The U.S. Department of Energy has approved a groundbreaking approach to powering AI data centers: start with natural gas, then transition to advanced nuclear reactors as they become available. This "gas-to-nuclear bridge" model addresses a fundamental problem in the energy sector: AI computing infrastructure needs power immediately, but building and licensing nuclear reactors takes years. By pairing immediate generation with long-term nuclear deployment, the strategy lets data centers operate while nuclear technology matures through licensing, financing, and construction phases.

What Is the Gas-to-Nuclear Bridge Model?

Amentum, a major infrastructure and engineering firm, was selected by the Department of Energy's National Nuclear Security Administration to negotiate a phased lease at the Savannah River Site in South Carolina. The proposal centers on a 1-gigawatt AI data center paired with approximately 2 gigawatts of on-site power generation. Initially, natural gas turbines would supply the electricity. As advanced nuclear reactors are built, licensed, and brought online, the facility would gradually shift to nuclear power, with gas equipment eventually serving backup and resilience functions.

This model separates the timeline for "first power" from the longer-term energy mix. Rather than waiting for a reactor to be constructed and licensed before opening a data center, companies can begin operations immediately while nuclear infrastructure develops in parallel. The Department of Energy emphasized that dedicated on-site power could meet the data center's needs without shifting costs to existing utility customers, potentially freeing up grid capacity for other uses.

Why Are Multiple Companies Pursuing This Strategy?

The Savannah River proposal is not an isolated experiment. Similar gas-to-nuclear projects are advancing across the country, signaling that this approach is becoming a repeatable development model for the nuclear and AI energy sectors.

• Blue Energy and GE Vernova in Texas: The companies unveiled a proposed 2.5-gigawatt project where two GE Vernova gas turbines could provide approximately 1 gigawatt as early as 2030, with small modular reactors ramping up to approximately 1.5 gigawatts by 2032. They are targeting a final investment decision in 2027.
• Oklo's Multiple Partnerships: The advanced reactor developer has pursued the same strategy through partnerships with RPower and Liberty Energy, combining immediate gas power with future deployment of their Aurora reactor technology.
• Shared Investment Logic: A large customer load, such as an AI data center, can support site work, equipment reservations, licensing, and financing before a reactor begins operation, creating commercial activity across the nuclear supply chain years before reactors generate power.

Together, these projects demonstrate that the gas-to-nuclear model is emerging as a standard approach for pairing AI infrastructure with advanced nuclear energy. The model works because it aligns the fast timeline of data center deployment with the slower, more complex timeline of nuclear reactor development.

How Does This Benefit the Nuclear Supply Chain?

If projects like Savannah River advance, they could create earlier commercial opportunities across the nuclear industry. Engineering firms, site developers, equipment manufacturers, and project services companies can begin work before reactors are operational, generating revenue and building expertise while nuclear technology matures.

Companies positioned at different points in this development model stand to benefit. Amentum brings nuclear operations, infrastructure development, and program delivery experience. GE Vernova participates through both gas turbines and small modular reactor technology, specifically their BWRX-300 design. Oklo offers exposure to advanced reactors and new commercial pathways. This diversification across the nuclear value chain reduces dependence on any single project or deployment schedule, creating multiple revenue streams as the sector expands.

What Regulatory Progress Is Accelerating Advanced Nuclear Technology?

Beyond the gas-to-nuclear projects, the regulatory landscape for next-generation nuclear technology is shifting rapidly. In a historic first, the Department of Energy approved a Nuclear Safety Design Agreement for a molten salt reactor under development by researchers at Abilene Christian University in Texas. This approval represents a major milestone for advanced reactor technology and signals federal commitment to commercializing innovative nuclear designs.

Molten salt reactors (MSRs) represent a fundamentally different approach to nuclear power. Unlike traditional reactors that use water for cooling, MSRs use molten salt as both the fuel medium and coolant. This design offers several advantages: they require less fuel, produce shorter-lived radioactive waste, and can process fuel online without lengthy refueling outages. The molten salt approach also addresses a critical vulnerability of traditional nuclear plants: water stress in dry regions. France recently took multiple reactors offline during a heat wave when river temperatures became too high for safe cooling, a problem MSRs would avoid.

The Abilene Christian University approval establishes "baseline parameters required for federal authorization of facility construction and system testing," according to recent reports. Oak Ridge National Laboratory has made critical advances in modeling molten salt behavior to improve reactor design for practical application. These developments reflect a broader Trump administration push to "produce lasting American dominance in the global nuclear energy market" through accelerated testing and commercialization of advanced nuclear technologies.

What Are the Key Takeaways for Energy and AI Infrastructure?

The convergence of AI data center demand, natural gas infrastructure, and advanced nuclear development is reshaping how the U.S. approaches energy security. The gas-to-nuclear bridge model solves a critical timing problem: it allows companies to deploy AI infrastructure immediately while nuclear reactors move through the lengthy process of licensing, financing, and construction. This approach is no longer theoretical; it is being implemented at federal sites with Department of Energy backing.

Simultaneously, regulatory approval of molten salt reactor designs signals that next-generation nuclear technology is moving from research labs toward commercial deployment. These two trends, combined, suggest that the U.S. nuclear sector is positioning itself to meet the massive power demands of artificial intelligence while advancing cleaner, safer reactor designs.

The Savannah River Site lease negotiation remains subject to permitting, safety reviews, security approvals, and other regulatory steps, so the project is not yet finalized. However, the fact that the Department of Energy selected Amentum and is advancing this model across multiple sites indicates that gas-to-nuclear is becoming a standard playbook for pairing AI infrastructure with long-term nuclear deployment. For investors and energy professionals, this represents a shift from viewing gas and nuclear as competing technologies to seeing them as complementary phases in a coordinated infrastructure strategy.