The US Department of Energy’s (DOE’s) National Nuclear Security Administration (NNSA) has selected global engineering firm Amentum to enter negotiations for a phased land lease at the Savannah River Site (SRS) in South Carolina. This public-private partnership aims to develop a 1 GWe artificial intelligence (AI) data centre powered by an integrated, co-located 2 GWe energy generation facility. It will initially utilise natural gas as a bridge fuel before permanently transitioning to advanced nuclear energy.
“Artificial intelligence, energy resilience, and national security are becoming increasingly interconnected, creating new opportunities to strengthen America’s strategic advantage,” Amentum CEO John Heller said. “Amentum is uniquely positioned at that intersection, combining deep engineering expertise, trusted mission execution, and decades of partnership with the US government. This agreement gives us the opportunity to partner with the NNSA to deliver infrastructure that will help power the technologies shaping our nation’s future.”
DOE also selected data centre developer DC BLOX as the primary digital infrastructure partner and lead developer within the Amentum-led consortium. DC BLOX, alongside Milliken & Company, Google and Nucor, is a founding member of the Palmetto Nuclear Coalition, launched in July 2025, dedicated to accelerating nuclear energy deployment across South Carolina.
While Amentum manages federal operations, technical integration, and overall consortium leadership, DC BLOX provides the direct technical blueprint for the technology campus. The two companies will now enter negotiations for a phased lease for the site, which, if successful, will lay the foundation for the campus development.
Selection for negotiations does not constitute a final lease award. Any agreement will be subject to successful negotiations and all applicable permitting requirements, safety and security evaluations, and other federal approvals.
In April 2025, DOE identified 16 federal sites that could support the construction of AI data centres and associated energy infrastructure. SRS and three other locations – Idaho National Laboratory, Oak Ridge Reservation, and Paducah Gaseous Diffusion Plant – were selected to move forward with efforts to invite private-sector development.
Operating under the federal “Ratepayer Protection Pledge”, the co-located power generation ensures the massive electrical demand of the AI centre is met entirely on-site. This prevents price spikes or grid strain for existing local utility customers. It directs the Environmental Protection Agency (EPA) and the Nuclear Regulatory Commission (NRC) to revamp restrictive rules to support independent power providers operating adjacent to high-demand computing facilities.
The SRS project supports the Trump administration’s goals of utilising Federal lands to lower energy costs and help power the global AI race, as outlined in a series of Executive Orders (EOs).
EO 14318 (July 2025), Accelerating Federal Permitting of Data Center Infrastructure bypasses legacy red tape to rapidly scale up critical AI processing hubs. It directs the Departments of Energy, Commerce, and the Interior to unlock underutilised federal lands for private co-location data centres. It instructs agencies to aggressively reduce timeframes for National Environmental Policy Act (NEPA) and water permitting processes and authorises the federal government to offer structured loans, grants, and offtake guarantees to support private digital infrastructure layouts.
EO 14299 (May 2025), Deploying Advanced Nuclear Reactor Technologies for National Security uses national security mechanisms to speed up the rollout of advanced nuclear designs. It grants the Secretary of Energy power to designate specialised AI data centres located at or coordinated with DOE facilities as “critical defense facilities” shielding them from certain standard public utility delays. It mandates DOE to use all legal pathways to site and build privately funded advanced reactors at federal hubs, targeting operation of the first commercial advanced reactor on-site within 30 months.
EO 14154 (January 2025), Unleashing American Energy, prioritises fossil fuel production, fast-tracks traditional energy projects, and pauses certain climate-related fund disbursements. It targets energy dominance by combining fossil-fuel bridging with next-generation power infrastructure.
The Amentum selection at SRS is a direct, practical application of this combined policy architecture. By utilising a 310-square-mile federal nuclear reservation, the project capitalises on EO 14318 to fast-track data centre environmental reviews, leverages EO 14299 to securely deploy advanced nuclear power directly onto a protected DOE site and uses EO 14154 to build a 2 GWe on-site gas-to-nuclear generation bridge that protects South Carolina electricity consumers.
“This proposed partnership represents an opportunity to strengthen America’s leadership in artificial intelligence, expand reliable energy generation, and strengthen our national security,” said NNSA Administrator Brandon Williams. “By working with the private sector, we can move faster, apply innovative technologies, and make productive use of federal land while maintaining our commitment to mission delivery.”
The Savannah River Site is a 310-square-mile (803-square-kilometre) site in Aiken, South Carolina, which was focused on the production of plutonium and tritium for use in the manufacture of nuclear weapons from the early 1950s until the end of the Cold War. In 1992, the focus at SRS turned to environmental cleanup, nuclear materials management, and research and development activities. It is home to Savannah River National Laboratory, NNSA Savannah River Field Office, the University of Georgia Savannah River Ecology Laboratory, and the US Department of Agriculture Forest Service Savannah River along with several major contractors. Existing facilities in the site include L Area where used nuclear fuel from domestic and foreign research reactors is stored. Ten tracts of land totalling 3,103 acres (1,256 hectares) have been identified for AI infrastructure projects.
