The third-generation in-house chip is designed to be Microsoft's primary AI inference engine, improving cost and efficiency across its products.
Microsoft unveiled Maia 300, its third-generation custom AI chip, built to serve as the primary inference system powering Microsoft's own products and services. The company says the chip improves both cost and operational efficiency versus prior generations.
The move continues a broader hyperscaler pattern of building proprietary silicon to reduce reliance on Nvidia GPUs, following similar bets from Google and Amazon. Microsoft frames Maia 300 as infrastructure strategy, not just a hardware refresh.
Every point of margin Microsoft claws back from GPU costs is margin it can redeploy into pricing Azure AI services more aggressively, or into pure profit. For enterprise buyers, custom silicon from cloud vendors means inference costs tied to one vendor's roadmap, a leverage shift worth watching in contract negotiations.
The daily signal, curated. Get it in your inbox.
Subscribe on LinkedIn →