Microsoft bets big on in-house Maia silicon: multi-gigawatt rollout, 40% better tokens/watt
mustafasuleyman · x · 2026-10-12
Microsoft announced its biggest bet yet on in-house silicon: starting next year, the Maia family scales to multi-gigawatt deployments as the primary inference system for its own services. New Maia 300 delivers 40% more tokens per watt than leading GPU-based systems, while Maia 200 is already deployed across the US, Europe, and Australia with up to 1.7x tokens-per-watt gains in Teams and Copilot.
Related event: Microsoft Unveils Maia 300, Betting Big on In-House AI Silicon(10 posts)→
More from Companies & People
- Hone, founded by ex-Cognition and Mercor staff, builds business-running agents — nmasc_ · 2026-10-13
- Google DeepMind's Philipp Schmid to speak on building an autonomous coworker at AI Engineer NYC — _philschmid · 2026-10-13
- The Ankler's Future of Entertainment 25: Runway CEO Says AI Is Eating Hollywood — DavidmComfort · 2026-10-13
- Tim Cook Shows Up at Zhejiang University Apple App Incubator Demo Day — op7418 · 2026-10-12
- OpenAI drops 700+ math manuscripts; human mathematicians group urges boycott — thione · 2026-10-12
- OpenAI's three fired safety researchers dispute misconduct claims, warn of chilled safety work — thione · 2026-10-12