Microsoft plans to make its in-house Maia chips the go-to for running AI across its products and services starting in 2027, CEO Satya Nadella said on Monday.
"Today we are announcing our biggest bet yet on our own silicon," Nadella wrote on X.
The company plans to expand the existing Maia chip line to "multi-gigawatt deployments," according to a Monday release. It did not give an exact target for capacity or when the rollout could be completed.
As part of the announcement, Microsoft also introduced Maia 300, the third generation of its AI chips, slated to be deployed in 2027. The company said it expects Maia 300s to process 40% more AI tokens than other leading GPU systems.
Maia 300 is an "inference-first" chip, meaning it is best suited for heavy AI training and workloads. The company's Maia 200 chips currently run Microsoft Copilot workloads across the U.S., Europe, and Australia.
The expansion could ultimately reduce Microsoft's reliance on Nvidia chips for running its AI services. For now, however, Microsoft remains an Nvidia customer and partner. In July, it announced an expanded partnership with Paris-based Mistral that included thousands of Nvidia's Vera Rubin GPUs to be deployed in Europe to support its AI services.
Nvidia shares (NVDA) fell roughly 1.7% during the first hours of trading on Monday, while Microsoft (MSFT) rose around 1%.