Is the solution to the AI power shortage "retiring nuclear reactors"? The US launches military nuclear power plant recycling test

Artificial Intelligence(AI) Data Center Power Demand Surges, a Texas-based energy developer in the United States has proposed a unique plan to repurpose decommissioned U.S. Navy nuclear propulsion units as power sources. In the context of private data infrastructure requiring high-performance, 24/7 power supply, such solutions are gaining attention as new alternatives.

Led by HGP Smart Energy, the plan requests loan guarantees under the U.S. Department of Energy’s government energy finance program by submitting an application titled “Core Holding Project.” The specific proposal involves installing two decommissioned naval nuclear reactors near Oak Ridge National Laboratory in Tennessee to supply power to large-scale data centers.

The reactors utilized by HGP are pressurized water reactors that have provided decades of power to U.S. Navy aircraft carriers and submarines. These devices, manufactured by Westinghouse Electric and General Electric, have long proven their operational performance and safety, featuring highly sealed structural designs and high management efficiency. If implemented, each reactor could provide 450 MW and 520 MW of base load power, sufficient to support the growing electricity demands of AI training and inference tasks.

The total estimated investment ranges from $1.8 billion to $2.1 billion (approximately 2.592 trillion to 3.024 trillion KRW), with an investment cost of about $1 million to $4 million per MW. Analysts believe this approach is more cost-competitive than building new large-scale nuclear power plants or small modular reactors. However, regulatory hurdles remain for the project to materialize. Military reactors typically use highly enriched fuel and do not align with the current U.S. Nuclear Regulatory Commission’s civilian licensing system, requiring regulatory adjustments.

In response, HGP CEO Gregory Forero stated, “We have the professional capability to safely implement this large-scale project and have laid the foundation for cooperation with investors and partners.” In fact, as the U.S. power grid has recently undergone restructuring around AI computing, cloud companies and infrastructure operators are viewing nuclear power as a reliable, 24/7 carbon-neutral energy source. This is especially emerging as a solution to address climate change while ensuring power stability.

Meanwhile, the investment market is also reflecting this trend. X-energy, which is developing microreactors, raised $700 million in November (approximately 1.08 trillion KRW), and Radiant Industries, which is producing portable, mass-produced microreactors, has raised over $300 million this month (about 432 billion KRW). As AI infrastructure power and nuclear energy continue to intersect, its development is expected to depend on how regulation, technology, and finance can synergize into a trinity, leading to a pivotal turning point.