HGP joins DOE nuclear AI program to help reactors track data center demand

HGP Intelligent Energy has been named a commercial partner in Prometheus, a Department of Energy-led nuclear AI program backed by national labs, universities and 25 industry partners. The Dallas company says its digital twin and variable-speed reactor coolant pumps could help reactors follow data center load without pushing the cost onto the public grid.

Why it matters: - Prometheus is designed to cut nuclear deployment timelines in half and reduce operating costs by 50%, aiming to make new reactor projects faster to build and cheaper to run. - HGP's technology targets a fast-growing problem in the AI buildout: how to power data centers with dedicated nuclear generation instead of adding load to the public grid. - The program also seeks to demonstrate at least seven continuous days of autonomous operation for a commercial microreactor at half the required staff.

What happened: - HGP Intelligent Energy said July 22, 2026, that it was named a contributing commercial partner in Prometheus, the Idaho National Laboratory-led program selected by the U.S. Department of Energy under the Genesis Mission. - Secretary of Energy Chris Wright announced Prometheus at the Genesis Mission Summit in Washington. - The program is the largest of 278 projects selected by the Department of Energy and the only Phase II award. - Idaho National Laboratory lists HGP among the Prometheus partners.

The details: - Prometheus brings together four DOE national laboratories: Idaho, Oak Ridge, Argonne and Sandia. - The program also includes four research universities and 25 commercial partners, including Microsoft, Amazon Web Services, NVIDIA, Westinghouse, GE Vernova, TerraPower, Oklo, X-energy and Aalo Atomics. - The award directs $60 million in federal funding over three years to the participating national laboratories and universities, matched by more than $250 million committed by commercial partners. - HGP is participating in the AI for Nuclear Energy Consortium and is working with Argonne National Laboratory to validate and verify its technology. - HGP's NTH-Sim digital twin provides real-time core monitoring and predictive safety analytics. - HGP says NTH-Sim's solvers have been validated across pressurized water, sodium fast and microreactor classes. - HGP's variable-frequency-drive reactor coolant pumps enable continuous variable-speed operation. - The pumps are intended to let proven reactor designs load-follow and track minute-to-minute demand without moving control rods. - HGP says the two technologies address the program's stated gap that reactor operations remain labor-intensive and manually driven. - The company said its approach would support on-site nuclear generation for data centers under private contract rather than through shared transmission and distribution infrastructure. - HGP said it is not currently operating a facility under that model. - The U.S. is estimated to need 300 gigawatts of new nuclear capacity by 2050. - Over the last eight years, the cost of building nuclear power has risen from under $10,000 per kilowatt to more than $21,000 per kilowatt, driven by late-stage licensing and constructability surprises. - Prometheus targets a 10x reduction in design and licensing workflow time, a 3x reduction in manufacturing cycle time and a 50% reduction in operating staffing, benchmarked against documented non-AI baselines. - HGP said its technologies are aimed at proven naval-derived pressurized water reactor designs, with fabrication anchored by Texas manufacturing partners. - HGP said it has pending patent applications covering variable-speed reactor coolant pumps, AI-powered digital twin systems and extended fuel life technology. - HGP says it has operated power generation assets in ERCOT since 2013 and managed multi-gigawatt fleets since 1998. - The company also said its team includes experience with a First of a Kind battery energy storage resource approved by ERCOT in 2019 and in operation since. - More information is available at the company's website.

Between the lines: - The announcement positions HGP's tools as infrastructure for a new nuclear business model: reactors sized and controlled to serve data centers with private generation, not just the bulk grid. - The program's mix of federal labs, universities and major industrial players suggests DOE wants AI-centered nuclear development to move from research into commercialization. - HGP's emphasis on Texas manufacturing and ERCOT experience signals a bid to turn the program into deployable hardware, not just software.

What's next: - The consortium's charter and membership agreements are scheduled for completion in late 2026. - HGP said its participation remains subject to definitive agreement. - The program will continue work toward the design, licensing, manufacturing and staffing targets set by DOE. - HGP said any real-world deployment of its dedicated data center model is still ahead, not already in operation.

The bottom line: - HGP is putting its AI control systems into one of DOE's biggest nuclear modernization bets, with the goal of making reactors flexible enough to match AI data center demand without burdening the public grid.