Anthropic locks in 5GW of AWS compute with up to $25B Amazon bet
Amazon drops $5B into Anthropic now, with $20B more on the table — plus 5 gigawatts of compute to train and serve Claude at scale.
Amazon just made its biggest AI infrastructure commitment yet. It's investing an additional $5 billion in Anthropic today, with up to $20 billion more to follow — stacked on top of the $8 billion it had already put in. In exchange, Anthropic commits to spending $100 billion securing up to 5 gigawatts of AWS compute to train and run Claude.
The timing isn't coincidental. Anthropic's run-rate revenue has surpassed $30 billion, up from ~$9B just four months ago. That kind of growth puts serious pressure on infrastructure — the company admitted a "sharp rise" in consumer and enterprise usage has caused "inevitable strain" on reliability. Nearly 1GW of Trainium2 and Trainium3 capacity is expected online by end of 2026, with Anthropic already running 1M+ Trainium2 chips in production.
The deal also expands international inference capacity across Asia and Europe, signaling Claude is pushing hard into global enterprise markets.
Why it matters: compute is now the real moat — and this deal makes Amazon Anthropic's engine for the next decade of AI scaling.
Sources
Primary: the company, paper or repository
Independent coverage
Written by an AI pipeline from the sources above. Methodology · Report an error
Feed, daily deep-dive and bytes — readable offline, with push alerts for the topics you follow.