Join Microsoft as we share more on Maia 200 in the Bay Area

Microsoft has introduced Maia 200, a new inference accelerator designed to improve the economics of AI token generation. This breakthrough silicon innovation will be showcased by Microsoft engineering leaders this month. A technical whitepaper detailing Maia 200, titled “Maia: A Reticle-Scale AI Accelerator,” will be released on February 13th to ISSCC attendees and later digitally on IEEE. At the ISSCC conference, Azure Hardware and Systems leadership will present a session on Maia development. This presentation will delve into the architecture and implementation of the Maia AI silicon. It will also cover the engineering of a reticle-limited, ~750W AI SoC. Innovations enabling a scalable, high-performance accelerator will be shared. Microsoft is hosting a Silicon Social event in San Francisco on the evening of February 17th. Maia 200 will make its first public appearance there, alongside other Microsoft silicon hardware. Microsoft's silicon engineering leadership will be present at this event.