[Microsoft Launches Third-Generation Maia 300 AI Chip, Aims to Reduce Dependency on NVIDIA] According to Deep Tide TechFlow, on October 12, CNBC reported that Microsoft officially unveiled its third-generation self-developed AI accelerator chip, Maia 300, on October 12. The company announced that the Maia series will serve as the primary architecture for running AI models, gradually reducing its reliance on NVIDIA GPUs. The Maia 300 is planned for production in 2027, with each data center rack capable of housing 96 chips (double the previous generation), a 7x increase in AI token processing per rack, and a 40% improvement in token processing efficiency per megawatt compared to GPUs. Microsoft aims to deploy multi-gigawatt-level Maia chip clusters, with Microsoft 365 Copilot being the first to benefit from the performance enhancements. Evercore ISI analysts estimate that by Q1 2027, Maia chips will account for approximately 10% of Microsoft's AI chip expenditures. It is worth noting that the Maia accelerators are currently for internal use only and are not available for external rental. NVIDIA and AMD will continue to participate in Microsoft's model training operations.