What's happening
The U.S. Department of Energy's National Nuclear Security Administration announced on July 20, 2026 that Amentum Holdings (NYSE: AMTM) has been selected to enter phased lease negotiations for a 1-gigawatt AI data center at the Savannah River Site in South Carolina. The facility would be supported by approximately 2 gigawatts of on-site energy generation capacity, structured as a public-private partnership between Amentum and the federal government. The project is designed in phases, beginning with natural gas-based power generation before transitioning to advanced nuclear energy technologies as those systems become commercially available and deployable at the site.
Amentum Chief Executive Officer John Heller stated: 'Artificial intelligence, energy resilience, and national security are becoming increasingly interconnected, creating new opportunities to strengthen America's strategic advantage.' Mark Whitney, president of Amentum's Energy & Environment business, added: 'For generations, Amentum has helped solve some of the nation's most complex engineering and nuclear challenges, including at the Savannah River Site.' Amentum, which carries a market capitalization of approximately $5.17 billion and reported revenue of $14.20 billion, provides engineering, technical, and mission support services to U.S. government agencies including the Department of Defense. The Savannah River Site has a long history as a federal nuclear facility, making it a logistically and regulatorily established location for this type of advanced energy and computing infrastructure.
Why it matters for markets
The scale of the proposed facility — 1 gigawatt of AI data center capacity backed by approximately 2 gigawatts of on-site generation — places it among the largest announced public-private AI infrastructure projects tied to a federal site. For Amentum, which operates with a price-to-earnings ratio of 35.3 and a 52-week trading range of $19.80 to $38.11, a long-term federal lease of this magnitude would represent a structurally significant addition to its government services portfolio. The phased approach, starting with natural gas and transitioning to advanced nuclear, reflects the broader challenge facing AI infrastructure developers: sourcing sufficient, reliable, and increasingly carbon-constrained power at the gigawatt scale required by modern GPU-dense compute environments.
The project's planned transition to advanced nuclear energy directly engages the commercial small modular reactor and advanced nuclear sector, which has been positioning itself as the long-term power solution for hyperscale AI compute. The intersection of national security mandates, AI compute demand, and nuclear energy policy embedded in this single project illustrates the convergence of federal priorities that is reshaping how large-scale digital infrastructure is financed and sited in the United States. The use of an existing federal nuclear site with established regulatory infrastructure at Savannah River may reduce permitting timelines compared to greenfield advanced nuclear deployments, a factor that has historically constrained the sector's commercialization pace.
For the broader AI infrastructure supply chain, a 1-gigawatt data center represents substantial downstream demand for GPU accelerators, power management systems, thermal cooling, and cloud compute capacity. NVIDIA's (NASDAQ: NVDA) H100 and A100 data center accelerators, with the company reporting $253.49 billion in revenue, and Vertiv Holdings' (NYSE: VRT) power management and thermal systems — the company reported $10.84 billion in revenue — are among the categories of infrastructure that facilities of this scale require in significant volume.
Sectors and assets to watch
The most directly affected ticker is Amentum (NYSE: AMTM), which as the selected party for lease negotiations holds the primary contractual relationship with the NNSA. The advanced nuclear transition component of the project places advanced SMR developers in a relevant position: Oklo Inc. (NYSE: OKLO), which develops the Aurora small modular reactor targeting data center and industrial power customers, and NuScale Power (NYSE: SMR), whose 77 MWe pressurized water reactor modules are designed for scalable, factory-built nuclear deployment, are both active in the commercial SMR market that a federal site transition of this kind would draw upon. Constellation Energy (NASDAQ: CEG), which operates the largest nuclear fleet in the United States with $29.87 billion in revenue, represents the established nuclear power sector that has already begun pursuing data center power agreements.
On the compute and infrastructure side, NVIDIA (NASDAQ: NVDA) supplies the GPU accelerator hardware — including H100 and A100 clusters — that underpins AI data center deployments at this scale. Vertiv Holdings (NYSE: VRT), with its power management, UPS, and precision cooling product lines serving hyperscale and colocation environments, is a key infrastructure supplier category for gigawatt-class facilities. CoreWeave (NASDAQ: CRWV), which operates GPU-accelerated cloud infrastructure built on NVIDIA hardware and reported $6.23 billion in revenue, represents the cloud compute layer that increasingly sources capacity from large, purpose-built AI data centers, making federal-site projects of this type relevant to its supply and competitive landscape.
What to watch next
Key developments to monitor include the outcome and timeline of the phased lease negotiations between Amentum and the NNSA, which will determine the contractual structure, financial terms, and construction schedule for the Savannah River facility. The specific advanced nuclear technologies selected for the site's transition phase — and whether Amentum partners with existing SMR developers such as Oklo or NuScale — will be a critical signal for the commercial advanced nuclear sector. Regulatory milestones at the Nuclear Regulatory Commission for any SMR deployment at or adjacent to the Savannah River Site will also warrant close attention, as licensing timelines remain a primary variable in advanced nuclear commercialization. Additionally, any further NNSA or DOE announcements regarding similar public-private AI infrastructure leases at other federal sites would indicate whether the Savannah River model is being replicated as a broader federal policy framework.