Meta has entered a multiyear, multi-billion-dollar agreement with AWS to deploy tens of millions of Graviton5 CPU cores, positioning itself as one of the largest Graviton customers. The rollout will support CPU-intensive agentic AI workloads including real-time reasoning, code generation, and multi-step task orchestration.

Each Graviton5 chip features 192 cores, a cache five times larger than its predecessor and 3-nanometer chip technology, reducing inter-core communication delays by up to 33% while delivering 25% better performance. These chips leverage the AWS Nitro System and Elastic Fabric Adapter for low-latency, high-bandwidth communication across instances.

The deal diversifies Meta's compute infrastructure beyond GPU-centric training to CPU-optimized inference and orchestration at scale, enhancing efficiency and cost management for large-scale AI deployments. It builds on Meta's existing chip partnerships with Nvidia, AMD, and collaboration on Arm-based architectures, reinforcing long-term infrastructure resilience.