Microsoft plans to transition the majority of its AI workloads from Nvidia and AMD GPUs to its own Maia AI accelerators, aiming to improve performance per dollar in its datacenters. Despite being late to the custom silicon market, Microsoft CTO Kevin Scott expressed confidence in the company's ability to optimize compute with their homegrown chips, although complete replacement of existing GPUs is unlikely. A second-generation Maia accelerator is expected next year to enhance competitive performance.