According to CNBC, Microsoft will increase use of its in-house artificial intelligence chips after years of relying on Nvidia graphics processing units, and said its third-generation Maia AI accelerator will mainly handle incoming requests. The company said production of the Maia 300 starts in 2027, expects to fit 96 Maia 300 chips in each data center rack, and said the new version can process up to seven times more AI tokens per rack and deliver 40% more tokens per megawatt than GPUs. Microsoft said parts of Microsoft 365 Copilot should speed up with Maia 300, while Nvidia will remain important for training models and for some other Microsoft products. Evercore ISI analysts estimated in a Wednesday note that Maia will account for about 10% of Microsoft's AI chip spending by the first quarter of 2027.