Though the two countries are now in a race to develop atomic technology, China’s most advanced reactor was the result of collaboration with American scientists.
A high-temperature solvent-extraction column at the Molten-Salt Reactor Experiment, in Tennessee, in 1970.Photograph from US Department of Energy / Smith Collection / Gado / Getty
This April, in a speech given at the Shanghai branch of the Chinese Academy of Sciences, the physicist Xu Hongjie announced a breakthrough. For over a decade, his team had been working on an experimental nuclear reactor that runs on a lava-hot solution of fissile material and molten salt, rather than on solid fuel. The reactor, which went online two years ago, was a feat in itself. It is still the only one of its kind in operation in the world, and has the potential to be both safer and more efficient than the water-cooled nuclear plants that dominate the industry. Now, Xu explained, his team had been able to refuel the reactor without shutting it down, demonstrating a level of mastery over their new system.
As dazzling as that was, the timing of Xu’s speech also freighted the topic with geopolitical import. Only a few months earlier, DeepSeek, the Chinese artificial-intelligence company, had set alarms ringing through the U.S. tech world when it became clear that the relatively small Chinese startup, operating under U.S. export controls, had created a large language model that rivalled anything devised by the behemoths of Silicon Valley. Xu cast his team’s molten-salt reactor in the same light: yet another sign that the technology gap between China and the U.S. had closed.
Xu explained that his team had based their design on an experimental reactor that had been built in Tennessee in the nineteen-sixties. Known as the Molten-Salt Reactor Experiment, or the M.S.R.E., that project hit a dead end in the early seventies, when it lost federal funding. Xu’s team had learned everything they could about the M.S.R.E. so that, decades later, they could bring the project back to life. Xu compared their labors to the story of the tortoise and the hare: whereas the United States had “gotten lazy and made a mistake,” China had seized the “chance to overtake” it.
In reality, China’s molten-salt reactor was less the product of a race than a collaboration. Less than ten years earlier, Xu’s team had been working with an array of American nuclear scientists. M.I.T. had irradiated graphite samples for the Chinese scientists. Nuclear engineers from Berkeley flew to Shanghai to review the original design. And by 2015, at what was perhaps the peak of U.S.-China amity in the sciences, Xu’s home institution, the Shanghai Institute of Applied Physics, or SINAP, had signed a coöperative research-and-development agreement with Tennessee’s Oak Ridge National Laboratory, the site of the world’s first molten-salt reactor.
These agreements could be seen as products of Reagan-era neoliberalism. They allow national labs to hire out their facilities and staff to outside entities that, in exchange for funding, can secure a proprietary claim to any technologies U.S. national labs discover while working on the designated project. For the most part, this has facilitated technology transfer from public institutions to the private sector. But the agreement between O.R.N.L. and SINAP created an unprecedented situation: a Chinese state-owned lab was paying an American lab millions of dollars to develop materials and plumbing for molten-salt reactors.
From the start, the American side operated under the belief that the Chinese would be the first to build a molten-salt reactor. China was spending the money to do it, after all. There was some funding for molten-salt research in America, but much less than was needed, and this was why the Oak Ridge researchers were willing to accept support from the Chinese. Through the partnership, the American researchers were hoping to advance work on a less complex reactor, in which molten salt would be used as a coolant rather than a fuel line. “The budget is what the budget is,” David Holcomb, Oak Ridge’s principal investigator for the agreement, explained during a conference appearance at the time.
Ten years later, the armature of assumptions and policies that enabled such a partnership has been blown apart. After Donald Trump won the 2016 Presidential election, the Department of Energy severed ties with SINAP and threatened to revoke licenses from American companies that exported nuclear technology to China. During Trump’s second term, the Administration’s hostility toward China has only increased. “If you write about the coöperative research agreement, the new Administration will fire everyone for being evil collaborators,” a senior figure at a U.S. nuclear company told me. He was only half joking. When I requested information about molten-salt research at Oak Ridge, the media-relations manager there told me “we won’t be able to help you this time,” and later minimized the extent of the facility’s coöperation with SINAP. After a brief back-and-forth with Idaho National Laboratory and Los Alamos about interview requests for David Holcomb, who has since moved to the private sector, and for Thomas Mason, who had been the head of Oak Ridge during its partnership with SINAP, both laboratories stopped responding to my messages. This reticence stands in contrast to the atmosphere a few years ago, when Holcomb gave multiple interviews to the press, and Oak Ridge allowed tours of the facility that had originally housed its molten-salt reactor.
