Microsoft launches Maia 300 AI chip to boost efficiency
4.4 Tech Microsoft unveiled the Maia 300 AI chip, promising 40% better energy efficiency and up to 7x the throughput of its predecessor. Launching in 2027, it will power key AI services and reduce reliance on Nvidia.
Microsoft has introduced the Maia 300, its third-generation custom AI chip, aiming to reduce dependence on Nvidia and AMD for AI workloads. Scheduled for production rollout in 2027, the Maia 300 is designed to serve as the primary AI inference system for Microsoft’s products and services. The chip is projected to deliver 40% more tokens per megawatt compared to leading GPU-based systems, enabling greater AI output for the same energy cost. A rack with 96 Maia 300 chips could achieve up to seven times the throughput of the previous Maia 200 generation. Targeting demanding inference workloads, such as conversational AI and long-context reasoning, the Maia 300 will power applications like Microsoft 365 Copilot and GitHub Copilot. Microsoft is negotiating large-scale manufacturing with TSMC and plans a global, multi-gigawatt deployment in Azure data centers, reflecting the industry’s shift toward in-house AI hardware to manage rising infrastructure costs and growing computing demands.
