Anthropic just dropped a bombshell announcement that makes their previous deals with Amazon look like pocket change. The company has signed a new agreement that commits over $100 billion to AWS technologies over the next decade, securing up to 5 gigawatts of compute capacity for training and running Claude. That’s not a typo — $100 billion.
Let’s break down what’s actually happening here, because there’s a lot to unpack.
The compute numbers are staggering
Anthropic is already using over a million Trainium2 chips to train and serve Claude. They launched Project Rainier together, one of the largest compute clusters in the world. Now they’re planning to scale that up dramatically. New Trainium2 capacity is coming online in Q2 of this year, and nearly 1GW total of Trainium2 and Trainium3 capacity should be online by the end of 2026. The commitment spans Graviton and Trainium chips through the Trainium4 generation, with options to purchase future custom silicon as Amazon iterates.
This is the kind of infrastructure spending that makes you realize how capital-intensive frontier AI really is. $100 billion over ten years is roughly $10 billion a year — more than most countries spend on their entire tech sectors.
The revenue story is even more interesting
Anthropic’s run-rate revenue has surpassed $30 billion, up from approximately $9 billion at the end of 2025. That’s triple-digit growth in less than 18 months. The company says enterprise and developer demand for Claude has accelerated in 2026, and consumer usage across their free, Pro, and Max tiers has spiked sharply.
But here’s the honest part they admitted: “Growth at this pace places an inevitable strain on our infrastructure; our unprecedented consumer growth, in particular, has impacted reliability and performance for free, Pro, Max, and Team users, especially during peak hours.” Translation: Claude has been struggling under the load, and this deal is meant to fix that.
Claude Platform is coming directly to AWS
This is a big deal for enterprise customers. The full Claude Platform will be available directly within AWS — same account, same controls, same billing. No additional credentials or contracts needed. Organizations can access Claude while meeting their existing governance and compliance requirements. It’s coming soon, and you can request access through your AWS account team.
Claude remains the only frontier AI model available on all three major cloud platforms: AWS Bedrock, Google Cloud Vertex AI, and Microsoft Azure Foundry. That’s a smart hedge — don’t put all your eggs in one basket, even if Amazon is your primary partner.
Amazon is putting more money where its mouth is
Amazon is investing $5 billion in Anthropic today, with the option for up to an additional $20 billion in the future. That builds on the $8 billion Amazon had previously invested. Andy Jassy, Amazon’s CEO, highlighted the performance and cost advantages of their custom AI silicon, which is clearly in hot demand.
Dario Amodei, Anthropic’s CEO, framed it as building infrastructure to keep pace with demand: “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.”
What this means for the AI landscape
This deal cements Anthropic as a serious player in the infrastructure arms race alongside OpenAI and Google DeepMind. The $30 billion run rate is eye-popping, but the real story is the commitment to compute. Training frontier models requires massive clusters, and Anthropic is locking in capacity years ahead.
The announcement also signals that Amazon is all-in on custom silicon. Trainium is their bet to compete with NVIDIA’s GPUs, and having Anthropic as an anchor customer gives them credibility. If Trainium can deliver high performance at lower cost, it could shift the economics of AI training for everyone.
One thing I appreciate about this announcement: they didn’t sugarcoat the reliability issues. Consumer growth has been a double-edged sword, and they’re being transparent about the strain. That’s refreshing in an industry that often pretends everything is running smoothly.
Overall, this is a massive bet on the future of AI infrastructure. Whether it pays off depends on whether Claude can maintain its competitive edge and whether the economics of training and inference continue to improve. But with $100 billion on the table, Anthropic and Amazon are clearly betting big.
Comments (0)
Login Log in to comment.
Be the first to comment!