Azure Maia

Azure Maia is Microsoft's in-house family of AI inference chips and systems. According to Microsoft's official blog, the multi-generational Azure Maia project has become a foundational part of Microsoft's AI infrastructure strategy: Maia 200 is in production, being deployed in the US, Europe, and Australia, and serves inference for the built-in AI capabilities of Microsoft Teams, Copilot, and Cowork, delivering up to 1.7 times token output per watt compared with leading GPU infrastructure, according to the official announcement. The next-generation Maia 300 is planned for 2027 and targets emerging AI workloads such as long-context inference, latency-sensitive conversational experiences, and multi-step agent tasks; Microsoft expects it to deliver 40% more tokens per megawatt than leading GPU-based systems and, in its full 96-accelerator configuration, up to 7 times the rack-level token throughput of Maia 200. Microsoft has also announced that from 2027 it will scale the Maia family to multi-gigawatt deployments, making it the primary inference system for Microsoft's products and services. Note that this is an infrastructure-level announcement: Azure Maia is not a standalone tool intended for direct use by individual users.