Innovation is a messy business. With the benefit of hindsight, successful innovators are visionary seers while those who fail are tragically flawed: myopic, delusional, or incompetent, victims of forces they never saw coming or intractable problems they should have anticipated.
Post-hoc, it all seems so obvious. But that is not the case when one’s gaze is directed in the other direction, from the present toward the future. For innovators, and for those who would evaluate their prospects, whether investors, policymakers, or advocates, it’s not so clear what about the world is contingent and what is immutable, which trends are hard-wired into our sociotechnological reality and which might, at any moment, tip into a new equilibrium.
I was reminded of this truism again upon hearing news that Leslie Dewan and Mark Massie, founders of the nuclear startup Transatomic Power, had announced that the company was suspending operations. Not so long ago, the pair were the face of the advanced nuclear community—millennials and recent graduates of MIT nuclear engineering who had chosen to forego careers in both academia and the nuclear industry to launch a startup that promised to change the world with a simple salt-cooled reactor that burned nuclear waste and couldn’t melt down. In the brief that argued that a nuclear future didn’t need to look like the past, that entrepreneurs and innovators—not national laboratories, state-owned enterprises, and behemoth incumbents like Westinghouse and General Electric—might rescue the technology from decades of stagnation, Leslie and Mark were Exhibit A.
It didn’t work out for Transatomic. Much of the company’s promise was predicated on a calculation in an early design whitepaper that turned out to be in error. Subsequent engineering showed that the reactor still worked. But it wouldn’t run on unreprocessed waste, which had been the company’s biggest selling point.
Yet in failure, Transatomic is itself proof of concept. For those of us who have argued for a different nuclear future, less centralized and state-dependent, companies like Transatomic will need to fail so that an advanced nuclear industry can thrive. A competitive and innovative advanced nuclear industry, independent of both the national security state and significant public subsidies, must allow for companies and technologies to fail without catastrophic political and economic consequences for the industry. And that is exactly what happened. Transatomic’s investors are out of luck, but taxpayers and ratepayers are not. That is as it should be.
Today, there is an industry, not a single technology, with dozens of companiesdeveloping different combinations of fuels, coolants, and reactors in different sizes and targeting different markets. NuScale Power, a small modular light water reactor company, is well on its way to being licensed by the Nuclear Regulatory Commission (NRC), has a memorandum of understanding in place to build its first twelve reactors, and just announced that it has awarded its first contract to manufacture its reactor. Several other reactor designers have initiated discussions with the NRC to begin licensing. They won’t all succeed. But given a fair shot at licensing and commercialization, it’s not unreasonable to think that at least a few of them will.
Serious political and economic obstacles remain. Licensing remains burdensome and commercializing a new nuclear reactor is a costly business. But policymakers have begun to respond. Earlier this month, Congress passed the Nuclear Energy Innovation Capabilities Act (NEICA), directing the Department of Energy to share information on advanced reactors with NRC and providing funding to help offset some of the licensing costs. The Department of Energy’s GAIN program is providing support for advanced reactor companies to access the treasure trove of National Laboratory experiments, data, and expertise on advanced nuclear materials and fuels. Legislation just introduced in the current Congress directs DOE to construct a fast reactor test facility, where developers could test new materials, and allows the federal government to enter into long-term power purchase agreements with nuclear generators.
But for advanced nuclear to succeed, we will need a lot more failure. This week, we’ve released a new whitepaper with the R Street Institute and ClearPath, two right-of-center Washington-based policy centers, called “Planting the Seeds of a Distributed Nuclear Revolution.” We call for expedited licensing and commercialization of micronuclear reactors (reactors that are less than 10 MW-thermal in size, or less than a hundredth the size of a typical large light-water reactor). Because they are so small, microreactors are much simpler than larger reactors. Accidents are both much less likely and much less consequential. Dozens of small research reactors have operated for decades in university and National Laboratories around the country, some of them in the middle of major cities.
But for the fact that research reactors are not connected to the electrical grid, they are not different from a number of microreactor designs that companies around the world are seeking to commercialize. There is no reason that these very small reactors should have to go through a licensing process that can take a decade and cost a billion dollars. Nor should they be subject to security and safety requirements that similar sized reactors sited on college campuses in the middle of Cambridge or Austin are not subject to.
