Ryan Smith

Google unveiled Ironwood TPU, its first chip explicitly designed for large-scale AI inference, offering 2x perf-per-watt improvement over its previous generation Trillium and scaling up to 9,216 chips with 42.5 Exaflops performance.

“Author of "Google Ironwood TPU Swings for Reasoning Model Leadership at Hot Chips 2025" in ServeTheHome”