Amazon's $25 Billion Anthropic Investment Accelerates Custom AI Chip Growth — Trainium at the Centre of a Decade-Long Infrastructure Deal
AI / AI Tech Trends | 4 min read
Amazon has announced an expanded partnership with Anthropic — committing up to $25 billion in new investment, with $5 billion effective immediately and up to $20 billion more tied to certain commercial milestones. This adds to the $8 billion Amazon had already invested in Anthropic — bringing total potential investment to approximately $33 billion. In return, Anthropic has committed to spending more than $100 billion on Amazon Web Services technologies over the next decade, specifically covering current and future generations of Trainium — Amazon's custom AI chips. Anthropic has secured up to 5 gigawatts of computing capacity for training and deploying its Claude AI models, with nearly 1 gigawatt of Trainium2 and Trainium3 capacity expected to come online by the end of 2026. Anthropic already operates more than 1 million Trainium2 chips, and will gain priority access to Trainium3 (released December 2025) and Trainium4 in future generations.
"Our custom AI silicon offers high performance at significantly lower cost for customers, which is why it's in such hot demand. 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, as we continue delivering the technology and infrastructure our customers need to build with generative AI."
— Andy Jassy, Chief Executive Officer, Amazon
Trainium at Scale — AWS Infrastructure and the Custom Chip Bet
The deal places Amazon's custom Trainium silicon at the centre of one of the most significant AI infrastructure commitments ever announced. Anthropic named AWS its primary cloud provider in 2023 and primary training partner in 2024 — Claude now runs natively in AWS accounts with the same billing and controls as other AWS services. More than 100,000 customers build on Amazon Bedrock using Claude. Amazon has budgeted approximately $200 billion in capital expenditure for 2026, the majority earmarked for AI infrastructure. The Anthropic deal directly supports Amazon's strategy of using major AI partners to establish Trainium as a credible alternative to third-party processors in the training and inference market. Trainium is already reported to be a multibillion-dollar business growing at triple-digit rates. Expansions of Trainium capacity are also planned in Asia and Europe.
"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. Our collaboration with Amazon will allow us to continue advancing AI research while delivering Claude to our customers, including the more than 100,000 building on AWS."
— Dario Amodei, Chief Executive Officer, Anthropic
Key Takeaways
- • Amazon has announced up to $25 billion in new investment in Anthropic ($5B immediate; up to $20B tied to commercial milestones), adding to $8B previously invested — total potential commitment approximately $33 billion. In return, Anthropic commits to spending $100B+ on AWS technologies over 10 years, covering current and future Trainium generations. Deal announced 20 April 2026.
- • Trainium at the centre: Anthropic has secured up to 5 gigawatts of compute capacity on Amazon's custom Trainium silicon. Already operates 1M+ Trainium2 chips. ~1 gigawatt of Trainium2 and Trainium3 capacity coming online by end of 2026. Priority access to Trainium3 (released December 2025) and Trainium4 in future. Trainium already a multibillion-dollar triple-digit growth business for Amazon. Asia and Europe Trainium expansion planned. Amazon's $200B 2026 capex — majority for AI infrastructure.
- • Partnership history and integration depth: Anthropic named AWS primary cloud provider in 2023 and primary training partner in 2024. Claude runs natively in AWS accounts with the same billing and controls. 100,000+ customers build on Amazon Bedrock with Claude. The expanded deal deepens what is already one of the most integrated AI model-cloud provider relationships in the industry — moving from partnership to a multi-generational, decade-long infrastructure commitment.
- • Strategic rationale for Amazon: securing Anthropic as a long-term Trainium anchor customer validates the custom silicon strategy and provides a credible alternative to third-party processors for AI training and inference at scale. Citi analysts noted that "following Amazon's expanded partnerships with OpenAI and Anthropic, we believe AWS is increasingly well positioned to deliver AI solutions and we are incrementally confident that AWS can continue to deliver accelerating revenue growth." William Blair analysts linked the deal to Anthropic's need to grow compute capacity as Claude demand grows.
- • Broader AI infrastructure context: the deal follows Amazon's $50 billion contribution to a funding round for a separate AI company (OpenAI's $110 billion round, valuing OpenAI at $730 billion pre-money). With $200 billion in planned 2026 capex and now two major long-term AI model relationships anchored to AWS and Trainium, Amazon is making one of the largest concentrated bets on AI infrastructure of any hyperscaler — using custom silicon as the differentiating layer between itself and rivals in the AI infrastructure race.