Microsoft announced plans to deploy its Azure Maia AI chips at multi-gigawatt scale starting in 2027, marking its largest silicon bet yet. The company's executive Scott Guthrie stated that Maia chips will become the main inference system across its products.
Maia 200 already runs Microsoft Copilot in the US, Europe, and Australia, with GitHub Copilot and Foundry following by year's end.
Microsoft reported that Maia 200 pushes up to 1.7 times more tokens per watt than leading GPUs in Teams. Each Maia 300 chip packs eight HBM4 memory stacks totaling 288 GB with 18 TB/s of bandwidth and is rated at 950 watts. Its successor, Maia 300, ships in 2027 and targets long inputs, fast chats, and agentic tasks.
"This is our biggest bet yet on our own silicon," said CEO Satya Nadella. The chip handles OpenAI, MAI, and open-source models, and Microsoft claims 40 percent more tokens per megawatt than leading GPU systems. A 96-chip rack should produce up to seven times the tokens of a Maia 200 rack.
The announcement follows the release of the technical report covering the details, with more coming in November at Ignite 2026. Microsoft did not say when the full scale of the Maia 300 deployment will be revealed, and the company emphasized the need for further development before finalizing the rollout plan.
Source: thedecoder
