Meta has signed a deal to use millions of AWS Graviton chips, Amazon's ARM-based CPUs optimized for AI inference and agent workloads. The agreement reflects a technical shift: while GPUs train models, deployed AI agents rely on CPUs for compute-intensive tasks like reasoning, code generation, and multi-step coordination. The deal redirects Meta spending to AWS, away from its recent $10B six-year Google Cloud commitment.
Infrastructure
In another wild turn for AI chips, Meta signs deal for millions of Amazon AI CPUs
Meta commits to millions of AWS Graviton chips for deployed AI agents, signaling a shift from GPU-focused training to CPU-based inference workloads.
Friday, April 24, 2026 12:00 PM UTC2 MIN READSOURCE: TechCrunchBY sys://pipeline
Tags
infrastructure
/// RELATED
StrategyApr 22
The most interesting startups showcased at Google Cloud Next 2026
Google commits $750M to fund Cloud partners building enterprise AI agents, partnering with startups like Lovable, Notion, and Gamma to distribute Gemini-powered tools across the market.
Safety4d ago
Severe Linux Copy Fail security flaw uncovered using AI scanning help
AI-powered security scanning uncovered Copy Fail, a critical privilege escalation flaw affecting all Linux distributions since 2017 through kernel page-cache corruption in the crypto subsystem.