Topics
Sections
More
For IEEE Members
For IEEE Members
IEEE Spectrum
Follow IEEE Spectrum
Support IEEE Spectrum
Enjoy more free content and benefits by creating an account
Saving articles to read later requires an IEEE Spectrum account
The Institute content is only available for members
Downloading full PDF issues is exclusive for IEEE Members
Downloading this e-book is exclusive for IEEE Members
Access to Spectrum 's Digital Edition is exclusive for IEEE Members
Following topics is a feature exclusive for IEEE Members
Adding your response to an article requires an IEEE Spectrum account
Create an account to access more content and features on IEEE Spectrum , including the ability to save articles to read later, download Spectrum Collections, and participate in conversations with readers and editors. For more exclusive content and features, consider Joining IEEE .
Join the world’s largest professional organization devoted to engineering and applied sciences and get access to all of Spectrum’s articles, archives, PDF downloads, and other benefits. Learn more about IEEE →
Join the world’s largest professional organization devoted to engineering and applied sciences and get access to this e-book plus all of IEEE Spectrum’s articles, archives, PDF downloads, and other benefits. Learn more about IEEE →
Access Thousands of Articles — Completely Free
Create an account and get exclusive content and features: Save articles, download collections, and post comments — all free! For full access and benefits, subscribe to Spectrum .
The U.S. Nuclear Plant Fleet Is Leaning Into AI
AI assistants can help a shrinking specialized workforce maintain aging power plants
Andrew Moseman is the online communications editor at Caltech and a freelance contributor to IEEE Spectrum.
The team at Atomic Canyon, one of the leading companies integrating AI into the nuclear industry, stands in front of Oak Ridge National Laboratory’s Frontier supercomputer.
If there is any place that would be off-limits to AI, the control room of a nuclear power plant certainly sounds like one. The nuclear industry is historically cautious and risk averse—understandable given the possible catastrophic consequences of an accident or a mistake. Even well-trained managers struggle with the operational complexity of a nuclear reactor . It’s not the kind of setting that seems well suited to a powerful but error-prone new technology.
Reality tells a startlingly different story. A variety of companies have begun to offer AI solutions for nuclear power. This year, nearly the entire fleet of 94 U.S. nuclear reactors has been offered the chance to integrate AI into its operations, and most have taken it.
In August, California-based Atomic Canyon launched NIVA , the Nuclear Industry Virtual Assistant, which was developed in cooperation with nuclear-industry groups. The assistant was pilot-tested in nuclear plants run by Constellation Energy and is now available to the United States’ entire fleet of reactors. (Constellation declined to be interviewed for this story.) Meanwhile, the AI-for-nuclear startup Nuclearn says its products have gone to work with more than 65 U.S. partners, a total that includes many traditional plants but also an integration with NuScale Power Corp. that is working toward advanced small nuclear reactors . Microsoft and Nvidia have collaborated on a project to apply AI toward optimizing “the entire life cycle of a nuclear plant, spanning site permitting, design, construction, and continuous operations.”
The reason for the rapid adoption, according to Nuclearn CFO and cofounder Jerrold Vincent , is twofold. First, the technology companies banking their futures on AI are also keen on a revival of nuclear power to provide the electricity for all those data centers . Vincent says AI is also ideal for addressing some of the nuclear industry’s most pressing problems. Many of the things that make reactors so difficult to operate—intricate regulations, along with complex systems for management and maintenance—are exactly the kinds of tasks that AI is good at. The industry is staggering to deal with aging hardware and an aging workforce. If it is going to survive, and have a real shot at a nuclear revival, it will need help from machine intelligence , Vincent says.
“There’s this huge growth in demand from advanced-reactor developers and in new nuclear,” Vincent adds. “But the existing workforce has been spread really thin. So most of the time when we’re coming into a customer, [they're saying,] ‘I’ve had jobs open and I can’t fill these positions. We need to do the things required by regulation. How the heck are we going to do it?’”
AI Streamlines Nuclear Documentation
Right now, the principal output of a nuclear reactor (aside from electricity) is paperwork. Nuclear plants produce piles of documentation to demonstrate regulatory compliance and to record how they respond to any potential problem, from a crack in the sidewalk to a fault in the reactor.
“Every engineering decision has to be evaluated, independently verified, and checked,” Vincent says. Nuclearn’s AI products help plants search through large repositories of data more efficiently to more quickly file paperwork with the federal government or figure out how a particular problem was solved in the past. “If you think of it as ChatGPT for nuclear,” he says, “it’s actually not a bad place to start.”
Rob Austin , the Electric Power Research Institute’s leader for the NIVA project, says the project started a year ago when Constellation chairman Joseph Dominguez issued a challenge to the nonprofit industry groups that act as storehouses of information about regulatory compliance, maintenance and repair, and mechanical history. He asked the groups to find a way to make that industry knowledge available with the convenience of a large language model like ChatGPT, Copilot, or Gemini but keep it secure at the same time.
The task fell to Atomic Canyon, whose CEO Trey Lauderdale previously worked to integrate AI into health care, another conservative industry. Living and working downstream from Diablo Canyon, California’s last operating nuclear plant, he began to see how nuclear energy and AI could form a mutually beneficial relationship: A revival of nuclear power could help feed electricity to energy-hungry data centers, but only if AI could help the industry thrive again.
