Microsoft bets on homegrown AI chip to challenge Nvidia

Microsoft plans to begin a massive rollout of its in-house Maia artificial intelligence chips in 2027 to serve as the primary hardware for running AI across its products and services. According to Punch and Punch Nigeria, the move aims to reduce the company's reliance on Nvidia. Quartz reports that Microsoft intends to rely on its own silicon to handle incoming AI model requests starting with the third-generation Maia 300.

The Maia 200 chip is already operational and is used for the Microsoft 365 Copilot assistant, according to Straits Times Business. The upcoming Maia 300 is designed to improve the cost and efficiency of cloud AI workloads while better handling reasoning models and agentic tasks. The Verge reports that Microsoft cloud group executive vice president Scott Guthrie expects the Maia 300 to deliver 40 percent more tokens per megawatt than leading GPU-based systems.

Straits Times Business notes that Microsoft plans to deploy enough chips to draw several gigawatts of electricity. While Nvidia has maintained a near-lock on chips used to train AI models, the market for inference is expanding as companies deploy AI assistants and agents.

Brief written by urgent.news from Punch, Punch Nigeria, Straits Times Business, Quartz, The Verge — 5 reports on this story. Machine-written — may contain errors; check the original before relying on it.