well as regional workforce information.”
OpenAI’s in-house lobbyist, Chan Park, also pushed for federal tax breaks and low-interest loans.
“With the Chinese Communist Party (CCP) making fast progress toward its goal of leading the world in AI by 2030, we believe that AI built on a foundation of American soil, American ingenuity, and democratic principles can continue to prevail over CCP-built AI,” Park wrote.
The DOE said last month that it received “enormous” interest in response to the RFI.
The agency said it would release “more details regarding project scope, eligibility requirements, and submission guidelines” for each named site “in the coming months,” and “partners could be selected by the end of the year.”
The July 24 announcement came one day after the Trump Administration released its new “AI Action Plan,” a 28-page document that vows to fast-track the environmental permitting process for data centers, increase semiconductor manufacturing and export American AI tech.
The action plan was released alongside two other AI-related executive orders signed by Trump on July 23, including an order preventing “woke AI” in the federal government and a directive to EPA Administrator Lee Zeldin to roll back or modify parts of the Clean Air Act, the Clean Water Act, the Toxic Substances Control Act and the federal Superfund program to accelerate AI data center construction.
According to an April 2025 report from the International Energy Agency, electricity demand from data centers across the globe is on track to “more than double by 2030 to around 945 terawatt-hours (TWh), slightly more than the entire electricity consumption of Japan today.”
Power consumption by U.S. data centers should “account for almost half of the growth in electricity demand between now and 2030,” the report added. 
