Microsoft will scale its in-house Azure Maia AI chip family to multi-gigawatt deployments from 2027, aiming to make Maia the primary platform for AI across products and services. Maia 300 promises higher energy efficiency and throughput versus GPUs and Maia 200, targeting demanding long-context and agent workloads with a 96-accelerator, 167-kilowatt system built on TSMC N2P process.
