Nuclear Notes is produced by the Breakthrough Institute, and covers the policy, regulation, technology, and business of advanced nuclear energy in the United States and beyond.
Every Thursday, we track what’s really going on in the field: the new rules moving through the Nuclear Regulatory Commission, the advanced reactors moving from design to deployment, the companies going public and the utilities merging, and the executive orders and acts of Congress reshaping how the country builds. We explain not just what happened, but why it matters and how the pieces fit together.
The Bad Guesses That Led to Yucca Mountain
Advocates for burying spent nuclear fuel at Yucca Mountain say it was chosen because the site, about 100 miles from Las Vegas, was remote, dry, and capable of holding in the radioactive material for millennia.
Those statements about Yucca are true, but the four other sites under consideration before Congress chose Yucca also comfortably met the safety standard established by the Department of Energy. According to a re-analysis by Dr. Adam Stein, the Breakthrough Institute’s Director of Nuclear Energy Innovation, Yucca outranked the other sites because the estimate was that it would be less expensive to build there than at the other finalist sites: Deaf Smith County in Texas, Richton Dome in Mississippi, Davis Canyon in Utah, or Hanford, in Washington State.
And the cost estimates were far from precise. The Energy Department’s 1986 estimate put the cost of building and operating the Yucca repository at $7.5 billion, but by 2008 it was over $96 billion. Even with inflation, the original estimate was not meaningful.
The cost estimates for the other sites were also just that, estimates. They were stated in ranges, and if the analysts had used the upper bound of the cost estimates, the ranking would have changed.
Estimates of repository costs accounted for nearly 88 percent of the differences among the sites, which is important because the estimates were probably of very poor quality. The DOE had not characterized any of the sites when it made the cost estimates, so the estimates themselves were unreliable.
“The cost estimate was too uncertain to serve as a decision point,” Dr. Stein concluded. The weighting was “concentrated in a single uncertain attribute,” he wrote, and the result was to “indicate robustness that the model does not have.”
Congress passed the Nuclear Waste Policy Act in 1982, and the Energy Department studied nine sites. By 1985, the Energy Department had winnowed this down to three sites, but in 1987, Congress amended the law to tell the Energy Department to evaluate just one, Yucca. Yucca was ranked first, which let lawmakers from other states go home to tell their constituents that they had picked the top-ranked site and that the repository would not be built in their states. But Nevada eventually fought off Yucca, largely through the efforts of Senator Harry Reid, who became the Senate Majority Leader. Reid maneuvered to cut off funding from the program, and Senator Barack Obama, running for president, pledged to kill it. The result is that the government’s high level waste program has been dormant for years, and spent fuel is stored at the reactor sites, at a cost to the federal government approaching $1 billion a year.
Now, through the Nuclear Lifecycle Innovation Campus initiative, the Trump administration is proposing to establish several sites where used fuel could be reprocessed, recovering plutonium produced in reactor operations plus unused uranium. The residual material could potentially be disposed of at those sites.
On Tuesday, the House Energy and Commerce Committee held a hearing on nuclear spent fuel policy and to examine the NLIC program. Ted Garrish, DOE Assistant Secretary for Nuclear Energy, repeatedly stated that the law needs to be updated to allow the agency to characterize and identify potential sites; currently, only Yucca can be the U.S.’s designated waste site. In this anticipated next round of siting studies and evaluations, DOE needs to take a different tack in its comparison of sites. Dr. Stein identifies that the failed step last time was between the analysis and the ultimate decision, and recommends how exactly DOE should solve that.
A Nuclear Company Sees Uncertainty on Wall St.
Last week Holtec, following a filing with the Securities & Exchange Commission, announced plans for an initial public offering that could raise $900 million and would value the company at about $10 billion. But this week the company delayed it, amid fears of a bubble in a closely-related industry, artificial intelligence.
But even if the IPO is delayed, the documents are worth a look. Both stress how well-established Holtec is, with business lines in numerous aspects of the nuclear energy industry–a contrast to other nuclear entrants.
Holtec does not fit the legacy model embodied by Westinghouse and GE, which have decades of experience in servicing reactors, and a mostly long-past history of reactor construction, and are only now beginning to sell a new generation of designs with a cobbled-together supply chain. Nor does it fit the start-up model, with teams of smart engineers who have innovative designs for reactors with inherently better safety and performance characteristics, but no construction or operation experience.