“The goal was to make decades of operating experience, hundreds of thousands literally of operating experience reports, tens of thousands of technical documents, and a whole lot of information available for people to find.” —Rob Austin, Electric Power Research Institute
A major hurdle is the public’s opinion of nuclear power, which soured decades ago after notorious accidents and worries about waste and only recently has begun to recover . But the resulting halt in construction of new nuclear plants has created another hurdle of its own: a loss of institutional knowledge, Lauderdale says. As the industry stalled in the U.S., fewer people entered careers as nuclear scientists or technicians. The result is a people problem. Nuclear power has a remarkably skilled but quickly shrinking workforce tasked with keeping old reactors online and possibly starting up new ones to meet surging demand for electricity.
“There is absolutely no way we as an industry will be able to have this nuclear resurgence or nuclear renaissance without augmenting the capacity of this knowledge-based workforce,” Lauderdale says.
For him, the answer is an AI that can intelligently search the repository of data that the nuclear workforce has created over the decades. When Atomic Canyon won the NIVA contract in 2025, the first step was to download 53 million pages of publicly available data from the Nuclear Regulatory Commission to train the model on the operations, vernacular, and regulations particular to nuclear power. Lauderdale then struck a partnership with Oak Ridge National Laboratory to build the first nuclear-specific AI models .
“Technically, they’re sentence-embedding models,” Lauderdale says. “Fancy way of saying, we taught AI how to seek the nuclear language. Nuclear is almost like its own vernacular.” The goal is that today’s nuclear workers can ask questions about every aspect of plant operations, from filling out regulatory forms to procuring the right tools and materials, simply by writing a query and letting the AI sort through all of the accumulated institutional knowledge.
Unifying Nuclear Data for AI Searches
When Lauderdale began to work with Diablo Canyon on integrating AI into the nuclear industry, it became apparent that one bottleneck was simply getting decades’ worth of a nuclear plant’s documentation—procedures, best practices, safety recommendations—into an easily searchable format. Information across the industry was scattered and not standardized, which is why three of the biggest nonprofit entities in the industry (the Institute of Nuclear Power Operations , the Electrical Power Research Institute, and the Nuclear Energy Institute) moved in 2025 to create unified datasets, an idea that would birth the NIVA AI assistant developed at Atomic Canyon.
For example, EPRI’s Austin says, nuclear engineers might run into a problem with one of the house-size, thousand-horsepower vertical pumps that circulate vast quantities of water used for cooling purposes. In their search for a solution, the workers have access to plenty of information—too much information, perhaps. They can read through EPRI’s maintenance guides about those components. They can also call upon tens of thousands of incident reports stored with the Institute of Nuclear Power Operations about things that have gone wrong with pumps in the past. AI’s task is to synthesize all those reports to create a guide to the main failures that strike a pump, and, crucially, to show all its work so engineers can find the documentation it used to make its recommendations.
“I want to know all the other nuclear power plants that have that same part,” Lauderdale says. “Have any of them seen a similar issue? How did they fix it? What’s the troubleshooting look like? If you think about nuclear, we have a tremendous amount of information. The challenge is, how do you collectively gather knowledge from that information dataset? And that’s what AI is really, really good at doing.”
“If you think of it as ChatGPT for nuclear, it’s actually not a bad place to start.” —Jerrold Vincent, Nuclearn
Secrecy is a major hurdle to modernizing nuclear operations. AI systems that chiefly help with documentation and data search aren’t subject to the strictest rules, as they would be if they interacted with components that control the power plant. But nuclear data itself is largely proprietary or protected by federal law.
Nuclearn offers only single-location solutions to deal with this, so that a power plant’s data can never leave that site. NIVA can only search databases that already required plant-level security clearances, Austin says, so the same rules hold true for using the AI assistant to hunt through that deluge of data. Technology cannot wipe away these human barriers, but it does provide new incentives to find ways to share data more openly.
“The goal was to make decades of operating experience, hundreds of thousands literally of operating experience reports, tens of thousands of technical documents, and a whole lot of information available for people to find, and do so in a way that’s going to be much more accessible and easy to use, but at the same time preserving data security and that critical information, keeping control over it,” Austin says. That’s exactly what NIVA does.
Nuclearn’s “Equipment AI” system lets engineers pose queries about nuclear-equipment operations and generates answers that draw on decades of industry experience. Nuclearn
AI-Assisted Nuclear-Plant Inspections
So far, both the Nuclearn and NIVA large language models used in the nuclear world report to humans and don’t go anywhere near power plant operations. Still, the inevitable question arises: What if they could?
Vincent and Austin both state that such a feat is far away, but they can envision the pathway there. Consider the litany of welds throughout a nuclear plant, Austin says. All of those welds currently must be inspected for quality and endurance by a human. One pilot project at EPRI is testing the ability to have AI look at the mountains of data generated by such inspections to see whether it can flag areas of concern. A human remains in the loop with that project, Austin says, to review all of the AI’s recommendations, and no power plant is allowing AI to take over its inspections. But the project marks a step toward using AI on direct plant operations rather than only guidance for the human operators.
“I cannot see, in the short or the long term, an environment where AI is making decisions around the operation of the plant, because of the regulatory nature and also because of the risk nature,” Lauderdale says. “You don’t want machines making decisions that could impact the operations. So, I believe we’re going to have humans in the loop for a very, very long time.”
UPower's Truck-Size Nuclear Power Plant ›
The Forgotten History of Small Nuclear Reactors ›
U.S. Plans Largest Nuclear Power Program Since the 1970s ›
Nuclear Power Plants | US EPA ›
Andrew Moseman is the online communications editor at Caltech and a freelance contributor to IEEE Spectrum.