On April 20, 2026, Anthropic announced a major expansion of its strategic partnership with Amazon. Over the next ten years, Anthropic will commit more than 100 billion USD (about 15.7 trillion yen) to AWS technologies and secure up to 5 gigawatts (GW) of Trainium-based compute capacity for training and serving Claude. Amazon, in turn, is investing an additional 5 billion USD (about 785 billion yen) immediately, with the option to invest up to a further 20 billion USD (about 3.14 trillion yen) over time. For anyone who has been using Claude as a daily companion, this is a thrilling sign that the long-running compute crunch is finally being addressed at the right scale.
More than 100B USD to AWS, locking in Trainium2 through Trainium4
The core of the new agreement is Anthropic's commitment of more than 100 billion USD (about 15.7 trillion yen) to AWS technologies over the next ten years, along with up to 5 gigawatts of new capacity dedicated to training and running Claude. The deal spans Graviton and Trainium2 through Trainium4 silicon, plus an option to purchase future generations of Amazon's custom chips as they become available. Source: Anthropic, "Anthropic and Amazon expand collaboration for up to 5 gigawatts of new compute" (https://www.anthropic.com/news/anthropic-amazon-compute)
The ramp is moving faster than expected. Significant Trainium2 capacity is coming online in Q2 2026, and scaled Trainium3 capacity is expected to come online later this year. By the end of 2026, nearly 1GW of combined Trainium2 and Trainium3 capacity is set to be online. Combined with Project Rainier, where Anthropic already runs Claude training and serving on more than one million Trainium2 chips, this puts Anthropic on track to operate one of the largest AI compute footprints in the world. Amazon CEO Andy Jassy noted that "Anthropic's commitment to run its large language models on AWS Trainium for the next decade reflects the progress we've made together on custom silicon."
Amazon adds 5B USD now, total commitment grows to roughly 33B USD
The capital side of the deal is just as large in scale. Amazon is investing 5 billion USD (about 785 billion yen) in Anthropic immediately, with up to an additional 20 billion USD (about 3.14 trillion yen) committed in the future. This builds on Amazon's previous 8 billion USD (about 1.256 trillion yen) of investment, bringing Amazon's total potential commitment to Anthropic to as much as 33 billion USD. Source: CNBC, "Amazon to invest up to another $25 billion in Anthropic as part of AI infrastructure deal" (https://www.cnbc.com/2026/04/20/amazon-invest-up-to-25-billion-in-anthropic-part-of-ai-infrastructure.html)
To put the numbers in perspective on the demand side, Anthropic's run-rate revenue has now surpassed 30 billion USD (about 4.71 trillion yen), up from approximately 9 billion USD (about 1.413 trillion yen) at the end of 2025. With demand growing more than threefold in such a short period, a multi-billion-dollar annual commitment alone clearly would not be enough, which makes this scale of agreement easier to understand. ※Exchange rate as of May 4, 2026: 1 USD = 157 JPY
Claude Platform on AWS completes the enterprise on-ramp beyond Bedrock
Another headline of the announcement is the upcoming launch of "Claude Platform on AWS." Customers will be able to access the full Claude Platform directly inside their AWS accounts, with the same account, the same access controls, and the same billing, and without needing additional credentials or contracts. More than 100,000 customers already run Claude on Amazon Bedrock, and Claude Platform on AWS layers on top of that to make full-feature enterprise adoption smoother for organizations that have to meet strict governance and compliance requirements. Source: Anthropic, "Anthropic and Amazon expand collaboration for up to 5 gigawatts of new compute" (https://www.anthropic.com/news/anthropic-amazon-compute)
International expansion is also baked in. Anthropic will scale up inference capacity in Asia and Europe to improve responsiveness and availability for users in each region. Claude remains the only frontier AI model available across all three of the world's largest cloud platforms — AWS (Bedrock), Google Cloud (Vertex AI), and Microsoft Foundry — while doubling down on AWS as its primary home.
A direct response to the reliability strain, and a hardware diversification play
On the user side, Anthropic openly acknowledges that recent rapid consumer growth has affected reliability and performance for free, Pro, Max, and Team users, especially during peak hours. Today's agreement is set to deliver meaningful additional compute within the next three months and bring nearly 1GW online by the end of the year — a direct lever for improving the everyday Claude experience. Source: Anthropic, "Anthropic and Amazon expand collaboration for up to 5 gigawatts of new compute" (https://www.anthropic.com/news/anthropic-amazon-compute)
Strategically, this Amazon deal builds on Anthropic's multi-gigawatt expansion of its partnership with Google and Broadcom announced on April 6, 2026. By distributing workloads across Trainium, TPU, and GPU silicon, Anthropic is reducing single-vendor risk while taking advantage of the strengths of each chip generation. Anthropic CEO Dario Amodei said, "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," underscoring that infrastructure expansion is now a top priority alongside frontier model research. Source: TechCrunch, "Anthropic takes $5B from Amazon and pledges $100B in cloud spending in return" (https://techcrunch.com/2026/04/20/anthropic-takes-5b-from-amazon-and-pledges-100b-in-cloud-spending-in-return/)
Summary
This agreement marks a clear shift in the Anthropic-Amazon relationship: Amazon is no longer just a major investor, it is locked in as Anthropic's primary training and cloud provider. With a 100B USD commitment over ten years, 5GW of Trainium capacity, the upcoming Claude Platform on AWS, and up to 25B USD of additional investment headroom, the scale of the numbers points to a meaningful change in both the Claude user experience and the broader AI infrastructure landscape. The scaled rollout of Trainium3 and the launch of Claude Platform on AWS are now firmly on the horizon to watch.
