House hearing spotlights nuclear power as primary option to meet surging AI data‑center demand
Get AI-powered insights, summaries, and transcripts
Subscribe
Summary
Witnesses and members at a House Science Committee subcommittee hearing emphasized nuclear energyas a reliable, low-carbon baseload for growing AI data-center loads while flagging fuel, regulatory and funding obstacles.
The Subcommittee on Energy of the House Science, Space, and Technology Committee met for a hearing titled "Powering Demand, Nuclear Solutions for AI Infrastructure" to examine whether nuclear energy can supply the large, continuous electricity needs of artificial-intelligence data centers.
The hearing opened with the subcommittee chair announcing the panel and the topic. "Welcome to the hearing titled Powering Demand, Nuclear Solutions for AI Infrastructure," the chair said, framing the session around AIand electric demand. The chair and multiple members emphasized that projected growth in data-center electricity use — cited repeatedly in testimony as rising from about 25 gigawatts today to about 80 gigawatts by 2030 — is driving interest in nuclear generation.
Why it matters: witnesses and members said AI-specific data centers will require reliable, often 24/7 power, and some corporate buyers prefer baseload generation that can minimize downtime and ensure continuous operation. "Nuclear's capacity factor of 92.5%... is the highest of any energy source," a committee opening statement said, and the hearing repeatedly framed that reliability as essential because data centers "can afford no more than 5.25 minutes of downtime annually."
Panelists described how nuclear fits into a broader energy mix. Pat Schweiger, chief technology officer of Oklo, told the subcommittee that advanced reactors and fast reactors can provide dependable power and that Oklo has a pipeline of customer commitments. Kathleen Barone, executive vice president and chief strategy and growth officer at Constellation Energy, said existing nuclear plants can be kept online, uprated, or restarted to meet near-term demand while advanced reactors add supply longer term. Dr. Jeremy Renshaw, executive director of AI and Quantum at the Electric Power Research Institute (EPRI), emphasized both the scale and uncertainty of future AI energy demand and described EPRI initiatives to coordinate the energy and technology sectors.
Members and witnesses stressed a combination of approaches: maintain and optimize the current nuclear fleet, accelerate advanced reactor demonstrations and commercialization, address fuel availability (including HALEU), and use data-center flexibility where possible to reduce peak stress on the grid. Members also raised state-level policies and public‑private arrangements that can accelerate deployment.
Remaining questions: witnesses and members flagged several constraints that could limit nuclear's near-term role, including HALEU fuel availability, licensing and permitting timelines, the cost of financing first-of-a-kind projects, and potential budget cuts to DOE programs that support reactor demonstration.
The hearing concluded after back-and-forth between members and witnesses on implementation steps, with members reserving written questions for the record.