Can America get to yes on a new reactor design?

Everyone says they want the NRC to approve advanced nuclear reactors, but it keeps not happening

Back in November, I wrote about Oklo, a California-based company that was midway through the process of (hopefully) getting the first-ever license to operate a commercial nuclear microreactor.

Unfortunately, what happened instead is that the Nuclear Regulatory Commission denied their application and said they need more information before they can reconsider.

I spoke with the Oklo co-founders, and they seem confident they can provide the requested information and that the NRC will grant the license in the end. But they also say they are frustrated by a lack of communication over the course of this very long process. The 1.5 megawatt Aurora reactor in question is very small, but it’s not as if it’s being constructed in someone’s backyard. Oklo’s business plan is to build and sell lots of small reactors, but the reactor in question is being built in partnership with the Energy Department’s Idaho National Lab with the idea that INL will provide Oklo with the reprocessed nuclear waste that serves as Aurora’s fuel.

Stepping back from the particulars of the regulatory process, it is the policy of the United States government to promote a new generation of advanced nuclear reactors. That’s why Congress passed the Nuclear Energy Modernization and Innovation Act in 2019 directing the NRC to develop a pathway for licensing small reactors. It was the Trump administration’s policy to promote advanced nuclear reactors, it was the Obama administration’s policy to promote advanced nuclear reactors, and it is the Biden administration’s policy to promote advanced nuclear reactors.

Secretary Jennifer Granholm @SecGranholm

Nuclear energy is essential to confronting the crisis of our lifetime: climate change. It’s time we #BuildBackBetter with nuclear innovation.

But the NRC has not actually created the new regulatory process. Nor have they approved any applications for advanced nuclear reactors. And what gives people in the industry heartburn is that the NRC has, in fact, never approved a nuclear reactor from start to finish since its creation in 1975.

The worry is that without either a clear process — you must show X, Y, and Z for approval — or a successful regulatory precedent, the NRC will always have more questions. Unless something is actually built and operating, there are going to be unknowns. But we approve (through separate regulatory processes) things that we know kill people with pollution all the time. We really need a clear way of getting to yes on nuclear.

The NRC has no track record of yes

Back when the United States built nuclear power plants, they were regulated by an institution called the Atomic Energy Commission.

The AEC had the dual role of developing nuclear weapons and promoting the safe use of civilian nuclear power. That was a slightly odd combination, so in the 1974 Energy Reorganization Act, Congress divided military and civilian responsibilities. Nuclear defense is now handled by the Energy Research and Development Administration (an arm of the Energy Department), while the Nuclear Regulatory Commission serves as an independent regulator of civilian nuclear plants. But as part of the reorganization, NRC was given a pure safety mandate rather than a mandate to promote the safe use of nuclear power.

The NRC really takes this single mandate both seriously and literally, and safety regulations became even more stringent in the wake of the Fukushima disaster.

But rather than demanding extra safety redundancies without conducting any cost-benefit analysis, I would have asked for regulatory ideas that would save lives on balance — taking into account the fact that raising the cost of nuclear and inducing utilities to switch to gas creates its own pollution and safety risks. Any idea that would increase the cost of operating nuclear plants by so much that some of them shut down in favor of burning natural gas has a real downside. The extra deaths due to burning more gas might be worth it in safety terms. But that’s the question that needs to be answered: does this idea save lives on net or does it only minimize lives lost specifically due to nuclear accidents? Because saving one person from one kind of accident only to kill 10 in another kind of seems like a bad idea to me. But that’s not the NRC approach; they just made things way more expensive without doing cost-benefit analyses.

Thirteen people have died in nuclear power accidents across the history of the United States.

And most of these are accidents that happened to occur at a nuclear power plant, not anything to do with reactor malfunctioning. In 1987, a safety inspector in Kansas touched a mislabeled wire and died by electrocution. On March 31, 2013, a team of workers was moving a large piece of machinery at Arkansas Nuclear One and it fell, killing one worker and injuring two others. Every death is a tragedy, but these accidents aren’t unique to or particularly common at nuclear facilities. There were about 4,700 fatal workplace injuries in 2020, a historic low that includes 19 deaths at utilities.

