Amazon's custom AI chips gain major validation: Meta and Anthropic commit billions to AWS silicon, justifying Amazon's massive AI spending spree

Meta and Amazon announced Friday that Meta will use AWS Graviton5 CPU chips (the processors that power AI workloads) in a multiyear deal covering 'tens of millions of cores,' while Anthropic committed more than $100 billion over 10 years to AWS technologies and Amazon invested an additional $5 billion in Anthropic (with up to $20 billion more possible). This follows Amazon's reveal that its chip business already has an annual revenue run rate exceeding $20 billion and is growing at triple-digit year-over-year rates.

Unlike Nvidia's GPUs (which train AI models), Amazon's custom chips excel at inference and agentic workloads—the part where trained AI models are actually deployed and used in real applications. This means customers can run their AI systems at lower cost than before, addressing what Amazon CEO Andy Jassy called the market's hunger for 'better price performance.'

For business professionals and enterprises: these deals signal that Amazon's $200 billion capital expenditure plan for 2026 has genuine customer demand backing it. If you use AWS or are evaluating cloud providers for AI workloads, Amazon's chips now represent a concrete cost-saving alternative to relying solely on expensive GPU-based solutions.

However, Amazon's stock already trades at a price-to-earnings ratio of 37 and has jumped 25% in 30 days—meaning much of this good news may already be reflected in the share price. New investors should be cautious about entry timing, especially given the company's massive spending commitments.