Anthropic and Amazon Team Up for Up to 5 GW of New Compute for Claude!

Hey everyone, it's Shiichan! Today I've got a huge piece of news from Anthropic and Amazon to share with you.

Anthropic News anthropic.com

What was announced?

Over on Anthropic's News, Anthropic and Amazon announced a new agreement that deepens their partnership and secures up to 5 GW of compute for training and running Claude. New Trainium2 capacity is coming online in the first half of this year, and nearly 1 GW of Trainium2 and Trainium3 capacity in total is set to come online by the end of 2026.

The story so far

Anthropic and Amazon have worked closely together since 2023, and today more than 100,000 customers run Claude on Amazon Bedrock. Together they launched Project Rainier, one of the largest compute clusters in the world, and they currently use over one million Trainium2 chips to train and serve Claude. This new agreement builds on all of that and expands things in three directions.

What changes

The biggest thing is a big jump in the compute available to run Claude. With demand rising so fast, reliability and performance have taken a hit for free, Pro, Max, and Team users, especially during peak hours. This agreement adds meaningful compute within the next three months, with nearly 1 GW in total arriving before the end of the year, so using Claude should keep feeling smoother over time.

Dive Deep

There are three main pieces here.

First, infrastructure at scale. Anthropic is committing more than $100 billion to AWS technologies over the next ten years, securing up to 5 GW of new capacity. That spans Graviton and Trainium2 through Trainium4 chips, with the option to buy future generations of Amazon's custom silicon. Significant Trainium2 capacity arrives in Q2, and scaled Trainium3 is expected later this year. Inference is also expanding into Asia and Europe to serve Claude's growing international customer base, and AWS remains the primary training and cloud provider for mission-critical workloads.

Second, Claude Platform on AWS. The full Claude Platform will be available directly within AWS, with the same account, same controls, and same billing, and no extra credentials or contracts needed. That keeps Claude the only frontier model available across all three of the world's largest clouds: AWS (Bedrock), Google Cloud (Vertex AI), and Microsoft Azure (Foundry). You can learn more on the Anthropic on Amazon Bedrock page. It's coming soon.

Third, continued investment. Amazon is investing $5 billion in Anthropic today, with up to an additional $20 billion in the future, on top of the $8 billion it had already invested.

Dario Amodei, CEO and co-founder of Anthropic, put it this way:

Our users tell us Claude is increasingly essential to how they work, and we need to build the infrastructure to keep pace with rapidly growing demand.

Behind all this is record demand. Anthropic's run-rate revenue has now passed $30 billion, up from around $9 billion at the end of 2025. Alongside this agreement, Anthropic is combining additional capacity expansions and a diversified strategy that spreads workloads across a range of chips to keep Claude at the frontier and serve customers reliably.

Wrap-up

• Anthropic and Amazon expand their partnership, securing up to 5 GW of new compute for Claude
• Anthropic commits more than $100 billion to AWS over ten years, covering Graviton and Trainium2 through Trainium4
• The full Claude Platform is coming to AWS soon as "Claude Platform on AWS"
• Amazon invests $5 billion today, with up to $20 billion more in the future
• Nearly 1 GW of capacity arrives by year-end, which should improve reliability during busy hours

This is a story about big money and chips, so it's not the kind of news you can go and try yourself, but if you use Claude at work or you're planning to run AI on AWS, it's an announcement that quietly shapes your future reliability and options.