Valar Atomics Proposes 456 Reactors in Utah. A Reality Check on Project Beehive

Valar Atomics has proposed Project Beehive, a campus of about 456 gas-cooled, high-temperature reactors near Price in Carbon County, Utah, to power a large data center complex. Data Center Dynamics (DCD) reports an estimated 9.4 GW in its headline and 9.6 GW in its text; Interesting Engineering reports 9.6 GW. Either way, it would be more than double Utah's average electricity demand of about 4 GW, according to DCD, in a state whose summer peak is around 10 GW.

This article separates what has been reported from what has been proven. We drew on DCD and Interesting Engineering and did not see the application itself.

What Has Been Reported

• The site. More than 9,000 acres of land managed by the federal Bureau of Land Management (BLM), plus roughly 640 to 650 acres of Utah state trust land. The state trust lands agency unanimously approved a lease on Sept. 17.
• The reactors. About 456 units using TRISO fuel, a particle fuel designed to tolerate high temperatures, cooled by gas.
• The demonstrator. Valar's Ward 250 reached zero-power criticality in June 2026, according to the coverage. Zero-power means a self-sustaining chain reaction at negligible heat output, a first step before running at power. Reports differ on where it was: DCD says a Utah laboratory, while Interesting Engineering describes a tent about 30 miles from the Beehive site. Interesting Engineering also reports the Department of Energy selected Valar under a pilot to reach criticality by July 4, 2026, and that on July 1 the unit powered a single Nvidia chip.
• The timeline. Reporting citing NPR says non-nuclear construction could begin in late 2026, the first reactors in 2028 and full build-out in 2032. Valar says no construction begins before permits and approvals.
• The approvals. BLM is reviewing the application for completeness. Next come a Notice of Realty Action, then environmental analysis under the National Environmental Policy Act with a public comment period.
• The opposition. Neal Clark of the Southern Utah Wilderness Alliance is among the critics quoted.

What Is Missing

None of the coverage we read names a customer for the power, names investors, or describes a Nuclear Regulatory Commission construction application. That is the reason to treat this as a proposal, not a project under way.

The Numbers Do Not Reconcile

This section is analysis, not reporting. Dividing 9.6 GW by 456 reactors gives about 21 MW per reactor. DCD describes a goal of 25 MW reactors and calls the Ward 250 a 5 MW design, and Interesting Engineering does not give a megawatt figure for the Ward 250. The coverage does not say whether those figures are thermal or electric power, or whether they describe different design phases. Thermal and electric figures can differ by a factor of two or three for a reactor, so the gap may simply be unit confusion. Until the filing is public, the headline gigawatts are an estimate, as DCD itself calls them.

What Would Have to Happen

Between a June zero-power test and 456 operating reactors sit several steps that are not yet visible in the reporting:

• Land and environmental review. BLM completeness, a Notice of Realty Action and a NEPA process with public comment.
• Nuclear licensing. A regulatory path with the NRC or an accepted alternative that the coverage does not describe.
• A buyer and financing. No tenant or investor is named, and a campus this size needs both.
• Manufacturing and fuel. Building hundreds of identical units depends on supply chains that the reporting does not address.

For context on why companies are chasing this model, see our guide to nuclear power and small modular reactors, our look at how small modular reactors could change the grid, and how Silicon Valley is financing nuclear. The alternative approach, extending existing plants, is in our piece on Google's deal with Constellation.

Why It Matters

• Data centers are driving nuclear proposals at utility scale. A single campus is being pitched at more than double a state's average load.
• A demo is not a license. Zero-power criticality shows a design works at the smallest scale, not that it can be permitted or built at scale.
• The review is public. The NEPA step includes comment, so local and environmental objections will surface.
• Watch for the filing. The reactor power figures, the buyer and the licensing path are the details that will show whether this is a plan or a pitch.