UT Austin wins Google.org award for AI-assisted stellarator design research

A team led by Joshua Burby at the University of Texas at Austin was selected for the Google.org Impact Challenge: AI for Science. Announced on 15 October, the three-year project will develop neural-network surrogate models to accelerate design exploration for compact stellarator fusion reactors. The models are intended to approximate repeated calculations in high-fidelity simulations of heat and particle transport inside a reactor.

The team is targeting high-beta configurations, where plasma pressure is a relatively large share of magnetic pressure and the plasma itself affects the magnetic field configuration used to confine it. The goal is to search a wider design space while retaining physical consistency. The announcement describes research aims; it does not report a reactor prototype or a verified reduction in construction costs.

The team includes researchers from UT Austin, the University of Washington and New York University. UT Austin is one of 17 organizations selected globally from more than 3,000 applicants across 30 countries to share $36 million. The release does not state this project’s individual funding amount. Recipients also receive Google Cloud credits and technical support through a six-month accelerator.