A proposed 1-gigawatt artificial intelligence data center at the Savannah River Site is moving closer to construction after the U.S. Department of Energy selected Amentum to enter lease negotiations. The project, which pairs a massive data center with dedicated onsite power generation, signals a new wave of public-private collaboration on federal land - a trend with direct implications for construction timelines, utility infrastructure, and site selection in the real estate sector.
The National Nuclear Security Administration (NNSA) said in a statement that the project "pairs new AI infrastructure with dedicated onsite energy generation, helping ensure the project's electricity needs are met without shifting costs to existing utility customers." The plan calls for a 1-gigawatt data center and approximately 2 gigawatts of onsite energy generation, moving from natural gas to nuclear power over time.
Amentum, a firm that oversees logistics and technical services for many U.S. government agencies and specializes in nuclear security and cybersecurity, was chosen to enter negotiations for a phased lease. The government's selection does not guarantee the final lease award, but it marks a critical step toward breaking ground on the site.
Project scope and power generation
The scale of the planned energy generation - roughly twice the capacity of the data center itself - reflects the enormous power demands of modern AI computing. NNSA Administrator Brandon Williams said in September 2025 that "the work being carried out at NNSA to harness AI will have an impact on national security as the Manhattan Project did in the 1940s." By colocating the data center with its own power plant, the project aims to avoid drawing from the local grid and shifting costs to residential or commercial ratepayers.
The dual-fuel approach, starting with natural gas and transitioning to nuclear, is designed to provide immediate capacity while aligning with federal goals for advanced reactor deployment. The strategy also reduces the risk of construction delays tied to utility grid upgrades, a common bottleneck for large-scale data center projects.
Federal land as a development asset
The Savannah River Site, covering more than 300 square miles across three South Carolina counties, is one of four federal sites chosen for a program outlined in executive orders on accelerating data center permitting and deploying advanced nuclear reactors. The Department of Energy is using its extensive land holdings to attract private-sector investment in AI infrastructure, speeding up a process that might otherwise face years of local zoning fights and environmental reviews.
Data center construction is expanding rapidly as the global digital economy grows. For real estate and construction professionals, this trend is increasingly tied to AI-driven demands, as explored in resources on AI for Real Estate & Construction. The NNSA's role in the project underscores how federal agencies are integrating AI into their operations, a subject covered in AI for Government.
Why this matters for Real Estate and Construction
For developers and contractors, the SRS project demonstrates how federal land can become a fast track for large-scale AI infrastructure, bypassing the typical local opposition and permitting delays that have stalled similar projects in other markets. The colocation of power generation on-site also changes the calculus for site selection: a project of this size doesn't just need land - it needs energy independence, and that requirement is reshaping where data centers can be built. Firms that understand the intersection of federal land policy, utility infrastructure, and AI demand will be better positioned to compete for upcoming contracts, as the model used at SRS is likely to be replicated across other DOE sites.
Your membership also unlocks:
