Microsoft unveils Maia 300 AI chip at OCP Global Summit
Microsoft just introduced its Maia 300 AI chip at the OCP Global Summit, aiming to make AI tasks far more efficient.
Microsoft expects Maia 300 to deliver 40% more tokens per megawatt than leading GPUs, helping Microsoft build a self-reliant AI setup and cut back on third-party chips.
Production is set for 2027.
Maia 300 to support Microsoft AI
With Maia 300, Microsoft plans to support a range of Microsoft's AI services, including coding tools, cloud-based model deployment through Microsoft Foundry, model development by the MAI team, and clinical workflow applications through Dragon Copilot:
a rack containing 96 accelerators is also expected to deliver up to seven times the rack-scale token throughput compared to the older Maia 200 (already running in places like the US Europe, and Australia).
This move keeps Microsoft in step with tech giants like Amazon and Google, who are also rolling out their own custom chips for smarter, faster services.