Microsoft plans to scale its own Maia AI chips to gigawatt levels by 2027
Microsoft is betting its silicon roadmap on inference at scale. By 2027, Maia 300 chips will run most Azure workloads—40% more efficient than GPU alternatives.
Why it matters
Microsoft is shifting its AI compute strategy from GPU-centric to custom silicon for inference workloads. Maia 300's claimed 40% token-per-megawatt advantage over leading GPU systems signals a strategic move to reduce inference costs and Nvidia dependency at multi-gigawatt scale by 2027—a material change for enterprises running large-scale deployments on Azure.
The key facts
5 to knowMicrosoft Maia 300 claims 40% more tokens per megawatt vs leading GPU systems
Deployment target: multi-gigawatt scale by 2027
Maia chips become primary inference system across Microsoft products starting 2027
Satya Nadella: 'our biggest bet yet on our own silicon'
Focus: inference workloads, not training
Go to the source
The Decoderthe-decoder.com
Publisher excerpt: Microsoft will make its in-house Azure Maia chips the main AI inference system across its products starting in 2027, deploying them at multi-gigawatt scale. The upcoming Maia 300 should deliver 40 percent more tokens per megawatt than leading GPU systems. CEO Satya Nadella calls it "our biggest bet…