The Nuclear-AI Power Play: Why Data Centers Are Moving to Reactor Sites
The U.S. Department of Energy is fast-tracking partnerships that place artificial intelligence data centers directly at nuclear sites, combining energy-hungry computing infrastructure with next-generation small modular reactors (SMRs) to address America's looming power shortage. Two major projects announced this week signal a strategic shift: the DOE is investing in AI data center campuses at the Paducah and Savannah River sites, paired with advanced nuclear energy infrastructure designed to power the computational demands of modern AI systems.
This convergence reflects a fundamental reality facing the tech industry. U.S. electricity demand is forecast to soar in the coming years driven by consumer needs, data center growth, increased AI use, and the industrial sector's need for constant power. Small modular reactors could provide reliable power for these energy-intensive sectors, with the added benefit of flexible deployment thanks to their compact size and modular design.
What Makes Small Modular Reactors Ideal for Data Centers?
Small modular reactors represent a departure from traditional nuclear plants. Unlike conventional reactors that require massive infrastructure and decades to build, SMRs are designed for faster deployment and can be manufactured in factories, then transported to sites. Light-water small modular reactors can also leverage the existing service and supply chain sectors supporting the country's current fleet of light-water reactors, helping to speed up near-term deployment of new nuclear reactors.
The DOE's Generation III+ Small Modular Reactor Program is accelerating this transition. In March 2025, the agency reissued a $900 million solicitation to deploy Gen III+ SMRs, with funding available through two tiers designed to support both pioneering projects and follow-on deployments.
How Are Federal Investments Supporting SMR Deployment?
- Tier 1 Funding: Two major projects received $400 million each. The Tennessee Valley Authority plans to advance deployment of a GE Vernova Hitachi BWRX-300 at the Clinch River Nuclear site in Tennessee, while Holtec plans to deploy two SMR-300 reactors at the Palisades Nuclear Generating Station site in Michigan.
- Tier 2 Supply Chain Support: The DOE allocated approximately $100 million to address gaps in design, licensing, supply chain, and site preparation, funding projects like fuel fabrication facility expansions and large component manufacturing equipment.
- Site Preparation Grants: Constellation SMR Development and Nebraska Public Power District each received tens of millions to pursue Nuclear Regulatory Commission approved Early Site Permits, laying groundwork for future SMR deployments.
The Paducah partnership exemplifies this strategy. The DOE announced a partnership with Big Rivers Electric Power Corporation, Brookfield Corporation, Jackson Purchase Energy Cooperative, NextEra Energy, and Paducah Power System aimed at redeveloping parts of the former Paducah Gaseous Diffusion Plant into a data center campus with energy infrastructure. This approach transforms abandoned nuclear sites into modern computing hubs, leveraging existing infrastructure and expertise.
Similarly, ARC Clean Technology entered into a Strategic Partnership Project agreement with Battelle Energy Alliance, the management and operating contractor of Idaho National Laboratory. That agreement establishes a multiyear framework for ARC and INL to collaborate on the design, demonstration, and first deployment of the company's ARC-100, a 100-megawatt sodium-cooled fast reactor.
Why Is This Partnership Model Gaining Momentum?
The pairing of AI data centers with nuclear power addresses a critical bottleneck. Data centers require constant, reliable electricity to operate continuously, and AI systems demand even more power than traditional computing workloads. Nuclear power provides carbon-free baseload electricity without the intermittency challenges of renewable sources like wind and solar. By co-locating data centers at nuclear sites, companies can secure long-term power agreements while reducing transmission losses and grid strain.
The federal government is also encouraging private-sector participation through voluntary commitments. In March, seven tech companies and hyperscalers signed on to President Trump's Ratepayer Protection Pledge, a voluntary agreement stating that in the pursuit of data center growth, signatories would not shift costs to electricity customers in the form of higher electricity bills. This pledge signals industry recognition that sustainable power solutions must not burden consumers.
The DOE's broader nuclear strategy reflects confidence in SMR technology. The Generation III+ SMR Program aims to speed the demonstration of Gen III+ nuclear reactor technologies, bridging the gap between the nation's current fleet of reactors and more advanced reactor designs that will be demonstrated through the Advanced Reactor Demonstrations Program. By funding both first-mover teams and fast-follower deployments, the government is creating a competitive ecosystem that accelerates innovation while reducing risk for individual projects.
Beyond data centers, nuclear innovation is expanding into new domains. The Department of Transportation's Maritime Administration signed a first-of-a-kind memorandum of cooperation with the Port of Long Beach to advance the development of small modular reactors for powering commercial vessels, ports, and other maritime assets, making Long Beach the first U.S. seaport to form such a partnership. Meanwhile, five states, Idaho, Louisiana, Oklahoma, Tennessee, and Utah, were named finalists to potentially host Nuclear Lifecycle Innovation Campuses, envisioned as federal-state partnerships to support work across the nuclear fuel life cycle.
The convergence of AI and nuclear power represents more than a technical solution to an energy problem. It reflects a strategic recognition that meeting future electricity demand requires deploying multiple technologies in concert. As data centers continue to expand and AI workloads grow, the nuclear-AI partnership model offers a pathway to reliable, clean power that can sustain both technological innovation and grid stability.