Microsoft Launches Maia 200 AI Chip but Will Keep Buying Nvidia, AMD Hardware: Nadella
Microsoft begins deploying its in-house Maia 200 AI inference chip in data centers as of January 29, 2026, touting performance exceeding Amazon’s Trainium and Google’s TPUs — yet CEO Satya Nadella confirms continued purchases from Nvidia and AMD. The Maia 200 will initially power Microsoft’s internal “Superintelligence” team, led by ex-DeepMind co-founder Mustafa Suleyman, to accelerate development of proprietary frontier AI models. It will also support OpenAI workloads on Azure. Despite vertical integration capabilities, Nadella emphasized sustained partnerships: “Because we can vertically integrate doesn’t mean we only vertically integrate.” Supply constraints for top-tier AI chips persist industry-wide, making multi-vendor sourcing essential even as Microsoft reduces long-term reliance on third-party model providers like OpenAI and Anthropic.