Valar Atomics Wants to Build 456 Nuclear Reactors in a Utah Coal Town. The NRC Hasn't Licensed One Yet.
Project Beehive would lease more than 9,000 acres of Carbon County, Utah to a startup proposing 456 helium-cooled microreactors generating 9.6 gigawatts — nearly matching the state's entire peak grid capacity. Valar Atomics has raised more than $1.5 billion since 2023. Its reactor design has not cleared the one agency that can actually let it operate.
Carbon County, Utah mined almost 10 million tons of coal in 2006. It hasn't mined a pound since 2023. Into that vacuum, a nuclear startup called Valar Atomics wants to lease more than 9,000 acres of federal land outside Price for Project Beehive — a data center complex that would run on 456 small nuclear reactors producing a combined 9.6 gigawatts. For scale: Utah's entire electric grid produces roughly 4 gigawatts on an average day and tops out around 10 gigawatts on the hottest afternoons of the year. One company, on one stretch of Bureau of Land Management ground, is proposing to build something close to the state's whole peak capacity, using reactors the Nuclear Regulatory Commission has not approved.
A county that needed the yes
Carbon County's coal mines once employed thousands; the last decade has been a slow unwind of that economy, and county officials have made clear they see data centers — nuclear-powered or otherwise — as the replacement. The Utah Trust Lands Administration Board didn't deliberate long on the piece of Project Beehive it controls: on September 17th it unanimously approved Valar's lease of more than 640 acres of state trust land. Carbon County isn't getting just one project, either. A second complex nearby, nicknamed "Jurassic Spark," would run on a conventional natural gas plant instead of reactors. Residents who've shown up to public meetings aren't reassured by the variety. "It's all descending, these data centers," Price-area resident Cindy O'Neil told the Salt Lake Tribune. "They're setting these things within a couple of miles of homes."
What's approved, and what's still a filing
It's worth separating what's actually happened from what's merely been proposed, because most of the coverage this project has drawn conflates the two. The state trust land lease — 640-some acres — is done. The much larger piece, the 9,000-plus acres of BLM land the campus and its reactors would actually sit on, is not. The agency's Utah office has confirmed only that it received Valar's application and is checking it for completeness, the preliminary step before a formal review even opens. A full National Environmental Policy Act process, with the public comment period that implies, comes after that — assuming the application clears the first hurdle at all. Nobody at BLM has approved a single acre of the federal land this project actually depends on.
The reactor technology is an earlier-stage bet than the land. Valar's design is a Generation IV, helium-cooled high-temperature gas reactor — gas instead of water for cooling, part of the pitch for siting it in the high desert, since it sidesteps the water-rights fights that have dogged other data center projects this year. It is also a design that has never been licensed anywhere for commercial operation; the NRC has not certified it. The only small modular reactor design that has cleared full NRC certification in the United States, NuScale's, took the better part of a decade to get there, for one design — not 456 units stacked across a single site, alongside a planned on-site fuel-production facility and nuclear waste storage, both of which carry their own separate and arguably harder licensing paths. Valar's own materials call for non-nuclear construction to start by the end of this year, first reactors online in 2028, and the full 9.6-gigawatt build-out by 2032. That schedule only works if the hardest regulatory step in the whole project — getting a brand-new reactor design through the NRC, then stacking hundreds of site-specific licenses on top of it — moves faster than any reactor design in American history has moved before.
Money is not a license
None of this is a knock on Valar's ability to raise capital, which has been remarkable. Founded in 2023 by 27-year-old Isaiah Taylor, the company has pulled in more than $1.5 billion in well under a year: a $130 million Series A last November, a $450 million round at a $2 billion valuation, and a $1 billion Series B in August led by Sequoia, with backing from Anduril founder Palmer Luckey and Palantir CTO Shyam Sankar. That's real, institutional-grade money from people who do serious diligence on where they put it. It is also not the same thing as a licensed reactor, in the same way Oracle's force majeure notice on its own New Mexico pipeline fight wasn't the same thing as a working power plan — a well-funded company's own description of its project is a claim, and claims sit at the bottom of the evidence ladder regardless of how many zeros back them.
GridPermit doesn't work Utah, and nothing in our Colorado footprint looks remotely like a 456-reactor campus in coal country. But the shape of the problem is one we check for on every site, at any scale: does the claimed capacity number survive contact with what's actually been licensed, permitted, or approved, versus what's merely been announced, leased, or funded? Project Beehive has a land deal for its smallest piece and a pending application for its biggest one. It has venture capital, not a reactor design the NRC has signed off on. Carbon County may well end up with 9.6 gigawatts of nuclear power a couple of miles from people's homes. Right now, what it has is a filing, a lease, and a very large number nobody outside the company has verified yet.
Evaluating a specific parcel for a 50–300 MW load? GridPermit's Preliminary Site Kill Test grades the power, zoning, water, and commercial-readiness evidence before you buy the land.
Start a Qualification ReviewSee state-by-state coverage · Read the methodology · See a sample report