Anthropic and Amazon expanded their partnership on April 20 with a deal that secures up to 5 gigawatts of new compute capacity for training and running Claude, alongside a fresh $5 billion Amazon investment and up to $20 billion more to follow. Anthropic is committing over $100 billion to AWS technologies over the next ten years, one of the largest cloud spending pledges ever disclosed. The expansion lands as Anthropic's run-rate revenue crossed $30 billion, up from roughly $9 billion at the end of 2025.
The capacity is front-loaded. New Trainium2 supply comes online in Q2 2026, and Anthropic expects nearly 1GW of combined Trainium2 and Trainium3 capacity operational by the end of 2026. Scaled Trainium3 capacity is slated for later this year, with Graviton and eventual Trainium4 chips also part of the roadmap.
Amazon's $5 billion top-up, plus the optional $20 billion tranche, stacks on $8 billion already committed in prior rounds. That puts Amazon's total potential exposure to Anthropic at $33 billion, a figure that rivals any strategic investment in the AI sector outside of Microsoft's position in OpenAI.
Key facts
- 01Anthropic and Amazon signed a deal for up to 5GW of compute capacity for training and running Claude.
- 02Anthropic will commit more than $100 billion to AWS technologies over the next ten years.
- 03Amazon is investing $5 billion in Anthropic now, with up to $20 billion more to follow, on top of $8 billion previously invested.
- 04Anthropic's run-rate revenue has passed $30 billion, up from about $9 billion at the end of 2025.
- 05Nearly 1GW of combined Trainium2 and Trainium3 capacity is set to come online by the end of 2026.
The two companies have worked together since 2023. More than 100,000 customers now run Claude on Amazon Bedrock, and the pair jointly built Project Rainier, which Anthropic says is one of the world's largest compute clusters. Anthropic already uses over one million Trainium2 chips to train and serve Claude.
“Anthropic's run-rate revenue has surpassed $30 billion, up from roughly $9 billion at the end of 2025, with over 1,000 business customers each spending more than $1 million a year.”— Jaeden Schafer
Andy Jassy, CEO of Amazon, tied the deal directly to AWS's custom silicon bet. "Our custom AI silicon offers high performance at significantly lower cost for customers, which is why it's in such hot demand," Jassy said, framing Anthropic's ten-year Trainium commitment as validation of the Trainium program.
For Anthropic, the driver is demand that is outrunning supply. CEO Dario Amodei said consumer growth has strained reliability on the free, Pro, Max, and Team tiers during peak hours. The new capacity is meant to deliver meaningful compute within three months and close the gap before the end of the year.
The Amazon deal follows Anthropic's April 6 agreement with Google and Broadcom for multiple gigawatts of next-generation TPU capacity beginning in 2027, which AI Chat Daily covered earlier this month. Between the two pacts, Anthropic is locking in compute across AWS Trainium, Google TPUs, and NVIDIA GPUs, a deliberate hedge that lets it route workloads to whichever chip fits best.
The commercial momentum behind the spending is steep. In February 2026, Anthropic disclosed that more than 500 business customers were each spending over $1 million annualized on Claude. By April that number had doubled to more than 1,000, with run-rate revenue roughly tripling inside four months.
Anthropic is also keeping its distribution advantage: Claude is the only frontier model available on all three major clouds, AWS Bedrock, Google Cloud Vertex AI, and Microsoft Azure Foundry. A new Claude Platform on AWS, giving customers direct access inside their existing AWS accounts and billing, is coming soon.
The obvious risk is concentration. Anthropic has now pledged over $100 billion to AWS over a decade and, separately, extended its November 2025 commitment to invest $50 billion in US computing infrastructure via the Google and Broadcom deal. Those are multi-year obligations underwritten by current growth rates that, however steep, have never been tested through a downturn in enterprise AI spend.
The read for the market is that Anthropic is behaving like a company that believes the compute crunch, not model quality, is the binding constraint on its growth. Locking in 5GW from Amazon and multiple gigawatts from Google gives it a supply position few competitors can match. If Claude usage keeps compounding from here, the $100 billion AWS commitment will look conservative. If it doesn't, Amazon's $33 billion of potential exposure becomes the most interesting number in AI.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
