This spring the Nuclear Regulatory Commission approved a design from Oregon’s NuScale Power for a 77 megawatt nuclear reactor. The company is already underway designing for a site in Romania and says they have received interest from numerous U.S. companies. Amazon has said it wants to build small-scale nuclear reactors along the Columbia River in the Pacific Northwest to power AI data centers.
We talk to José Reyes about the company’s design and their vision for what small-scale nuclear power could look like.
Note: The following transcript was transcribed digitally and validated for accuracy, readability and formatting by an OPB volunteer.
Dave Miller: This is Think Out Loud on OPB. I’m Dave Miller. The Oregon-based company NuScale Power had once been the clear leader in the small modular nuclear reactor world. In 2023, [it] became the first American company to get federal approval for one of its designs. It faced some challenges after that, including the demise of its plan to provide nuclear power to 46 utilities in Utah. Now, it seems to be riding high once again. In May, the company got federal approval for a second reactor design, one that would produce more electricity. The company is also moving forward on a site in Romania and says they have interest from numerous U.S. companies. One of the big reasons: power hungry AI-focused data centers.
Tomorrow we’ll hear from Columbia Riverkeeper, a group that is opposed to the expansion of nuclear power in the Northwest. But right now, I’m joined by José Reyes. He is a co-founder and the chief technology officer at NuScale. Welcome to Think Out Loud.
José Reyes: Thanks, Dave. Thanks for having me on your show.
Miller: What do you and others mean when you talk about the “small” and the “modular” of small modular reactors?
Reyes: That’s a great question. That term, small module reactors [SMRs], was first used and popularized by NuScale. So this is an industry that started right here in Oregon. And now SMRs are being developed across the globe and they all use the same term: small modular reactors. In terms of what we’re doing by “small,” we mean about 77 megawatts of electricity per module. So we build these factory assembled modules.
Just to describe it a little bit, it’s about 57 feet in length, about 10 feet in diameter – that would be the reactor. And everything you need to produce steam is inside that reactor. The reactor sits inside of a steel containment vessel. We evacuate that vessel, so it’s a vacuum. We take that reactor and containment, put it in a pool of water, a stainless steel lined pool, below ground. So it’s a reactor inside of a steel thermos bottle, basically, underwater, underground. That’s the basic technology.
Miller: How does 77 megawatts of electricity – and I should say that is the level of electricity that one of these reactors from the newer design that was just recently approved – compare to the kinds of nuclear power that we’ve had from, I don’t know, the ‘50s or ‘60s onward?
Reyes: You had two sizes of reactors in the past; they were large and extra large. So they were 1,000 megawatts and growing, up to almost 1,500 megawatts in the latest designs. This is much, much smaller. So six of our modules, for example, would be about 462 megawatts, about half the power output of one of the large reactors. Our designs come in four modules at 308 megawatts, six modules at 462 [megawatts] and 12 modules at 924 megawatts, so almost a gigawatt class. What’s nice about this, of course, is the flexibility. You can start with one module, add modules as modules are needed.
Miller: What does the most recent approval from the Nuclear Regulatory Commission – that came in May for your second slightly bigger design or slightly more powerful design – mean in practice?
Reyes: It’s a significant achievement. The teams worked very hard to get this design approved. We are the only SMR design that’s been approved by the Nuclear Regulatory Commission, so that gives us a big advantage because we’re ready to deploy now. So when we talk about projects in the future, we’re talking about an approved design that only requires a construction permit or a construction operating license application for the site. So once you have an approved site, we have an approved design. It’s a marriage made in heaven, basically, and we’re able to go ahead and start building that plant.
Miller: OK, but you make it seem like it’s easy peasy to build a new nuclear plant in the U.S. But everything we’ve heard for decades seems to show otherwise, that there are always cost overruns and it can take literally decades for these to happen. Sometimes they just don’t happen at all because of those delays or cost overruns. What is your best case scenario for when one of your designs would actually turn into an operating nuclear power plant in the U.S.?
Reyes: The reason I started the company to begin with was because of those challenges that I saw. By going to a factory manufactured reactor and containment vessel … and that’s key. So instead of the big concrete dome, we go to a very small factory manufactured containment vessel. We can do that in parallel with civil construction. Just by doing that, we go from a seven-year schedule down to about a three-year schedule from first support of safety grade concrete. So while you’re manufacturing your modules, in parallel, you’re doing all your excavation, you’re building your buildings. And once the building is done, then you can just install the modules. So you’re basically just installing equipment at that point. That greatly reduces the time and the cost of building a NuScale plant.