My two e-mails to Xu Hongjie similarly went unanswered. Then, in November, Xu passed away, reportedly dying while at work at his desk.
The days of “common interest” and “open science,” invoked by Holcomb in 2015, have given way to new mantras of “geostrategic influence” and “national security.” These talking points, which have been embraced by Republicans and Democrats alike, call on nationalism to reboot an industry withered by decades of retrenchment. It remains to be seen if this is enough to make the U.S. nuclear industry competitive in international markets, or even profitable domestically.
To make sense of what is happening in nuclear energy today, it helps to know about what was once called “the first nuclear era”—a thirty-seven-year stretch between 1942, when Enrico Fermi oversaw the first controlled fission chain reaction, and 1979, when the second reactor at the Three Mile Island Nuclear Generating Station, in southeastern Pennsylvania, partially melted down. At the height of this period, around 1960, the United States accounted for almost seventy per cent of global spending on R. & D. Nuclear energy, which sat at the nexus of defense and civil engineering, was a double beneficiary. From these investments came a series of ever more terrifying weapons alongside a fleet of experimental and commercial reactors that made the U.S. the world’s largest producer of nuclear energy. America still holds this title, but China is poised to assume the mantle, probably sometime around 2030.
The Molten-Salt Reactor Experiment epitomized the possibilities of this period. The concept originated in the late nineteen-forties, with a request from the Air Force to develop a nuclear-powered airplane. Alvin Weinberg, who later became the director of research at Oak Ridge National Laboratory, didn’t think that such an aircraft would fly, but he was willing to try to build one. He had helped develop the reactors that produced plutonium for the Manhattan Project and moved to East Tennessee, after the war. There, he presided over the development of O.R.N.L., which grew from a plutonium-production facility near the Clinch River. For Weinberg, the purpose of a national lab was to try “things too difficult or too risky for private industry to undertake.” An airplane that burned uranium was precisely that.
Weinberg wrote that the reactor would need to reach temperatures around fifteen hundred degrees Fahrenheit in order to power a jet engine. His team surmised that such heat would mangle any fuel rods small enough to install into an aircraft, so they decided to use fluoride salts. These melted into a liquid at around four hundred degrees Celsius and stayed stable above sixteen hundred degrees. With uranium fluoride mixed in, the molten salt itself could function as fuel.
The system went critical in November, 1954. In its brief life, it showed some remarkable properties, but the test also revealed some of the challenges of working with molten salt. Leaks were a constant problem, and the radiotoxicity of most of the apparatus made repairs next to impossible. As a stopgap, Weinberg’s team had to repeatedly off-gas the reactor compartment, bathing a nearby forest in radioactive xenon and iodine. At the hundred-hour mark, the project was shut down.
The Molten-Salt Reactor Experiment gave him another shot. By then, the Atomic Energy Commission was ready to make major investments in order to develop breeder reactors, or reactors that produce more fissile material than they burn. Breeder reactors promised energy on a scale far beyond what could be provided by the global supply of coal and oil, fuels that were projected to become scarce within a century and which were already suspected of warming the Earth. Planning began in 1960, and five years later Weinberg’s team loaded sixty-nine kilograms of enriched uranium into the salt. This time, the experiment was a success. The M.S.R.E. logged over thirteen thousand operational hours, during which the researchers ran countless tests. “They did, like, every calculation you could have done at the time to understand how you would build and run and fuel this reactor,” Katy Huff, the Assistant Secretary of Nuclear Energy under President Joe Biden, said. The most important finding was a simple one: the M.S.R.E. proved that a molten-salt reactor was viable.
Weinberg had hoped to move from the M.S.R.E. to a molten-salt breeder reactor. But in 1973 President Richard Nixon pulled federal funding for molten-salt research in order to go all-in on a competing breeder reactor that was cooled with sodium. In 1983, the sodium breeder, in turn, lost its funding. Plagued by budget overages, the project also fell victim to a conservative revolt, spearheaded by the Heritage Foundation. By then, the public had also soured on nuclear-energy projects, owing to the partial meltdown at Three Mile Island, in 1979.
There are various ways to index the downturn of the U.S. nuclear industry, but the starkest is probably by permits. From 1954 to 1978, regulators issued a hundred and thirty-three construction permits for civilian nuclear reactors. Between 1979 and 2012, they issued none. “There’s been almost no real work on nuclear power since the seventies,” Nathan Myhrvold, who sits on the board of TerraPower, a nuclear-technology company that he co-founded with Bill Gates, told me. “The Department of Energy still had some research programs going, and I don’t want to slight anyone who was in one of those things. But they stopped building plants. When you stop building plants, it makes it very hard for companies to justify the enormous number of engineers it takes.”
