Microsoft has unveiled its third-generation AI accelerator, the Maia 300. According to the company, the new chip delivers 40% higher performance than leading Nvidia GPU-based systems while maintaining the same power consumption levels. A full rack of Maia 300 units provides up to seven times the performance of the previous generation for tasks such as AI agent operations and processing long documents.
The chip is designed in-house by Microsoft and manufactured by TSMC. Mass deployment is scheduled for 2027. Unlike competitors such as Google and Amazon, Microsoft does not intend to rent these accelerators to third-party clients. Analysts at Evercore ISI estimate that Maia chips will account for approximately 10% of Microsoft's total AI chip spending by early 2027.
The current Maia 200 generation is already supporting Microsoft 365 Copilot and OpenAI models, with ongoing rollouts in the US, Europe, and Australia. GitHub Copilot and Microsoft Foundry are slated to transition to this hardware by the end of the year. The initial Maia chip was announced in 2023, while the second version launched in January 2026.
Despite the development of its own silicon, Microsoft continues its partnership with Nvidia. The company also plans to be among the first cloud providers to deploy systems based on Nvidia Vera Rubin and AMD Helios. Further details regarding the company's infrastructure strategy are expected to be revealed in November at the Ignite conference in San Francisco.