Miller: How much local opposition to one of the plants … because a lot of what you’re talking about here seems to be in the order of engineering or technical challenges, but my understanding is that there are also social challenges. To what extent could those get in the way or prevent the construction of one of these series of modular reactors?
Reyes: We’re seeing an overwhelming interest and support for nuclear power right now. It’s an exciting time because what’s driving it are two things. One is climate change, so people are looking for large quantities of base load carbon-free power. So that’s driving it. And more recently, it’s been amazing. Five or six years ago, we talked to data centers and they said, “Well we need 50 megawatts, 100 megawatts.” Now they’re all coming to us saying, “We need 1,000 megawatts. We need 2,000 megawatts.” So it’s really an unanticipated shift in the demand for power. So I think that’s what’s really driving it. And we’re seeing that.
In terms of the concerns of the past that people had with nuclear power, you have a generation now which understands what these new designs are like. So it’s certainly worth an effort to look at what’s different about the new designs versus what’s been built in the past, because it is very different. Just look at our safety, look at the cost, look at the way in which we would deploy a plant. So there’s a lot of great things that are happening, but I think a part of it involves getting folks familiar with what’s different with our design. A lot of progress has been made.
Miller: You say, “just look at our safety and now we have people who understand that these are different,” but these don’t exist yet, right? I mean, are there any sets of these fully up and running anywhere in the world right now?
Reyes: No, the first ones we’re looking at right now are Romania. But we’ve spent about $2 billion de-risking the licensing, de-risking the technology, de-risking the operations. We’ve built a full-scale control room simulation that allows us to actually start practicing operations with our plant. So we’ve done a lot to assure that this is a safe and an effective design.
The big thing, I think, in terms of getting people familiar with the plant though, we’ve got what we call our E2 Centers. These are Energy Exploration Centers. So we’ve deployed these now in 11 universities. So students are now getting hands-on experience with this design, even before it’s built. But it includes all the procedures, all the processes, everything that you’d see in the actual plant. So we’re learning as we’re going, in terms of the operations. We’re getting a lot of good feedback. But it’s all been reviewed and approved by the regulator.
Miller: I mentioned the big setback that came in November of 2023, when a project to provide nuclear power to dozens of utilities in Utah fell through. What went wrong?
Reyes: The contractual structure was a challenge. So you had these different municipalities that had to vote every time there was a change to the subscription. In the end, NuScale and its partners were able to meet the price targets that they had required. We did what’s called the economic competitive test every year and we met those every year. But they were not able to get enough subscriptions. So in the end, I think they only were able to get 120 megawatts of subscription for a 462-megawatt plan. So at that point it was mutually decided that, well, we need to go our separate ways. Because unless there’s enough subscription, they won’t be able to buy the plan.
Miller: What business lessons did you take from that?
Reyes: Yeah, one of the big lessons, of course, is that it’s very difficult to work with multiple municipalities, multiple utilities, especially if they can jump in and jump out of projects. And it was very, very challenging. On the very positive side is that we learned so much in terms of interacting with the customer. We learned about how to share this information with the public. I think we had well over 120 visits to town halls and to different locations to share about the technology. So that was very positive.
We also learned of and studied the mistakes of previous designs, the large plants, to make sure that you have a design that is very complete before you start building. So that was one of the lessons learned that we took away from previous construction projects of large plants.
Miller: Going forward, do you see your business model more being the direct provider of electricity to companies that run data centers as opposed to utilities?
Reyes: It’ll be a mix of both. What we’re seeing is that because of our unique approval, [being] the only design that’s been approved for off-grid operations, we don’t need any AC or DC power for safety. So we can actually operate in island mode and provide direct power to a data center. Or we can operate connected to the grid and provide power to the grid, which then provides power to the data center. So it’s tied a lot to the local regulations and what they’ll allow us to do.
But we’re getting interest from all the data centers. We’re also getting interest from petrochemical plants who need lots of processed heat, steam. So we can produce 800,000 pounds of steam per hour for their processes.
Miller: So nuclear energy could be a way to keep using fossil fuels?
Reyes: Well, many of these large companies have made commitments to reduce their carbon emissions by 2040 and 2050. So they’re looking for clean ways of producing chemicals as well as other petrochemical products that are needed, either for the pharmaceuticals or for a whole range of chemical systems.