Holtec is broader. Its capabilities include “reactor design, reactor operations, engineering, licensing, manufacturing, construction, commissioning, and decommissioning services,” says the company announcement. Its pitch to potential investors is that, with that experience, “Holtec believes that it is well-positioned to help catalyze the growth of the nuclear power industry in the U.S. and around the world.” Among its advantages is a steady stream of contracts for the massive casks used to store spent nuclear fuel.
Holtec lists its design capabilities, but perhaps its most distinctive feature is its vertically integrated approach to reactor delivery—a self-proclaimed leg up. Holtec designs, manufactures and fabricates, gets licenses, and executes. The company says it has “hundreds of completed projects delivered on-time,” although these are not complete reactors. Holtec is also on the verge of producing power from the Palisades reactor near Kalamazoo, Michigan, that it bought for decommissioning. The company wants to deploy two of its SMRs there, and the NRC just allowed preliminary site preparation.
The company has 40 years of experience in spent fuel management, holding about 90 percent of the U.S. market for components used for wet storage of spent nuclear fuel, and 76 percent of the market for dry cask storage. It also says it has 203 worldwide patents, customers in more than 40 countries on five continents, $577 million in annual revenue, and more than 4,150 employees worldwide.
As with all SEC filings, Holtec lists uncertainties and risks. For example, the company stressed its leadership in reactor decommissioning, a profitable business, but reactor owners can decide to extend their plants’ lives instead of decommissioning them. But Holtec has positioned itself in that business too. And Holtec’s SMR design is new but not audaciously different. It is seeking to launch a line of 300-megawatt units using the familiar pressurized water technology.
The IPO would have valued the company at $10 billion. Is that justified? Companies that have yet to build their first power reactors and have no industrial base or substantial lines of business generating income are valued in the billions. Oklo, which has a design for a more advanced reactor but has yet to sell one, is at $6.74 billion; NuScale Power, which has an NRC-approved design but no sales, has a market cap of about $3.5 billion.
Those companies will be commercial successes only if they can build their new reactors. Holtec seems certain to go forward whether its SMR-300 is built or not.
A New Link in the Fuel Supply Chain
An obscure but critical step in making reactor fuel is converting yellowcake, which is produced by uranium mines, into uranium hexafluoride gas, the substance that goes through centrifuges to sort the isotopes of uranium. This phase of the fuel cycle is known as conversion, before UF6 can then be enriched. The United States has one factory that does that work, in Metropolis, Illinois, and it closed abruptly in 2017 because of poor market conditions. It was overhauled and reopened in July 2023.
Now a start-up in Idaho Falls, Idaho, Raven-Flint Nuclear Corp., is pioneering a new chemical process for conversion that is supposed to have lower costs for capital and energy. Its process has not been publicly described but is potentially safer; it eliminates use of elemental fluorine, which is highly reactive and poisonous. Its application is now being evaluated by the NRC.
A Reactor to Maneuver a Satellite in Orbit
The Pentagon has awarded $161 million to Antares, a California startup, to develop a reactor that could provide propulsion to carry out “dynamic operations that were once impossible” in orbit.
Antares achieved initial criticality for its microreactor prototype on June 4th this year. That model and the space reactor will use TRISO fuel and sodium heat pipes. The space version will produce up to 1 megawatt. It is scheduled for testing on the ground next year. The award follows an executive order issued in December 2025 that calls for a reactor in orbit as soon as 2028.
The satellites in service today that can adjust their orbits use chemical rockets to do so. But they carry limited fuel. The United States has already deployed weapons in orbit, but their usefulness is constrained by the limited ability of satellites to maneuver. A reactor would vastly improve a satellite’s ability to chase other satellites. Such a reactor could also be a prototype for reactors on the Moon or Mars, or could provide propulsion for a rocket going to Mars, reducing the duration of the trip.
NRC Overhaul Will Stretch “Well into Next Year”
The NRC’s broad reform of its rules, mandated by President Trump’s executive order of May 2025, will continue “deeper into next calendar year,” the chairman, Ho Nieh, told Reuters. And, the Commission is getting closer to its goal of approving new facilities within 18 months, he said.
Thanks for reading!