Microsoft has unveiled Maia 300, its third-gen custom AI accelerator developed to run as the primary inference engine across Azure and first-party enterprise services. Built specifically to handle large-scale reasoning models and autonomous agent workflows, the hardware represents a major expansion of the company's internal hardware programme.By deploying purpose-built inference silicon, Microsoft seeks to curtail infrastructure expenses and reduce its heavy operational dependence on Nvidia GPUs amid sustained datacenter capacity demands.The processor will begin handling production inference tasks immediately across cloud regions, providing dedicated computing power tailored to complex multi-step reasoning workloads.Will in-house hyperscaler processors like Maia 300 successfully displace commercial GPU dominance for cloud AI inference?#Microsoft #AI #Silicon #Technology
Microsoft has unveiled Maia 300, its third-gen custom AI accelerator developed to run as the primary inference engine ac...
