AWS and national lab team up to deploy AI tools in pursuit of fusion energy

Amazon Web Services is teaming up with Lawrence Livermore National Laboratory — home to the world’s only facility to achieve fusion ignition — to develop artificial intelligence tools to advance the lab’s efforts, the two announced today.

AWS and LLNL’s National Ignition Facility are working together to build an AI-driven troubleshooting and reliability system, and have already deployed generative AI capabilities into the fusion lab’s operations. The focus is on using AI to produce real-time solutions to anomalies that arise in the research and addressing increasing operational demands.

More than two years ago, NIF reported that it had produced more energy from a fusion reaction than went into it, an accomplishment known as ignition. Since then, the facility has hit that mark seven additional times, most recently in April when it nearly tripled the amount of energy produced in December 2022.

Researchers internationally are trying to recreate the fusion reactions that power the Sun — developing “star in a jar” technologies that will allow humanity to produce nearly limitless clean energy on Earth. That power is increasingly in demand as data centers continue expanding and other sectors of the economy are electrifying their operations.

In the new partnership with the federal lab, AWS’s AI could help solve the very energy consumption problems it is helping to create.

“I’m excited to unleash the superpower that is AI on NIF operations,” said Kim Budil, director of LLNL, in a statement. “By leveraging our extensive historical data through advanced AI techniques, we’re solving today’s problems faster and paving the way for predictive maintenance and even more efficient operations in the future.”

Last week, Washington state companies Helion Energy, Zap Energy and Avalanche Energy participated in a Seattle-area summit to share their progress in working towards commercialized fusion. In the past they celebrated NIF’s experiments as a validation that their ambitions are possible. No other facility anywhere has demonstrated fusion ignition, and NIF’s objective is strictly research, as opposed to building reactors to put power on the grid.

One of the interesting applications being pursued at NIF is unleashing AI on more than 98,000 archived problem logs stretching back 22 years. The documents are a trove of lessons learned, including symptoms, causes and the steps taken to fix the problems.

A release from the California-based national lab said the partnership could “establish a new standard for AI application in high-stakes scientific facilities and may influence operational approaches at other national laboratories.”

David Appel, vice president of U.S. Federal Sales at AWS, called LLNL “an innovation and scientific powerhouse, and we’re extraordinarily proud of our partnership together.”