Anthropic and Amazon Expand AI Infrastructure Deal

Anthropic has signed a new agreement with Amazon to secure up to 5 gigawatts (GW) of compute capacity for training and deploying its Claude AI model. The deal includes over $100 billion in commitments to AWS technologies over the next decade, spanning Graviton and Trainium chips through Trainium4. Significant Trainium2 capacity is coming online in Q2 2026, with scaled Trainium3 capacity later this year, and nearly 1GW total of Trainium2 and Trainium3 by end of 2026. Amazon is investing $5 billion now with up to an additional $20 billion, building on $8 billion previously invested. Anthropic's run-rate revenue has surpassed $30 billion, up from $9 billion at end of 2025. The full Claude Platform will be available directly within AWS, and expansion of inference in Asia and Europe is planned. Anthropic currently uses over one million Trainium2 chips and launched Project Rainier, one of the largest compute clusters globally. CEO Dario Amodei stated the collaboration will advance AI research and serve over 100,000 customers building on AWS. Andy Jassy, CEO of Amazon, noted the high demand for custom AI silicon. Claude remains the only frontier AI model available on AWS, Google Cloud, and Microsoft Azure.