Meanwhile, air pollution is the cause of a bit over 100,000 premature deaths each year in the United States, so the cost of shifting at the margin from nuclear to fossil fuels is very high. And while natural gas is low-pollution compared to coal or oil, a single natural gas explosion in 1944 killed roughly 10 times as many people as all U.S. nuclear accidents combined. Even the high-end estimate of the worst nuclear accident ever is much smaller than the annual death toll due to fossil fuel plants operating as designed.

Nevertheless, the “safety first” approach at the NRC has been so successful that they’ve actually never approved a reactor that was then built.

A few designs that were already in the works when the AEC was dissolved were completed in the NRC era. And back in 2012, they approved a couple of new reactors at an existing facility in Georgia. But due to construction delays and spiraling costs, it’s not clear whether they will actually be built.

From a super literal view of making sure that nobody ever dies in a nuclear accident, making sure that nobody ever builds a nuclear plant does work. But this costs lives on net and makes climate goals very hard to achieve.

Small is beautiful

Consider all the different times that environmental groups end up opposing renewable energy projects. To an extent, this is them being short-sighted and foolish, like when the local chapter of the Sunrise Movement came out in favor of a moratorium on solar projects in Amherst, Massachusetts.

But utility-scale wind and solar plants really do take up a lot of space.

There is no truly free lunch in the power generation space. Wind and solar are renewable in the sense that you don’t need to dig stuff up out of the ground to create electricity. But you do need a much more expansive footprint to capture the wind or sunshine. And this is, I think, a bigger problem than a lot of people realize. America’s largest solar project was killed by local NIMBYs (some environmentally minded, others less so), not by mustache-twirling fossil fuel barons.

Personally, I am very worried about climate change and air pollution and fired up about energy abundance, but I’m not much of a nature lover. So at the end of the day, if we need to cover huge swathes of open space with solar panels, wind turbines, and transmission lines, I’m all for it. But realistically, not everyone feels that way, and the margin does matter. We should cut down some woodlands in New England for the sake of clean energy, but probably not all of it. It would be really nice to get a decent chunk of electricity from microreactors that have small footprints and could fit into the built landscape or be situated in natural settings with minimal disruption.

A hubristic effort to explain nuclear physics

In addition to being convenient, the small size of advanced reactors plays a key role (as I understand it) in their safety. Nothing in life is accident-proof. But the idea behind the microreactor’s design is that it really is immune to catastrophe. It could break, but the damage that could cause the external world is exceptionally limited.

Now of course I don’t have a PhD or an engineering degree. And if the people at the NRC who do have those things were saying that the engineers and PhDs at Oklo are totally full of shit and that’s not how high-assay low-enriched uranium works, then I’d have to be an honest journalist here and admit that I’m not fit to adjudicate. But scientists at the Idaho National Lab think the design has merit. And the NRC seems to as well. They didn’t reject the application out of hand or with prejudice the way they would if the basic claims were absurd or wrong. They just say they want more information. And if six months from now it’s all cleared up and we’re off to the races, then that’s fine.

But the concern, again, is that the NRC has actually never successfully shepherded a project to completion, and unlike Congress or the Obama, Trump, and Biden administrations, they don’t particularly seem to see an upside to promoting nuclear development or a downside to delaying. And that’s a mistake.

The big question — economics

I do think there are questions about this and other reactor designs where the answer genuinely is much more unclear, but they’re about economics, not safety.

The pitch of Oklo and others in this space is that they can build a reactor at INL to demonstrate that the technology works, then build a few for people who have eccentric use cases (remote communities dependent on diesel, companies that for PR reasons want to say they’re zero-carbon, even at an elevated price). As production scales up, unit costs fall. The notion is that since the reactors are small and standardized rather than giant and site-built, you can make them in a factory and improve your utilization. And there are lots of examples you can point to — like Tesla with electric cars and many different manufacturers with solar panels — where this has worked.

Something starts expensive to build but becomes cheaper thanks to learning-by-doing and mass production.

But of course if it were that easy to start with a high-end hardware product and end up with a cheap mass-market one, everyone would do it. The challenges involved in actually executing on this vision are extremely non-trivial, and for certain kinds of products, they might be insurmountable. I have no way of knowing if Oklo will in fact be able to scale up production in a cost effective way or whether any of their competitors will. But the only way to find out is to start letting some people build some stuff, and to recognize from a regulatory standpoint that the upside to doing that is actually quite large. Whatever the risks are, they need to be balanced against the benefits.