This company wants to use U.S. Navy nuclear reactors to power AI

HGP Intelligent Energy wants to put U.S. Navy nuclear reactors on land, and it wants to do it now.

Partnering with the Shaw Group, the company plans to deploy proven submarine and carrier reactor designs at the Department of Energy site in Paducah, Kentucky.

No waiting for next-gen commercial reactors.

No decade-long construction procedures.

Take what already works and move it ashore.

HGP pitched the idea to the federal government in December. The target customer is the government itself, starting with power-hungry programs like advanced computing and Project Genesis, US government's AI platform.

Shaw would handle engineering, procurement, and fabrication, leaning on its experience from Vogtle Units 3 and 4.

The pitch rests on one big credential. The U.S. Navy nuclear fleet has logged more than 7,500 reactor years of operation.

Paducah is not a random choice. The old gaseous diffusion plant shut down in 2013, but the site has come back as a nuclear hub. General Matter is building a new enrichment facility there after securing $900 million from the DOE. Global Laser Enrichment is also operating on site with laser-based enrichment. Across the river sits Solstice Materials’ uranium conversion facility. HGP wants to drop reactors into the middle of an already forming nuclear supply chain.

The disruption behind the news: This plan cuts straight through the biggest problem in nuclear power. Time.

Traditional nuclear projects take 10 to 15 years from concept to grid.

Costs balloon.

Political patience runs out.

By reusing naval reactors, HGP is trying to compress that cycle to something closer to three to five years for initial deployment.

That changes who can realistically buy nuclear power. Hyperscale computing centers, federal AI programs, and defense infrastructure don’t need merchant-market economics. They need guaranteed megawatts, fast. Naval reactors are designed for compactness, durability, and predictable output. Those traits line up perfectly with data centers and classified facilities that can’t afford downtime.

There’s also a regulatory jolt coming. Naval reactors live in a parallel safety universe run by the Navy, not the Nuclear Regulatory Commission. Putting them on land forces a collision between two regulatory cultures. If Washington greenlights even a limited deployment, it creates a fast lane for nuclear that commercial developers will demand access to next. That’s the part incumbents should fear.

For utilities, this is bad news. A federal customer that can self-supply nuclear power bypasses regional grids entirely. That weakens utility load forecasts and long-term investment planning. For reactor startups chasing novel designs, this raises the bar overnight. Why fund an unproven system when the government can point to decades of operational history and say it wants that instead.

What to watch next

First, watch how the DOE frames oversight.

If it keeps control in-house rather than defaulting to the NRC, expect a wave of copycat proposals at other federal sites within 18 months.

Second, watch Shaw’s contracting timeline.

If balance-of-plant work starts before 2028, that’s the signal this is moving faster than any commercial reactor in the U.S. has in generations.

Third, watch Paducah’s political alignment.

Kentucky could lock in a full nuclear lifecycle cluster, from enrichment to conversion to power generation. That would pull federal dollars and talent away from slower-moving states.

If US Navy reactors work once, it won’t stay small. Land-based naval reactors could start popping up everywhere.