01 What happened
Microsoft announced Maia 300, the next-generation AI inference platform, and said plans call for Azure MaIA to expand to multi-gigawatt deployments beginning in 2027. Microsoft said MaIA 300 will become the primary AI inference system for its products and services. At rack scale inside a Microsoft datacenter, MaIA 300 systems are expected to deliver up to seven times the token throughput of MaIA 200. The chip is designed around real-world Microsoft workloads and new AI workloads, including long-context reasoning, latency-sensitive conversational experiences and sustained, multi-step agentic tasks. Microsoft said MaIA 300 is expected to deliver 40% more tokens per megawatt. These are company-stated expectations, not independently verified benchmarks.
02 Key details
- Microsoft announced Maia 300 as its next-generation AI inference system.
- From 2027, Azure MaIA aims to expand to multi-gigawatt deployments.
- At rack scale in a Microsoft datacenter, Maia 300 is expected to deliver up to 7x token throughput versus Maia 200.
- Maia 300 is expected to deliver 40% more tokens per megawatt.
- Maia 300 is designed for long-context reasoning, latency-sensitive conversations and sustained, multi-step agentic tasks.
03 Why it matters
Maia 300 targets higher token throughput per watt and more efficient cloud inference for Microsoft services. The rollout matters for datacenter operators, cloud customers and Microsoft product teams working with Copilot, Teams, GitHub and Foundry. The 40% figure and 7x throughput claim come from Microsoft and should not be treated as independently verified measurements.
04 Who it matters to
Infrastructure engineers, corporate AI platform developers, AI agent developers, technical leads and operations directors.