Building 92 at Microsoft's headquarters in Redmond, Washington.

Microsoft unveils Maia 300 chip as it scales back reliance on GPUs

The third-generation inference chip begins production deployment in 2027.

Building 92 at Microsoft's headquarters in Redmond, Washington.

By West Coast News staff

Oct 12, 2026, 9:00 am PT · 1 min read

Microsoft says it will scale its Azure Maia chips to multi-gigawatt deployments from 2027, making them the main inference system for its products, according to a company blog post by cloud chief Scott Guthrie.

Maia 300 is built on TSMC's N2P process with eight HBM4 memory stacks. Microsoft says it will deliver 40 per cent more tokens per megawatt than leading GPU systems, and up to seven times the rack throughput of Maia 200.

Partner links From West Coast News's sister sites and stores

ARQS browser gamesOriginal browser games to play at home.
the ARQS Press roomReleases, editorials and new books from ARQS Press.
InvoiceFlow Billing Control StudioInvoice-data quality analysis with an ARQS-reviewed findings summary.
AI Academy Year 1Design controlled AI-assisted workflows for real operations.
ARQS Flow ConnectA workspace for integrations, approvals and governed workflows. 30-day starter access.
LinguaQuest language softwareBrowser language-learning software with adaptive review and speech support.
the ARQS Store's business softwareAI receptionist, business automation, readiness reviews, website audits and launch packages.
AI Academy's practical AI coursesSelf-study courses in practical AI workflows from Northline Technology Institute.