“The thing that struck me the first time we went to China, in particular, is they assigned an awful lot of people to the problem,” Charles Forsberg, a research scientist at M.I.T., told me. “And if you assign several hundred engineers to the problem you will learn very, very rapidly.” Forsberg spent his early career as a researcher at Oak Ridge National Laboratory before moving to M.I.T., where he is overseeing the construction of a molten-salt loop that will run along the side of the campus’s research reactor. He is also one of three engineers who, in 2002, hashed out the concept for a fluoride salt-cooled high-temperature reactor, or F.H.R. That involved taking Weinberg’s molten-salt reactor and swapping the liquid-fuel loop for a more conventional core design, while still using molten salt as a coolant. This change simplified the most vexing problems, of corrosion and containment, while preserving the high process heat that molten salt makes possible. The F.H.R. has played a significant role in rekindling interest in molten salt for fission reactors in the United States—which is the reason that Forsberg initially travelled to China to meet with the SINAP team.
Forsberg’s travel, and the relationship that he developed with Xu Hongjie and other researchers at SINAP, took place at the outset of a relatively recent period of collaboration between the U.S. and China. The partnership was formed under the framework of a 2011 “memorandum of understanding” between the Department of Energy and the Chinese Academy of Sciences, which provided for coöperation on nuclear technologies. That agreement was based on a previous agreement, from 2006, which had cleared the way for U.S. nuclear firms to sell reactors to China. Both stemmed from the desire of each country to leverage the other to revamp its own nuclear industry.
China had only a handful of reactors in the early two-thousands, but in 2007 its planners had vowed to massively increase nuclear-energy production by 2020. That meant building something like forty new reactors in about fifteen years—a pace and scale only matched by the U.S. nuclear industry in the twentieth century. To meet that goal, China intended to buy the first fleet of new reactors from foreign companies, under contracts that required significant technology transfer. Although this now looks like a mixed bargain, at the time, the U.S. nuclear industry was only too happy to take it. The industry had just weathered a quarter century of effectively zero domestic demand for new reactors, and had hundreds of experts unable to put their skills to use. These were “a bunch of old Navy nuke guys, or guys that studied nuclear engineering forty years ago, who knew a ton about aging management and cracking piping and corroding pumps and things like that,” David Fishman, then a partner at a boutique China-based nuclear consultancy, told me. “They were just so pleased to come over and find a young, eager market and industry that was planning to build dozens of reactors.”
U.S.-China coöperation on molten-salt research proceeded under conditions not so different from the commercial melee. Forsberg and his collaborators—Per Peterson, a nuclear engineering professor at Berkeley, and Paul Pickard, formerly of Sandia National Laboratories—had pursued their design through academia for years, using oil or water to simulate molten salt, which is expensive and difficult to acquire in the United States. Then, in 2011, they were awarded a major multi-university grant from the Department of Energy which ultimately allowed them to start running tests with the real thing. That became a useful point of connection for Xu and his team, who had recently received a major grant from the Chinese government. The SINAP group was created to build a liquid-fuel reactor, with hopes of eventually fulfilling Weinberg’s vision of making a thorium breeder. To create some common ground with the Americans, they also committed to building a salt-cooled reactor like the F.H.R.—the project that the D.O.E. was most interested in at the time.
A technician prepares salts for use in the M.S.R.E., at the Oak Ridge National Laboratory, in 1964.Photograph courtesy Oak Ridge National Laboratory
You can see the early dynamic of U.S.-China coöperation play out in a video of SINAP’s first presentation at Berkeley, delivered in August, 2012. As the institute’s representative, SINAP sent Kun Chen, who had done his Ph.D. at Indiana University and was still in his thirties. The audience skewed much older: about two-thirds of them looked to be in their fifties or sixties. The attendees tried to suss out the practicality of SINAP’s ambitious plan. One man asked about the budget, which was about three hundred and fifty million dollars, spread over five years. Another man asked where SINAP planned to get molten salt, since “to my understanding, there are no facilities in the world that can produce” it. Chen replied that China had several facilities that could.
