Amazon completed a $4 billion investment in Anthropic, the safety-focused lab founded by former OpenAI researchers, cementing a partnership under which Anthropic would use AWS as a primary cloud provider and Amazon's custom Trainium and Inferentia chips.

The deal crystallized the era's structure: each hyperscaler was underwriting a frontier lab. Microsoft had OpenAI; Amazon and, separately, Google were backing Anthropic; the cloud giants were effectively financing multiple foundation-model efforts to avoid ceding the category to a single rival.

For Anthropic, the capital and compute were existential — frontier training runs had become extraordinarily expensive — and the AWS distribution put its Claude models in front of enterprise buyers already spending on Amazon's cloud.