Great Sky’s High-Speed AI Chips Analyze Fusion Plasmas
aiinsider ·
Great Sky, in partnership with Brookhaven National Lab and Princeton Plasma Physics Laboratory, has demonstrated a superconducting AI chip capable of analyzing fusion plasmas at high speeds. Unlike conventional silicon-based processors, this chip utilizes analog neural networks and optical communication to process data directly from tokamaks. The team successfully processed DIII-D tokamak data at one million samples per second for anomaly classification and disruption prediction tasks. Additionally, the chip analyzed simulated data from Google DeepMind’s TORAX simulator to predict sawtooth instabilities, which are critical for reactor stabilization. This demonstration highlights the potential for real-time plasma control systems necessary for practical fusion power generation. Great Sky aims to address energy efficiency and speed challenges in both AI hardware and fusion technology through this innovation. The company plans to showcase a next-generation chip with a neural network ten times larger in December.
aiinsider