Meta signs deal for millions of Amazon AI CPUs in chip strategy shift
Original: In another wild turn for AI chips, Meta signs deal for millions of Amazon AI CPUs
Why This Matters
Demonstrates evolving AI chip requirements as industry shifts from training to inference workloads
Meta has signed a deal to use millions of AWS Graviton ARM-based CPUs for AI workloads, marking a shift from GPUs to CPUs for AI inference tasks. The deal brings Meta's spending back to AWS after its $10 billion Google Cloud agreement last year.
Amazon announced Meta will use millions of AWS Graviton chips to power its AI needs, representing a strategic shift in AI chip requirements. While GPUs remain preferred for training large models, AI agents are driving demand for CPUs that handle compute-intensive tasks like real-time reasoning, code writing, and multi-step coordination. AWS's latest Graviton version was specifically designed for AI-related computing needs. This deal redirects Meta's cloud spending back to AWS after the company signed a six-year, $10 billion agreement with Google Cloud in August. The timing coincided with Google Cloud Next conference's conclusion. Amazon also makes Trainium AI GPUs, but Anthropic secured most of those chips through a $100 billion, 10-year AWS deal announced earlier this month.