Microsoft’s new AI chip makes it less reliant on Nvidia’s GPUs
The Verge AI · 12 Oct 2026, 23:00 AEST
Microsoft has announced its third-generation Maia 300 in-house AI chip, aimed at reducing reliance on Nvidia's GPUs. The chip is designed to improve the cost and efficiency of AI workloads in the cloud and will begin production in 2027.
Key facts
- Microsoft is announcing its third-generation Maia in-house AI chip today, designed to be the 'primary AI inference system' for Microsoft's products and services.
- Maia 300 not only improves the cost and efficiency of running AI workloads in the cloud, it's also more capable of handling the latest reasoning models and agentic tasks.
- We expect Maia 300 to deliver 40 percent more tokens per megawatt compared to leading GPU-based systems.
Read aloud
Microsoft has announced its third-generation Maia three hundred in-house A I chip, aimed at reducing reliance on Nvidia's G P U s. The chip is designed to improve the cost and efficiency of A I workloads in the cloud and will begin production in twenty twenty seven. Microsoft is announcing its third-generation Maia in-house A I chip today, designed to be the 'primary A I inference system' for Microsoft's products and services. Maia three hundred not only improves the cost and efficiency of running A I workloads in the cloud, it's also more capable of handling the latest reasoning models and agentic tasks. We expect Maia three hundred to deliver forty percent more tokens per megawatt compared to leading G P U-based systems.
Sources in this cluster
- Microsoft’s new AI chip makes it less reliant on Nvidia’s GPUs
The Verge AI · 12 Oct 2026, 23:00 AEST