Microsoft Plans Maia 300 AI Chip Rollout in 2027

Microsoft announced the Maia 300, its third-generation custom AI chip, on October 12, 2026, saying it plans to begin deploying the processor in 2027 as the primary system for running artificial intelligence in its products and services. The chip is designed for inference, the computing work involved in running trained AI models. Microsoft said it is intended to handle long conversations, reasoning and multistep tasks performed by AI agents. The company plans deployments measured in multiple gigawatts starting next year. One report said that could mean more than one million chips in 2027, though that scale remains a projection. Microsoft expects Maia 300 to produce 40 more tokens per megawatt than leading GPU-based systems. Tokens are units of text processed or generated by AI models. The company did not identify the systems behind the comparison, and the figure is a Microsoft projection rather than an independently verified benchmark. A Maia 300 rack will contain 96 accelerators and draw about 167 kilowatts, according to Microsoft. The company says a rack can deliver up to seven times the token throughput of a Maia 200 rack.