The Department of Energy put $3.2 billion on the table six years ago to prove that a new generation of nuclear reactors could actually be built in America. Now, for the first time, concrete is being poured.

Kairos Power broke ground on its reactor in Oak Ridge, Tennessee, in April. TerraPower started construction on its plant in Kemmerer, Wyoming, shortly after. A third company, X-energy, expects to secure a construction permit by early 2027 for a reactor it is building for Dow Chemical in Texas. Three companies, three different reactor designs, three construction sites. After decades of nuclear promises that evaporated before a shovel hit dirt, that counts as a genuine turning point.

What the government actually built here

The program, called the Advanced Reactor Demonstration Program, works on a straightforward principle: the government covers up to half the cost of each project, but pays out only when developers hit predefined technical and schedule milestones. No delivery, no check. It also provides regulatory and design assistance, which matters enormously for technologies that have never gone through the licensing process before.

For TerraPower's sodium-cooled Natrium plant, total projected cost is $4 billion. The federal government is contributing $2 billion, with TerraPower and its investors matching that dollar for dollar. For X-energy, the DOE committed $1.2 billion under a 2021 agreement. For Kairos, the government agreed to provide up to $303 million. In each case, the public commitment appears to have unlocked the private one. As Adam Stein of the Breakthrough Institute told Reuters: "If there was no public side of the funding, then the private side of the funding probably wouldn't have been there either."

The private side, it turns out, includes some recognizable names. Google signed a deal to take 50 megawatts of clean electricity from the Kairos reactor through the Tennessee Valley Authority grid. Meta signed an agreement with TerraPower for eight Natrium units, which include heat storage systems that can temporarily push output to 500 megawatts. Dow Chemical is the anchor customer for X-energy's four Texas reactors. These are not speculative future offtake agreements. They are signed contracts, which is a different thing entirely.

Why this matters beyond the energy industry

The United States has not built a new commercial reactor design from scratch in decades. The only recent examples, the Vogtle 3 and 4 plants in Georgia, became operational in 2023 and 2024 after finishing seven years late and at more than double their original projected cost. That history is exactly why private capital has been so reluctant to fund nuclear alone.

Small modular reactors, the category all three of these projects fall into, are designed to be factory-built in standardized units rather than custom-engineered on-site. The theory is that standardization cuts construction time, reduces cost overruns, and makes nuclear financeable the way wind and solar eventually became financeable. The theory has never been tested at commercial scale in the United States. These three projects are the test.

TerraPower's Natrium plant is targeting completion in 2030. Kairos also expects to finish by 2030. Whether those timelines hold matters well beyond Wyoming and Tennessee. If these plants come in reasonably close to schedule and budget, they provide the proof of concept that could unlock a much larger wave of nuclear investment in the 2030s. If they run into the same overruns that plagued Vogtle, the commercial case for small modular reactors gets much harder to make.

The deeper story here is about what it takes to start an entirely new industrial sector inside an existing regulatory and construction environment. The federal program was, in Stein's words, "intentionally patient about getting to completion instead of rushing through something." That patience, combined with a milestone-based payment structure that kept pressure on developers, may turn out to be the design feature that matters most.

Whether three reactors in three states can rebuild American nuclear construction competence after a 40-year gap is a genuinely open question. But the groundbreaking has happened. The question is no longer theoretical.