Last week, the Nuclear Regulatory Commission gave the Bill Gates-backed atomic energy startup TerraPower permission to begin construction on its first plant in Wyoming, marking the first time the United States has greenlit a fourth-generation reactor that used a coolant other than water for commercial use in nearly 60 years.
The move comes as more than a dozen new nuclear companies race to roll out smaller, and potentially more efficient, designs for fourth-generation reactors meant to do everything from powering data centers and producing industrial heat to generating electricity in space for a moon base.
The approval last Wednesday highlighted what analysts described as real changes to the way the NRC functions following President Donald Trump’s series of executive orders last May, targeted at accelerating the regulatory approval process for new reactors. But experts also warned that TerraPower — a company backed by one of the world’s richest men with 20 years of work behind it — may be an exceptional case, one that could face stickier questions on the backend when seeking an operating license sometime in the future.
Brett Rampal, an analyst at the consultancy Veriten, told Latitude Media that the approval is “a great sign” for the NRC’s ability to get things done, and demonstrates “that people who put their head down and engage in work can get things accomplished in what is a much more timely timeframe than we have seen for commercial regulatory engagement at the NRC in decades.”
To Nick Touran, a nuclear engineer who spent 15 years at TerraPower before going independent, “it’s progress.” That said, “it still doesn’t say anything about whether the project is going to totally boondoggle or if the reactor will work well when it turns on,” he added.
The fourth generation revival
All of the 94 commercial power reactors in operation in the United States are light water reactors, which use either pressurized or boiling water to cool the heat generated from splitting uranium atoms and make steam to generate electricity. The overwhelming majority of the more than 400 reactors in operation worldwide are also cooled with water at present — the result of the market’s years of trial and error.
Throughout the 20th century, scientists in the U.S., Russia, Japan, and Europe designed all kinds of research reactors that used alternative coolants, allowing for fission to reach much higher temperatures than water could safely handle. That included solid graphite, molten salt, and high-temperature gases such as helium. The vast majority of those reactors were never commercialized for reasons ranging from the high cost of specialized fuel to low power output due to routine operational issues.
One notable example was the Tennessee Valley Authority’s Clinch River project, constructed in the early 1970s. Costs of the so-called “breeder” cooled with liquid sodium soared so high that Congress ultimately slapped a debt cap on the government-owned utility and the effort was ultimately abandoned. An Xcel Energy subsidiary made another attempt a few years later in Colorado — the Fort St. Vrain project — which also ended up being too expensive to keep running.
In the nuclear industry’s modern parlance, these reactors cooled with something other than water would be considered fourth-generation, or “advanced” reactors (though the latter also applies to third-generation water-cooled reactors like the Westinghouse AP1000). Today only one fourth-generation reactor is in commercial operation in the world: a helium-cooled, high temperature gas reactor that China hooked up to the grid at its Shidao Bay Nuclear Power Plant in December 2023.
For years, TerraPower’s Natrium design has been the leading fourth-generation effort in the U.S. Like Clinch River, Natrium is a sodium-cooled fast reactor, and is capable of churning out 345 megawatts of electricity. It comes with a molten salt energy storage unit that stores excess nuclear energy, raising its potential output to more than 500 MW when both the reactor and the battery are pumping out electrons.
Technically, TerraPower’s reactor doesn’t qualify as a small modular reactor; that term defines units that generate more than a microreactor’s maximum of 20 MW but less than 300 MW. Still, the technology is meant to be easily replicable and small enough that utilities would need to order the reactors in bundles, an approach that SMR boosters say will bring down the cost per unit — especially as compared to the large, bespoke, legacy plants.
Several U.S. competitors are developing small fourth-generation systems of their own. The Amazon-backed X-energy’s high-temperature gas-cooled units tap out at 80 MW of electricity each, and theoretically generate enough heat to replace fossil fuels in petrochemical and other industrial processes. Like X-energy, both Valar Atomics and Radiant Nuclear use helium to cool their five- and one-MW units, respectively. Meanwhile, the California-based company Oklo is pushing a 75-MW liquid sodium-cooled machine that is closer to TerraPower’s design.
Who’s next at bat
Given that the Clinch River and Fort St. Vrain reactors were approved under the NRC’s predecessor agency, the Atomic Energy Commission, TerraPower’s construction permit made history as the first fourth-generation reactor endorsed by the regulator.
“This approval came in well ahead of schedule and both the NRC and TerraPower deserve credit for that,” Madi Hilly, managing director of the consultancy Radiant Energy Group (which has no relation to Radiant Nuclear), told Latitude Media. “TerraPower has been at this for a long time, so it would make sense that they were prepared with a high-quality application. But it also seems that there’s this new culture of rigorous but efficient reviews taking hold at the NRC.”
As a result, she said, “I think we’re going to see a lot more high-quality projects moving faster than ever before.”
Of the projects likely to move next, Touran said, the X-energy and Valar projects, as well as the Google-backed Kairos Power’s reactor, are the most promising. Kairos is so far the only fourth-generation reactor company with a power purchase agreement with a utility.
New momentum on reactor approvals could help quell mounting criticism of the NRC itself. Over the past few years, a contingent of next-generation nuclear executives have lambasted the NRC as the chief impediment to building new reactors, arguing that its focus on protecting the public from the risks of atomic energy smothered the industry and inadvertently secured fossil fuels’ dominance over the American power grid. Among them was Isaiah Taylor, the founder and chief executive of Valar Atomics.
Last Wednesday, after news of TerraPower’s permit broke and Valar’s own preliminary safety analysis advanced at the Department of Energy, Taylor offered rare praise for the agency he has long fought against. In a post on X, Taylor said “it’s massively under-appreciated how much work goes into these reviews” on the U.S. government side. “One of the best parts of my job is meeting the insanely smart folks” from the Energy Department, the NRC and the nuclear Navy.
It’s a sign, Hilly said, of the shifting attitudes toward the NRC as the agency looks to support construction of a new wave of reactors.
“A lot of advanced nuclear companies have long thought that the NRC was a barrier to deployment, perhaps an insurmountable one,” she said. “What’s appearing to be true is that high-quality nuclear projects are going to move forward faster than ever before, including new designs, so that the NRC may actually be helpful in getting these projects built while also protecting them against the downside risk by making sure that these evaluations are rigorous and up to the standard that the NRC has always held.”
Still, speeding up one part of the process could just push scrutiny to another, Touran warned.
“This is just a construction license. It means go ahead and build it. It doesn’t mean you can turn it on,” he said. “There’s a question about how much was pushed into the future and how much that’ll come back when they apply for an operating license.”
Fort St. Vrain, he added, “is a cautionary tale.”
“We built this other super advanced reactor in the 1970s in Colorado. It was super safe. It was super high temperature. But it didn’t work that well and we shut it down,” Touran said. “That’s just the risk of doing anything in nuclear.”


