Microsoft & AMD Double Down on Custom AI Silicon
Microsoft CEO Satya Nadella announced Azure Maia will scale to multi-gigawatt deployments starting next year as the primary inference system for all MS products, alongside the new Maia 300 chip. AMD unveiled Helion, an open AI-agent server rack headed to the Open Compute Project.
What we know
Satya Nadella announced Azure Maia will scale to multi-gigawatt deployments next calendar year, replacing third-party silicon as the primary inference backbone for Microsoft's own products and services. The new Maia 300 delivers approximately 40% performance improvement. Separately, AMD unveiled Helion, an open server rack for AI agents headed to the Open Compute Project, with configurations supporting up to 32,768 EPYC cores or EPYC + Instinct MI350P GPUs. This comes against a backdrop of Broadcom CEO Hock Tan warning about an AI capital shortage and global smartphone shipments falling 20% in Q3 as AI data centers absorb memory chip supply.
How the story developed
Microsoft CEO Satya Nadella announced Azure Maia will scale to multi-gigawatt deployments starting next year as the primary inference system for all MS products, alongside the new Maia 300 chip. AMD unveiled Helion, an open AI-agent server rack headed to the Open Compute Project.
Microsoft announced it will scale Azure Maia to multi-gigawatt deployments in 2027, with its new Maia 300 chip delivering 40% improvements. AMD also launched the Helion Agentic AI Rack at rack scale, marking a major AI infrastructure push from both companies.
By the Global Eye 24 desk, with AI assistance. Updated as the story develops. Spot an error? Write to corrections@globaleye24.com.