Meta Scales AI Infrastructure with Massive AWS Graviton Chip Deployment

Meta is significantly expanding its AI hardware footprint by adopting Amazon Web Services' (AWS) Graviton chips in a deal spanning at least three years. The social media giant will deploy hundreds of thousands of these general-purpose processors to support its massive computing requirements across its global data center network. This strategic pivot comes as Meta continues an aggressive AI infrastructure build-out. The company recently committed $48 billion to CoreWeave and Nebius to secure access to Nvidia GPUs. While GPUs handle the heavy lifting of initial model training, the Graviton chips will be utilized for post-training refinements and the CPU-intensive tasks required for agentic AI. AWS claims the Graviton platform provides superior price-performance and consumes 60% less energy than alternative computing options. Meta's adoption places it among the top five Graviton customers, joining other major firms such as Apple, Adobe, and Snowflake, as well as AI builder Anthropic. To offset the immense capital expenditure required for these AI investments, Meta is simultaneously streamlining its operations. The company recently announced a workforce reduction of approximately 8,000 employees, representing 10% of its total staff. The shift highlights a growing trend where CPUs are re-establishing their importance as the foundation of AI architecture. While Nvidia remains dominant in training, the move toward ARM-based efficiency suggests a maturing infrastructure phase focused on operational costs and energy sustainability.