AI NewsWaveMicrosoft Unveils Third-Generation AI Chip 'Maia 300,' Plans 1 Million or More Next Year
Key points
- Microsoft unveils its own AI chip, Maia 300
- Maia deployments to expand to multiple gigawatts from 2027
- A move to reduce reliance on Nvidia GPUs
Microsoft unveiled its third-generation in-house AI chip, the Azure Maia 300, on the 12th (local time). According to Platum, from 2027 it plans to expand deployment of the Maia lineup to a multi-gigawatt scale and use Maia as the main AI inference system for its own products and services.
AI Newspaper (인공지능신문) reported that production deployment of the Maia 300 will begin in 2027, following the Maia 200 now used in real services such as Copilot.
On the scale of deployment, Industry News put deployment of 1 million or more next year in its headline. iNews24, citing Reuters, reported that Microsoft could deploy 1 million or more Maia 300 chips starting next year. IT Chosun wrote in its headline of deployment of up to 1 million or more next year.
Explanations of the background also overlap. Chosun Biz reported that Microsoft, which has relied on Nvidia GPUs until now, is declaring a shift of its center of gravity to its own chips. IT Chosun explained that it is a strategy to reduce dependence on Nvidia GPUs and improve the operating cost and power efficiency of AI services.