So that’s one aspect. The other is hydrogen. One of our modules could produce 50 tons of hydrogen per day. And today, clean chemical plants use 250 to 500 metric tons of hydrogen per day. So this would be a six-module plant or a 12-module plant that would work very well off-grid, with those types of sites.
Miller: I’m curious to get your response to one of the critiques I’ve seen, which is focused on data centers and on AI. I think it’s sharpened by the news six months ago or so that Chinese AI models seemed to be about as good as the super power-hungry American ones, even though they used a fraction of that electricity.
That led a lot of folks to say, “Why are these Silicon Valley companies demanding so much more electricity in the future?” And the anti-nuclear argument, which is tied to that, is why give them all this electricity that they’re asking for if it’ll mean that they won’t be forced to be as energy efficient as they might otherwise be?
What’s your response?
Reyes: Yeah, I know. These data centers are very focused on energy efficiency because that gives them the competitive edge. So they want to reduce their costs as much as possible. But it will require that they be as efficient as possible, not only in terms of the power usage, but in terms of water usage, in terms of the designs of their facilities. AI is very energy intensive and as I speak to different data centers, they show me plots of their energy requirements. You see big spikes and base load requirements, and it really can only be provided by a stable source of power like nuclear.
Miller: Did the news out of China, about their models that did use a lot less electricity, change the way you think about your business?
Reyes: No, I don’t think it did. We still look and we track just data centers in terms of what’s being used on the Cloud and internet providers. That continues to grow at a tremendous rate, even without AI. So AI is just kind of the icing on the cake. But eventually, it’ll become a dominant part of the cake, let’s say.
Miller: There’s also the much more longstanding argument against more nuclear power, which is that until we have a really good permanent containment plan for radioactive waste, we shouldn’t be creating a lot more of that waste. What’s your response?
Reyes: If you look at what’s being done in other countries, in France they’ve been recycling their fuel for it must have been 30 years at least. So the technology there to do the recycling is available. The U.S. has excellent technology. Right now, it’s the best practice to use the fuel, typically in three cycles. So one fuel assembly will go into a core and then it’ll get shuffled and then used again, and shuffled and used again. Then it goes into a pool of water for about five years. And then it moves to a dry cast storage, which is just air cooled. It’s contained in a concrete steel shell and just air cooled. That’s the best practice today in the United States.
But the volume that we’re talking about is relatively small. So for example, one of our six module plants will run for 60 years producing power at 462 megawatts. At the end of 60 years, all the used fuel from that plant will fit on 0.8 acres. So it’s a very small amount, a very small footprint, in terms of the amount of used fuel – and that’s 60 years producing power continuously, on 0.8 acres. So we’re looking at very small quantities.
My own personal opinion is that eventually, when the cost of uranium goes up, someone will have the “aha” moment, “oh, we should recycle.” Because, while uranium is inexpensive, I think recycling then doesn’t look as favorable. But when the cost of uranium goes up, then I think suddenly someone will say, “well, we should be recycling.”
Miller: That’s a six-module plant that you’re talking about. In your dream, or your hope, how many of those would there be, say, 30 years from now?
Reyes: Oh my, well … I don’t know if I can count that high.
Miller: Maybe that’s the answer. You mean like thousands more?
Reyes: I think thousands more, yeah. As we become efficient … Right now, we’ve got the first six modules being manufactured in Tucson. We’ve ordered the long lead components for the next six. I think the interest is significant, globally. So we’ve been working with our developer, ENTRA1, which is an energy developer. What they do is to provide the financing. They set up for the construction of the project and then they’ll build it. They can either own and operate, or they can own and transfer, depending on what the customer wants.
And for data centers, that’s key. Many data centers don’t want to own or operate a nuclear power plant. So having this ENTRA1 model allows us to say you don’t have to own it or operate it. We just sign a power purchase agreement and we’re good to go. So, I think there’s opportunities globally because of the ease at which now, you don’t have to own or operate the plant, and you can just sign a power purchase agreement.
Miller: José Reyes, thanks very much.
Reyes: Thank you, Dave.
Miller: José Reyes is the chief technology officer and one of the co-founders of Oregon-based NuScale.
“Think Out Loud®” broadcasts live at noon every day and rebroadcasts at 8 p.m.
If you’d like to comment on any of the topics in this show or suggest a topic of your own, please get in touch with us on Facebook, send an email to thinkoutloud@opb.org, or you can leave a voicemail for us at 503-293-1983.