It’s hard to tell from the video what the Chinese side got out of these exchanges, but when I spoke with Chen he stressed how helpful it was to have interlocutors in the U.S. “From the start, we didn’t believe we could get this far,” he said. Molten salt was no less niche in China than it was anywhere else. Chen estimated that, back in 2011, there were only thirty or forty people in the whole world working seriously on using the substance for fission reactors. Connecting with some of those individuals in the U.S. made the project seem possible.
For the Americans, there was the curiosity of seeing how far the Chinese could go with resources that simply didn’t exist here. Coöperating with SINAP was also a way to prod the U.S. federal government. The logic was “If the Chinese are doing it, it must be relevant,” Forsberg said.
In that sense, the coöperative research-and-development agreement that Oak Ridge signed with SINAP cut out the middleman. To fund the molten-salt loop, SINAP paid Oak Ridge around four million dollars, according to Chen. With such a loop, researchers could test materials and all the plumbing components needed to circulate molten salt. The project also gave a focal point to people working on molten salt in the U.S. Speaking to a reporter from the MIT Technology Review, David Holcomb explained his motivations. “One of the important things to realize is that a number of key people in molten-salt reactors are retiring very fast or passing away,” he said. “China is providing the funding that allows us to transfer that knowledge, to gain practical experience at building and operating these reactors.”
That article ran in August, 2016. By 2018, the U.S. had withdrawn from almost all coöperation with China. “I wouldn’t say it’s a total surprise,” Chen told me. He and the SINAP team figured that the relationship would probably deteriorate under Trump. “But it was just happening very suddenly. It’s similar to what we have learned in the tariff issue.”
I asked Chen if he ran into any challenges once his team was going it alone. “The challenges are, I think, mostly, first of all, if you have the money,” he said. But the SINAP team certainly had that. The Chinese Academy of Sciences had been extending the project’s grant every year. By 2018, China promised three billion dollars for molten-salt reactors over the next two decades, while Chinese planners have called for a $1.3 trillion investment in nuclear energy as a whole by 2050.
During Chen’s first presentation at Berkeley, in August, 2012, one of the few young people to ask him a question was a man with a shock of dark brown hair and an ample goatee. I had watched the recording several times before I realized the man was Mike Laufer, who would go on to help found Kairos Power, a privately held nuclear company that is attempting to commercialize the fluoride salt-cooled high-temperature reactor originally designed by Forsberg, Pickard, and Peterson, who is also a co-founder of Kairos. Once I recognized Laufer, his question to Chen, about “the biggest challenges or obstacles to overcome” in order to build a salt-cooled reactor, had a new resonance. Was Laufer, who at the time was a graduate student at the university, already putting together a business plan?
Kairos represents a new era for the U.S. nuclear industry. Inspired by SpaceX, it is effectively trying to rebuild U.S. industrial capacity within a single company. The business model calls for a vertically integrated network of facilities that can fabricate fuel and salt for Kairos, and can manufacture a large share of what the company needs to build its reactors. The hope behind all this is that by running things internally Kairos will be able to offer nuclear energy at a competitive price in the market. And it has had some success. Last year, Google committed to buying five hundred megawatts from the firm by 2035. Kairos is also one of only two U.S. companies with a permit from the Nuclear Regulatory Commission to build a new reactor. Construction of the reactor building, located in Oak Ridge, broke ground last year. “We’re working to get that reactor up and running this decade,” Laufer told me.
In getting to this point, Kairos initially benefitted from U.S. partnership with China on molten-salt research, and is now reaping the rewards of the recent pro-nuclear turn in American domestic industrial policy. The money that China put into U.S. research in the early twenty-tens pushed development of the fluoride salt-cooled high-temperature reactor from theoretical work into practical experimentation, and the salt loop that SINAP paid for at Oak Ridge National Laboratory yielded a report of molten-salt pumps, which dovetailed with one of Kairos’s early priorities. For several years after the Trump Administration ended nuclear coöperation with China, there was little to replace Chinese money in the U.S. nuclear industry. But big public spending eventually started coming, along with growing private investment. In 2020, Kairos was awarded a three-hundred-and-three-million-dollar grant from the Department of Energy, and with other young nuclear companies it benefited tremendously from a thirty-per-cent-investment tax credit for clean energy contained in the 2022 Inflation Reduction Act. Trump’s so-called One Big Beautiful Bill sunsetted credits for solar and wind early, but the Senate ensured that nuclear energy would keep them.
I asked Laufer if he was worried about competing with China. “At the moment, what we’re trying to do is challenging enough,” he said. ♦