Microreactors are also less costly to commercialize and promise much faster technological learning. Our work on nuclear costs and learning rates, along with that of many others, has demonstrated that standardized designs and building in multiples are the key to reducing the cost of nuclear reactors. The more you do the same tasks, over and over again, the better you get at doing them.
Historically, standardization and learning by doing have been limited by the large size of reactors. Nuclear developers assumed that economies of scale would make larger reactors cheaper. Because of everything else that came with the package—large security perimeters, a dedicated control room, cooling towers and backup generation—larger reactors would cost less. The nations that have most successfully deployed nuclear energy, like France and South Korea, did so by settling on a single national reactor design that was both large and standardized. The catch, however, was that to get lots of experience building large reactors, you functionally needed to nationalize your electrical grid.
Moreover, the theory that larger reactors would prove to be cheaper hasn’t worked out. The complexity and safety systems necessary to operate large reactors safely outweighed the benefits of scale. It also appears that today at least, first-of-a-kind (FOAK) small reactors won’t cost any more than first-of-a-kind large reactors. NuScale expects that its first twelve 50 MW modular reactors will cost no more per kilowatt of capacity than will the two large AP1000 reactors currently under construction in Georgia.
In places like the United States, western Europe, Japan, and even India, where nationalization of the power sector is unlikely, the only way to get experience building multiple reactors, and to fully benefit from manufactured components and mature supply chains, is to go small, not large. All the more so given the reality that any FOAK reactor, large or small, is likely to cost well above the market rate for new electrical generation and require some significant public cost sharing. Even if a FOAK small reactor costs substantially more per KW deployed, the public subsidy necessary to get that reactor to market will cost substantially less than the subsidy necessary to build a large reactor.
The case for microreactors is even stronger. At the same public cost of building a single FOAK large reactor, we could deploy dozens of microreactors, supporting commercialization and learning for multiple reactor designs, and allowing the best technologies and companies to succeed, simultaneously deploying multiples of each microreactor in order to support robust technological learning. And because of the safety characteristics inherent to very small reactors, the balance-of-system costs historically associated with large reactors should be much less significant with small reactors. The NRC just agreed with small modular reactor developers that Emergency Planning Zones can be scaled to risk. For the NuScale reactor, that could simply be the perimeter of the power plant site, which would eliminate the twelve square kilometer perimeter required for large light-water reactors. Smaller microreactors should benefit from these dynamics even more so.
With an expedited licensing pathway and competition for public deployment contracts for microreactors, it is possible that some advanced nuclear developers will downsize their builds to fit the microreactor criteria. We make no claim that microreactors will prove to be the optimal size, only that they are the optimal size for initial commercialization. At the point at which microreactors have been commercialized and their design, manufacturing, and supply chains well-established, licensing and commercialization of larger designs should be much more straightforward.
At the end of the day, of course, we and our coauthors might be wrong. Perhaps the United States will manage to build transcontinental transmission capacities that will make high shares of intermittent renewable energy technologies possible. Or we will crack the code on long-term seasonal electricity storage. Or very cheap carbon capture, storage, or removal technologies will materialize, allowing fossil fuel combustion to play an ongoing role in low-carbon power systems. Or climate change or some other geopolitical imperative will lead us to reconsider the wisdom of liberalized electricity markets and start building large publicly financed light-water reactors centrally—the way we have historically built other critical public infrastructure.
But it is also possible that none of those things will happen. Long-distance transmission has proven no easier to site and build than large nuclear plants. Storage or carbon capture at scales consistent with fully decarbonizing the power sector are still more speculative than real. Fossil fuels are cheap and abundant and climate change remains a relatively low priority for both policymakers and the public.
Absent one or more of those things changing, an innovative advanced nuclear sector, responsive to the economic and institutional realities of the US power sector, is among our highest hopes. That will require policy action, both to reform the licensing process and provide modest public support for the commercialization of first-of-a-kind reactors. It is heartening that our coauthors, both with strong libertarian priors, have recognized that the latter as well as the former will be necessary and that microreactors offer the most cost-effective means of doing so.
In the face of deep uncertainties, about the long-term consequences of climate change and the future of clean energy technology, offering the advanced nuclear sector a chance to be part of the solution is the least we can do. But without real companies building real reactors, that option will remain entirely speculative. A focused effort to commercialize small reactors, in our view, is likely the most viable path to making that promise a reality.