The US Department of Energy (DOE) has detailed 16 possible sites for the development of data centers on federal land.
The DOE has released a Request for Information (RFI) about the sites, and said that it will help inform the department on whether it will add further sites. The locations include in-place energy infrastructure with the ability to fast-track permitting for new energy generation, such as nuclear.
The RFI hopes to lead to the construction of AI infrastructure at select DOE sites with a target of commencing operation by the end of 2027. Companies have a 30-day window to respond to the RFI.
“The global race for AI dominance is the next Manhattan project, and with President Trump’s leadership and the innovation of our National Labs, the United States can and will win,” Secretary of Energy Chris Wright said.
“With today’s action, the Department of Energy is taking important steps to leverage our domestic resources to power the AI revolution, while continuing to deliver affordable, reliable, and secure energy to the American people.”
The move to open up federal land for data centers was begun during the Biden Administration, and continued with President Trump‘s Removing Barriers to American Leadership in Artificial Intelligence and Unleashing American Energy Executive Orders.
White House Office of Science and Technology Policy Director Michael Kratsios said: “The Trump Administration will unleash Federal resources to build out the data resources needed for an AI-powered future.”
The 16 sites are as follows:
- Idaho National Laboratory
- Paducah Gaseous Diffusion Plant
- Portsmouth Gaseous Diffusion Plant
- Argonne National Laboratory
- Brookhaven National Laboratory
- Fermi National Accelerator Laboratory
- National Energy Technology Laboratory
- National Renewable Energy Laboratory
- Oak Ridge National Laboratory
- Pacific Northwest National Laboratory
- Princeton Plasma Physics Laboratory
- Los Alamos National Laboratory
- Sandia National Laboratories
- Savannah River Site
- Pantex Plant
- Kansas City National Security Campus
The RFI varies in the level of detail given about each site. Some include specific locations within the federal land, information on grid availability, and water capacity, while others are a little broader.
Of note is Argonne National Laboratory, which the RFI notes "could accommodate a 110-acre developable site for a future 1,000MW AI data park on US Department of Energy (DOE) land with an early target for operations by 2028."
The land is 23 miles from Chicago, the facility would benefit from Illinois’ data center tax exemption that has helped push Chicago to its position as the fourth largest data center market in the US by capacity.
Over in New York state, Brookhaven National Laboratory is in close proximity to the existing Caithness Long Island Energy Center, "which could be the site of a new 750MW gas turbine plant for a data center."
The Colorado-based NREL Flatirons Campus "has enough land, power, water, and broadband capability to host a 100MW data center that could be initiated as soon as this year (2025)," the RFI notes. "The site could support an integrated data center energy system test bed, that could be deployed later at scale at other locations."
Another smaller deployment could be found in New Jersey at the Princeton Plasma Physics Laboratory. Its Forrestal Campus "currently has 100MW of energy capacity with district upgrade potential available, and current water contract with NJ Water Supply Authority includes ~55 million gallons/year."
Home to some of the world's largest supercomputers, the Los Alamos National Laboratory in New Mexico is currently working on its own data center deployments.
LANL said in the RFI: "LANL is executing an upgrade of the Strategic Computing Complex (SCC) to 70MW, which requires the Electrical Power Capacity Upgrade (EPCU) and the SCC Electrical Upgrade (SEU) GPP; these projects are funded and near start of construction currently.
"LANL’s strategy is to leverage new off-premises power and water capabilities to supplement its enduring on-premises capabilities. In responding to this call, LANL recognizes that this new on-premises commercial data center would expand our mission further, and without new power sources or exercising EPCU options, would be limited to a total of 100MW for HPC+AI infrastructure.
"This on-premises 100MW limit would need to be operationally managed with the newly upgraded SCC (70MW) mission. The SCC could host low-density data systems, reducing its peak power needs to below current levels, and still take advantage of 70MW total capability by re-configuring its electrical distribution back to its original configuration of 2N power. This would enable shifting power from the SCC to the new AI facility. A better approach would be to identify and deploy new on-premises power sources such as gas turbine (exercise options to expand the existing steam plant), or nuclear small modular reactors."
In March 2025, DCD exclusively reported that the US government was similarly assessing interest in developing a 500MW data center on an Air Force base in Tucson, Arizona.
