AMD targets system-level AI infrastructure optimization as agentic workloads reshape enterprise compute

Agentic AI workloads are pushing AMD to optimize infrastructure at the system level, balancing performance and cost across diverse compute engines. CTO Mark Papermaster said the company expanded through acquisitions of Xilinx, Pensando, and ZT Systems, evolving into a rack-level optimizer. AMD's ROCm software stack runs identically across data centers, edge, and AI PCs.