Illustrative photograph by İsmail Enes Ayhan on Unsplash .

On 12 October 2026, Microsoft announced Azure Maia 300, the third generation of its in-house AI accelerator, and said it will scale the Maia family to multi-gigawatt deployments starting in 2027, making Maia "the primary inference system for Microsoft products and services" ( Microsoft, Official Blog, 12 Oct 2026 ).

The headline numbers travel well: 40% more tokens per megawatt than "leading GPU-based systems," and up to 7x the rack-scale token throughput of the previous Maia 200 in a full 96-accelerator rack. Both figures are Microsoft's own, and both are forward-looking. What makes the announcement more interesting than a chip spec sheet is the unit Microsoft chose to compete in. Not tokens per second. Tokens per megawatt.

What Microsoft actually said

Guthrie's post is specific about the design targets.…