Generative AI Revives Diablo Canyon's Nuclear Potential

When PG&. E considered shutting down the Diablo Canyon Power Plant by 2025, few expected a technological intervention to extend its life. Yet, in a surprising twist, California Governor Gavin Newsom signed legislation to keep the plant running through 2030. As regulatory filings piled up, it became clear: the bottleneck wasn't the plant itself, but the task of sorting through mountains of paperwork.

Enter Neutron, a generative AI tool developed by the startup Atomic Canyon. Inspired by the burdensome documentation in healthcare, Trey Lauderdale, Atomic Canyon's CEO, turned his gaze to nuclear power. His solution? An AI that could sort through the billions of data points at Diablo Canyon, cutting through the chaos with precision.

AI Meets Nuclear

Launched in 2024, Neutron is the brainchild of a collaboration between PG&. E and Atomic Canyon. It connects directly to the plant's systems, bypassing the cloud to provide a secure, web-based search tool for its workers. The AI was trained on 53 million pages of publicly available data from the Nuclear Regulatory Commission, a vast trove that dwarfs typical machine learning datasets.

In a sector where regulatory compliance is king, Neutron's ability to speed up document retrieval is a breakthrough. PG&. E's senior vice president, Maureen Zawalick, is optimistic. She's seen firsthand how AI can achieve in days what once took months.

Operational Efficiency

Previously, workers needed to sift through multiple systems to find regulatory documents. Neutron consolidates this data, slashing search times and reducing the workforce needed for such tasks. The outcome? For projects like investigating safety valve issues, what took 180 days now takes just 40.

Why should this matter to the broader energy sector? Simple. As the US grapples with growing energy demands, especially from data centers and new tech industries, tools like Neutron might be the key to efficient regulatory filing for new power projects.

The Bigger Picture

But here's the question: If AI can speed up operations at a nuclear facility, what else can it optimize? Energy efficiency isn't just about cleaner sources, but better management of existing resources. The economics of nuclear power could shift dramatically with AI-driven efficiencies.

For PG&. E, Neutron is more than just a tech upgrade, it's a lifeline. As California's last operational nuclear plant, Diablo Canyon stands as a testament to how generative AI can redefine industry standards. The unit economics break down at scale, but with the right AI tools, the infrastructure bottleneck might just become a thing of the past.