Microsoft Just Took Its Biggest Step Toward Independence From Nvidia

Microsoft announced Azure Maia 300, its newest in-house AI chip, and said Maia will become the primary inference system for its products, with multi-gigawatt deployments starting in 2027. Microsoft claims Maia 300 produces 40% more tokens per megawatt than leading GPU systems, though its 96-chip rack draws about 167 kilowatts versus 142 kilowatts for Nvidia's 72-GPU GB300 rack. The chip is meant to lower the cost of Microsoft's own Copilot and OpenAI services, which it does not sell as chips, while the company expects capital spending above $175 billion this calendar year and will keep leasing Nvidia chips for training. Evercore ISI estimates Maia will account for about 10% of Microsoft's AI chip spending by early 2027. All performance figures come from Microsoft and have not been tested independently, and the next checkpoint is the October 28 earnings call.

Summary written by Upside AI from the original article. Read the source for full details.

Read the original on Yahoo Finance ↗

All news →