AI · SEMICONDUCTOR · FACT CHECK
SKD WIREPartly True
Microsoft has disclosed its self-developed AI chip, the Maia 300, and plans to deploy more than 1 million units of the accelerator within the next year, according to material published on the company's official blog.
The Maia 300 announcement was made through Microsoft's official channels, with the company publishing product specifications alongside images of the chip and a prototype version. The announcement marks the latest step in Microsoft's effort to design its own AI accelerators rather than relying solely on third-party GPUs for its data center infrastructure.
The chip is positioned as an in-house alternative for AI workloads, a segment where demand for compute capacity has grown rapidly alongside the expansion of large language model (LLM) services.
Microsoft stated that it intends to place more than 1 million Maia 300 units into operation next year. No additional figures on pricing, manufacturing partners, or production capacity were included in the published material.
Details on the chip's full specifications are available in the specification materials released through Microsoft's blog, though independent performance comparisons with competing accelerators have not yet been made public.
Large cloud operators have increasingly moved toward custom AI silicon to control costs and secure supply amid tight accelerator availability. Microsoft's Maia line is part of that broader shift, and the Maia 300 represents its newest generation.
The claim that Microsoft revealed the Maia 300 and will deploy over 1 million units next year is Partly True.
