Microsoft Launches Maia 300 AI Chip to Power Cloud Services and Reduce NVIDIA Dependence

Microsoft announced on October 12, 2026 the arrival of Maia 300, its third-generation custom AI accelerator designed to serve as the primary inference engine for its cloud services and reduce its reliance on external semiconductor suppliers for artificial intelligence workloads. Built on Taiwan Semiconductor Manufacturing Company's cutting-edge N2P process, the chip represents a major escalation in Microsoft's vertical integration strategy, following the earlier Maia 200 that launched in January 2026.