Meta has announced a partnership with AWS to deploy tens of millions of AWS Graviton5 cores, making it one of the largest Graviton customers globally. This move is part of Meta's broader multi-architecture compute strategy that spans Nvidia, AMD, Arm, and its own MTIA silicon. Analysts note the shift reflects the demands of agentic AI workloads, where CPUs serve as the control plane for orchestration, memory management, and scheduling alongside GPU accelerators. The strategy is about vertical integration and workload-aware infrastructure rather than competing with hyperscalers, with efficiency and total cost of ownership becoming increasingly important as inference workloads grow persistent.
Sort: