In a significant policy move aimed at bolstering U.S. artificial intelligence capabilities, President Donald Trump has signed a series of executive orders designed to double the nation’s nuclear energy development over the next 25 years. These directives form a cornerstone of his strategy to secure high-density, resilient power sources to meet the surging energy demands of AI technologies and to compete globally, particularly with China.
The new orders specifically call for the rapid construction and deployment of advanced nuclear reactors, classifying them as essential military electric infrastructure. These next-generation facilities are expected to serve as stable, secure energy sources immune to grid disruptions and external threats – qualities deemed critical for powering the data-intensive computing operations that underpin modern AI applications.
During the signing ceremony at the White House, Secretary of the Interior Doug Burgum emphasized the urgency and strategic value of this energy shift. “We have enough electricity to win the AI arms race with China,” said Burgum. “For the first time in history, electricity can be directly converted into intelligence. What we do in the next five years will determine the next fifty.”
One of the directives tasks Secretary of Energy Chris Wright with establishing AI data centers in 48 states and Washington, D.C., intentionally excluding Alaska and Hawaii. These centers will be operated in conjunction with Department of Energy (DoE) facilities and designated as vital to national defense infrastructure. The aim is to embed AI computing power within the existing federal energy framework, ensuring that it is both secure and scalable.
Chris Wright has also been directed to identify at least one DoE facility for the deployment of privately funded advanced nuclear technologies. These include Generation III+ reactors, small modular reactors (SMRs), microreactors, and both stationary and mobile systems. According to the order, these systems have the potential to deliver “resilient, secure, and reliable power” not just to AI data centers but also to broader defense infrastructure.
Department of Energy
One of the more controversial aspects of the executive orders involves the restructuring of the Nuclear Regulatory Commission. The NRC, which has overseen nuclear regulation in the U.S. for five decades, will see some of its powers redistributed to accelerate the approval of novel reactor designs. Burgum framed this as a landmark shift, stating, “Put this day on your calendar because it will reverse more than 50 years of excessive industrial regulation.”
The push for nuclear-powered AI data center infrastructure follows earlier Trump-era directives encouraging the use of coal to power data centers. The Department of Energy estimates that national electricity demand will rise between 6% and 12% by 2028, driven in large part by AI workloads. Currently, nuclear energy accounts for 19% of U.S. power generation, with fossil fuels providing 60% and renewables making up 21%, according to the U.S. Energy Information Administration.
Using nuclear energy to power AI is not a new concept. Earlier this year, Representative Gary Palmer (R-Ala.) proposed using SMRs as a scalable solution for data center power. At an April congressional hearing, Palmer noted that such reactors could power AI infrastructure for over a century but emphasized the need for regulatory reform to enable their deployment.
Despite the ambitious goals, questions remain about the practicality of implementing the directives. The U.S. nuclear sector has faced headwinds in recent years, with numerous plant closures and regulatory hurdles hampering growth.
Judi Greenland, President and CEO of the Nuclear Innovation Alliance, welcomed the administration’s renewed focus on nuclear energy but flagged concerns over capacity. “We applaud the Administration’s efforts to advance nuclear power development and deployment. However, achieving these goals will require adequate staffing and funding. Recent workforce reductions and planned budget cuts at the DoE risk undermining the very efforts these executive orders are intended to support.”
