AMD Pushes Rack-Scale AI for Agentic Workloads to Boost CPU Demand
AMD expects enterprises to run hybrid architectures combining cloud frontier models with on-prem open-weight systems. The company reiterated targets of over 60% data-center CAGR and over 80% AI growth over its strategic timeframe, citing a ~$2 trillion TAM by 2030. It plans to scale AI through rack deployments such as Helios systems and MI455, emphasizing agentic workloads that alter inference architecture and affect CPU needs.