Scott Nolan drops a truth nuke: the U.S. must rebuild uranium enrichment capacity to power nuclear's AI-era renaissance

Summary

Scott Nolan founded General Matter to solve what he calls the single most critical bottleneck in America's nuclear revival: uranium enrichment. The U.S. went from producing roughly 85% of global enriched uranium in the 1980s to less than 0.1% of global capacity today. Russia holds more than half of world enrichment capacity. Europe is second. China is a fast-growing third. The U.S. is effectively out of the game.

The collapse was deliberate. After the Cold War, the "Megatons to Megawatts" program had Russia downblending weapons-grade material into reactor fuel, and the U.S. gradually outsourced enrichment rather than maintain aging gaseous diffusion plants. That decision made sense under free-trade assumptions. It looks very different now, when every advanced reactor company Nolan met during a decade at Founders Fund gave him the same answer: we have no domestic fuel source.

The enrichment gap

The five-step nuclear fuel cycle runs from uranium mining through milling, conversion to gas, enrichment, deconversion back to solid, and finally fuel fabrication into pellets. The U.S. has functioning capacity at every step except enrichment. U.S. uranium deposits exist and are economical to mine, with new mines coming online in Texas and Wyoming. Honeywell operates a conversion facility directly across the river from General Matter's site in Kentucky — it was mothballed last decade and is now being restarted. Enrichment is the one gap with no commercially relevant U.S. capacity.

Nolan is explicit that enrichment is not a nuclear engineering problem. There are no nuclear reactions and no chemical reactions at an enrichment facility — it is a material separation operation. General Matter's team is built primarily from aerospace, applied physics, electrical, software, chemical, and mechanical engineers, with nuclear engineers as a minority. The company is not disclosing its enrichment process.

Paducah, Kentucky

General Matter signed a lease with the Department of Energy in August 2025 for 100 acres inside a 3,000-acre DOE site in Paducah, Kentucky — formerly the top enrichment facility in the U.S. before it was shut down in 2013. The plan is to build roughly one million square feet on that footprint, using depleted uranium already stockpiled on site as a potential feedstock. Engineering headquarters and G&A are based in California; construction and production are in Kentucky. General Matter is building its own EPC and general contractor function in-house, modeled on how Tesla ran Gigafactory construction.

The company is targeting commercial-scale domestic enrichment online by end of decade. Last week's DOE award, Nolan says, pulls that timeline forward by a few years. Facility capital costs are over $1 billion, rising to multiple billions at full scale. Nearly all technical staff require security clearances.

Market sizing and the Thiel logic

Nolan frames the opportunity in classic zero-to-one terms. HALEU — high-assay low-enriched uranium, the fuel required by most advanced reactor designs — is roughly a $100 million market today. Standard LEU enrichment for existing U.S. reactors is a couple of billion dollars annually in the U.S. alone. The near-term target is HALEU, because without a domestic source, advanced reactors cannot come online regardless of how well the reactor technology performs.

Reactor timeline

Nolan expects the nuclear growth curve to steepen sharply at the end of the decade. Several reactor companies are targeting first criticality tests in 2026. Under revised NRC executive orders, the subsequent licensing process is meant to take 18 months, putting commercial deployment approvals around 2028. Meaningful scale in deployed capacity, he says, realistically arrives in the early 2030s.

Solar versus nuclear

For terrestrial data centers, nuclear wins on base load reliability. Solar is intermittent and requires grid-scale storage that is not yet cost-competitive with nuclear. In orbit, the calculus flips: a sun-synchronous satellite gets sunlight roughly 99% of the time, eliminating the storage problem. The remaining engineering question for orbital data centers is heat dissipation from GPUs — Nolan sees that as the open materials science challenge for the next few years.

Competitive moat

Building a competing enrichment facility requires more than $1 billion in capital, a team with the right security clearances, a site with community support and regulatory history, and technical knowledge that atrophied across the entire industry. Nolan wants multiple U.S. players to emerge — he draws the parallel to SpaceX, which built the talent vortex that defined the commercial launch industry and eventually pulled Blue Origin forward. But the SpaceX comparison also captures his cost discipline: General Matter intends to stand on commercial revenue, not indefinite government subsidy.

The company is hiring more than 100 people across its California and Kentucky sites in 2026. The workforce follows a demographic bathtub curve — a cohort of young nuclear enthusiasts at one end, engineers who worked in the industry during its 1980s peak (some now in their 80s, serving as consultants) at the other, and a thin middle generation from the decades when nuclear stalled.