Third in an occasional series on companies leading the way on small nuclear reactors. Read the first one here and the second one here.
The people of Kemmerer, Wyoming, braced for the worst when the owner of the town’s coal-fired power plant scheduled its closure in 2020. Jobs would disappear, along with tax revenue.
Then came TerraPower, an advanced nuclear company based out of Bellevue, Washington, that needed a place to build. TerraPower pitched itself to local officials as future employment for the power plant workers about to lose their jobs.
Instead of closing the power plant, its owner, PacifiCorp, is retrofitting it to burn natural gas. And TerraPower is charging ahead with plans to build the first small U.S. nuclear reactor of its kind close by, with the support of local officials in western Wyoming.
“You’re shutting the lights off in too many places in America,” said Kent Connelly, chair of the Lincoln County Commission.
“We have no complaints about TerraPower,” Connelly added.
TerraPower is widely considered one of the strongest contenders to build advanced nuclear reactors for the U.S. grid. It also continues to have backing from Microsoft co-founder Bill Gates, its board chair who has poured hundreds of millions of dollars into the company.
In Wyoming, TerraPower has started preparing the site and is expected to get a construction license from the Nuclear Regulatory Commission in the first quarter. The only other small nuclear power plant with a permit to build is California’s Kairos Power.
“The significance of this milestone — particularly in a period when there is significant ‘buzz’ around a range of new advanced reactor designs — is that it positions TerraPower firmly as one of the leaders of the pack,” said Alan Ahn, deputy director for nuclear at Third Way.
Others are worried that TerraPower is moving too fast, with its design getting less scrutiny as the Trump administration pushes regulators to deliver federal permits more quickly.
TerraPower is looking to the next project. Last September, the company began studying for sites in the United Kingdom, and it is exploring a project in Kansas. Last month, it scored a deal to sell power to Facebook parent company Meta, which would help finance up to eight reactors across the country.
But even those with high expectations see reasons to be wary of the sodium-cooling technology TerraPower plans to deploy — a concept with a checkered history in the lab. And the company faces the same questions about the costs of building new reactors.
“The biggest challenge nuclear faces is proving it can build them in a way that makes them economically viable,” said former NRC Chair Richard Meserve.
TerraPower did not provide POLITICO’s E&E News with an estimate of how much it would sell power for, but CEO Chris Levesque said a 2023 report from DOE provided a good estimate for costs.
Applying that report’s formula to TerraPower’s first reactor — alongside current estimates for the price of fuel — suggests it will take 11 to 17 years of operation to break even if the company sells power at Wyoming’s average price. Operation is currently slated to start in 2032 and continue for up to 80 years.
Construction costs should come down by 40 percent with subsequent iterations, the DOE report says. Levesque thinks TerraPower can do one better, cutting more than 50 percent off its reactor price tag with time.
“Design effort, a thousand people for three-plus years: That is now behind us. That’s a nonrecurring cost that future customers won’t have to pay for,” he said, adding that nuclear power from developers like TerraPower will help with the affordability of electricity in the long run.
A sweet spot for utilities?
TerraPower breaks the mold of advanced reactors in many ways.
Most immediately apparent is size and output. The dozens of large conventional reactors that dot the U.S. today consistently produce more than 700 megawatts. That’s enough to power roughly 570,000 homes. Most advanced reactor startups are designing “small modular reactors” with less than 300 MW capacity.
Each TerraPower “Natrium” reactor will produce 345 MW around the clock, which is the approximate demand of some Amazon data centers. Levesque says that size is just right.
“It begins with: Where’s the demand?” Levesque said. “We can support a gigawatt-sized data center with the Natrium four-pack.
“We looked at going smaller. It didn’t make sense to us when you go to a smaller plant. You have negative economies of scale,” Levesque added. “We went smaller than a gigawatt because we think it works for a somewhat weak grid, which unfortunately we have in this country.”
In theory, a Natrium reactor could be slotted in at a retiring nuclear or even coal plant, providing a new heat source while leveraging existing mechanisms for converting the heat to electricity and grid connections.
A January report from the Nuclear Innovation Alliance found that medium-sized reactors are bigger than what most data centers are looking for, but perfect to meet demand from municipal utilities and to repower old plants.
On top of the baseload 345 MW, Natrium boasts an energy storage system that can temporarily add 155 MW to the grid for a little over four hours at a time. No U.S. nuclear plant today has anything close to that level of flexibility.
Natrium works by generating heat that is then transferred to a tank away from the reactor. When the grid has enough power, Natrium can store some of the heat for later. When operators need more, the plant can release stored heat to boil water, sending steam to spin a turbine and generate extra electricity.
“That pairs it very nicely with renewables,” said Chris Gadomski, a nuclear analyst at BloombergNEF.
Wind and solar are usually the cheapest forms of electricity. The downside? They can only produce when the wind blows or the sun shines. As such, utilities are turning to energy storage systems to allow them to maximize the cost reductions of renewables while ensuring enough supply on windless nights.
A rushed review?
NRC staff cleared the Wyoming project in a safety review announced in December.
But Ed Lyman, director of nuclear power safety with the Union of Concerned Scientists, isn’t convinced. He notes that the review was originally scheduled for completion in August 2026.
He believes that the risk of sodium fires and other “inherent instabilities” was overlooked by staff under pressure from the Trump administration. In a May executive order, President Donald Trump imposed 18-month deadlines for the NRC to decide on reactor applications, and DOE officials have been pressuring the agency to expedite its regulatory process.
“The only way the staff could finish its review on such a short timeline is by sweeping serious unresolved safety issues under the rug or deferring consideration of them until TerraPower applies for an operating license, at which point it may be too late to correct any problems,” Lyman said in a statement.
To date, practically every sodium-cooled fast reactor around the world has encountered significant operational problems and sometimes serious accidents.
“Sodium fires happen when sodium interacts with water of any kind, including dampness in ordinary air,” said Katy Huff, former assistant secretary of nuclear energy at the Department of Energy.
Japan tried the technology at the Monju plant in the 1990s, but before it even began serving the grid, it was forced to shut down due to a sodium fire.
“If you were to sort of characterize the Japanese nuclear safety culture and the American nuclear safety culture, they’re generally pretty similar,” Huff said. “Sodium fires aren’t unheard of, even in very well-managed machines.”
At the same time, Huff says, the fires come from hot sodium slowly leaking rather than a high-pressure explosion, meaning such accidents are rarely catastrophic. Arguably, it makes sodium cooling even safer than water cooling in today’s reactors.
Russia’s BN-600 — one of only two commercial sodium-cooled fast reactors — suffered 14 sodium fires in its first 17 years of operations but is still producing electricity 45 years after its inception. Moreover, the reactor experienced no sodium leaks between 1995 and 2017.
Levesque says that the historical trend should give people confidence.
“We’ve been working around sodium for 18 years. We’ve studied the sodium reactor designs around the world,” Levesque said. “We made it such that sodium and water are never in the same building.”
