Deep dive into Azure Maia 300 architecture

Microsoft has unveiled the Azure Maia 300, a third-generation AI acceleration architecture that integrates software, silicon, rack-scale infrastructure, networking, and datacenters to optimise large language model inference efficiency in terms of tokens per watt and cost. Key innovations include a multi-chiplet architecture with heterogeneous dies, advanced memory organisation shifting to non-uniform memory access, enhanced low-precision data formats, and expanded scale-up capability to 12,288 accelerators in a superpod.

This Wire brief sits within Fusion42's coverage of AI & ML, and 6 sources have reported it between 30 Sep 2026 and 12 Oct 2026.