Brief
Microsoft to Deploy Over a Million Maia 300 AI Chips Next Year
The new 2‑nm chiplet‑based processor promises 40% more work per megawatt than leading GPU systems.
By Felo News Desk · Published
Microsoft unveiled its third‑generation custom AI processor, the Maia 300, on Monday in San Francisco and said it intends to ship over a million units in 2027 to power its internal AI services.
The chip uses 2‑nanometer silicon for core compute chiplets and 3‑nanometer chiplets for data movement, sourced from Taiwan Semiconductor Manufacturing Co. Each rack holds 96 chips and consumes about 167 kilowatts, versus 142 kilowatts for Nvidia's latest GB300 GPU racks. Microsoft claims the Maia 300 delivers roughly 40% more useful computing per megawatt than leading GPU‑based systems.
Rani Borkar, president of Azure hardware systems and infrastructure, told Reuters the chip "delivers more intelligence for every watt, every dollar, and every byte of memory by optimizing across the entire system." Microsoft also said it expects to spend more than $175 billion in capital expenditures this calendar year and will begin "multi‑gigawatt deployments" of the Maia 300 next year.
Key facts
- Microsoft announced the Maia 300 AI chip on 13 Oct 2026. (deccanchronicle.com)
- The company plans to deploy over a million Maia 300 chips in 2027. (deccanchronicle.com)
- Maia 300 uses 2‑nm chiplets for core compute and 3‑nm chiplets for data movement. (deccanchronicle.com)
- A rack of 96 Maia 300 chips consumes about 167 kW, 25 kW more than Nvidia's GB300 GPU rack. (deccanchronicle.com)
- Microsoft claims a 40% efficiency gain per megawatt over leading GPU systems. (deccanchronicle.com)
Sources
- [1] deccanchronicle.com — originally reported as “Microsoft Plans Huge Deployment Of Its Own AI Chips, With An Eye On Saving Costs”
