Keep up with the innovative tech transforming business
Tech Brew keeps business leaders up-to-date on the latest innovations, automation advances, policy shifts, and more, so they can make informed decisions about tech.
Generative AI is coming to nuclear power plants—but it’s perhaps less scary than it sounds.
California’s last remaining nuclear power plant, the Diablo Canyon Power Plant, recently teamed with AI startup Atomic Canyon to handle office tasks like document search and retrieval. The goal is to ease management of the extensive regulatory documentation operator Pacific Gas & Electric is required to log on the facility.
The startup claims it’s the first time generative AI has been used onsite at a US nuclear power plant.
But use cases like these are nothing new for large language models—delegating certain rote paperwork and organizing internal information are among the most intuitive ways for all kinds of offices to put this type of AI to work. Atomic Canyon has eyed a niche in the nuclear power industry, which requires reams of datalogs and operational notes to stay compliant with safety rules, as well as plenty of specialized knowledge.
Nuclear energy also happens to be booming right now, thanks in big part to AI itself. Faced with a power crunch brought on by the massive data centers needed to train and run generative AI models, tech giants like Microsoft, Google, and Amazon have turned to an energy source that conveniently seems to be enjoying newfound public acceptance after decades of safety stigma.
The Diablo Canyon deal marks the first commercial deployment of Atomic Canyon’s Neutron Enterprise offering, which is designed specifically for the nuclear power industry and runs on Nvidia’s AI platform.
Neutron taps a family of open-source models called FERMI, specially developed by Atomic Canyon in collaboration with the Department of Energy’s Oak Ridge National Laboratory. The models are trained on millions of publicly available licensing and technical documents from the US Nuclear Regulatory Commission’s Agency-wide Documents Access and Management System (ADAMS).
The platform uses a technique called retrieval-augmented generation (RAG), which requires document citations to improve accuracy and cut down on potential hallucinations.
Atomic Canyon is hoping to ride the growing popularity of nuclear energy into more deals like these in the future.
“With skyrocketing energy demands and increased support from tech leaders, we are witnessing the growing excitement and need for nuclear energy in real time,” Trey Lauderdale, founder and CEO of Atomic Canyon, said in a press release. “This is the future of nuclear plant operations, and we’re just scratching the surface.”