The U.S. Department of Energy is betting big on nuclear power to keep America's AI ambitions plugged in. On Tuesday, the agency announced $17.5 billion in conditional loans for utilities and energy companies to purchase critical components needed to rebuild the domestic nuclear supply chain. The move signals a dramatic shift in how the government plans to meet soaring electricity demand from data centers powering artificial intelligence, with tech hyperscalers now emerging as key drivers of nuclear investment.

Why Is Nuclear Power Suddenly Critical for AI?

Data centers running AI models consume enormous amounts of electricity around the clock. Unlike solar or wind farms, nuclear plants provide what energy experts call "baseload power," meaning they generate consistent electricity 24/7 without weather interruptions. As companies like Microsoft, Google, and others race to build massive AI infrastructure, they're discovering that traditional power sources can't keep pace. Energy Secretary Chris Wright told reporters that the loans have attracted "strong interest from data center hyperscalers," the tech giants that operate global cloud and computing infrastructure.

The timing matters. The U.S. electricity grid is already strained, and AI's explosive growth is making the problem worse. By offering loans to help companies secure hard-to-find reactor components, the government is essentially removing a major bottleneck that has slowed nuclear development for decades. Components like reactor vessels and steam generators can take years to procure, and the new loans will help companies lock in fixed prices now.

How Will These Loans Actually Work?

• Loan Structure: The Department of Energy's Office of Energy Dominance Financing will back up to five loans, each supporting two 1.1-gigawatt Westinghouse reactors per site, with each project jointly owned by Westinghouse and a utility or energy company.
• Cost Sharing: Each utility or energy company must commit $500 million of their own capital before accessing federal funds, ensuring skin-in-the-game accountability from private partners.
• Supply Chain Focus: The loans specifically target "long-lead" items that take years to manufacture, including reactor vessels, steam generators, and other critical components that have historically delayed nuclear projects.
• Fixed Pricing: Westinghouse will partner with utilities to procure reactors and supply chain needs at fixed prices, protecting companies from cost overruns that have plagued past nuclear projects.

Seven utilities have already expressed interest in the program, though the Energy Department has not disclosed their names or project locations. Each project aims to deliver two 1.1-gigawatt reactors, which would generate enough power to support massive data center operations or supply electricity to millions of homes.

What Does This Mean for Nuclear's Future?

The announcement comes as part of a broader nuclear revival. Three shuttered nuclear plants are on track to restart in the coming years, including Three Mile Island in Pennsylvania, Palisades in Michigan, and Duane Arnold in Iowa. These restarts, combined with new construction, represent the most significant nuclear expansion in the U.S. in decades.

Energy Secretary Wright said the government expects these plants' timing and costs to "well outperform what was done on Vogtle," referring to Georgia's Vogtle Electric Generating Plant, which experienced significant delays and budget overruns. The new loan program is designed to prevent those problems by locking in supply chain certainty early.

"The money serves as a catalyst for nuclear, providing the certainty needed to enhance the domestic nuclear supply chain and accelerate construction of nuclear projects that will deliver reliable baseload power around the country for decades to come," stated Connor Teskey, CEO of Brookfield Asset Management, which co-owns Westinghouse Electric.

Connor Teskey, CEO, Brookfield Asset Management

The loans also reflect the Trump administration's aggressive nuclear ambitions. President Trump aims to quadruple U.S. nuclear power capacity to 400 gigawatts by 2050, a target that would require sustained investment and regulatory streamlining. The $17.5 billion announcement is one piece of that puzzle, but it also signals that private companies, not just government, are willing to bet on nuclear's role in powering the AI era.

The nuclear industry has historically struggled to attract investment because projects are capital-intensive, prone to cost overruns, and subject to complex regulations, making them riskier than cheaper alternatives like natural gas and renewables. By reducing supply chain uncertainty and offering federal financing, the government is attempting to level the playing field. Wright emphasized that "we are confident that these projects will be economic for utility shareholders, ratepayers and hyperscalers," suggesting the loans are structured to benefit all stakeholders.

Wright

For data center operators and AI companies, this announcement represents a major step toward solving one of their most pressing challenges: securing enough clean, reliable power to fuel the next generation of AI infrastructure. As competition for electricity intensifies, nuclear power is no longer a niche energy source; it's becoming central to America's AI strategy.