We're expanding our collaboration with Amazon to secure up to 5 gigawatts of compute for training and deploying Claude. Capacity begins coming online this quarter, with nearly 1 gigawatt expected by the end of 2026.
Anthropic Secures 5 Gigawatts of Amazon Compute Amid 30 Billion Revenue Surge
· Updated
Anthropic signed an agreement with Amazon to secure up to 5 gigawatts of compute capacity for training and deploying Claude. The deal includes an immediate $5 billion investment and a $100 billion commitment to use AWS custom silicon like Trainium (specialized AI training chips) over the next decade.
- Total compute secured
- Up to 5 gigawatts
- Immediate investment
- $5 billion
- Future investment option
- Up to $20 billion
- Infrastructure commitment
- $100 billion over 10 years
- Revenue run-rate
- $30 billion
- Immediate capacity
- Nearly 1 gigawatt by end of 2026
This expansion follows a revenue surge that saw the company's run-rate triple to $30 billion. The resulting strain on infrastructure has impacted reliability for users. By securing gigawatt-scale power, Anthropic is building a hardware moat that extends its multi-gigawatt deal with Google to ensure frontier model stability.
You can now access the full Claude Platform natively within AWS, using your existing account, billing, and security controls. This move mirrors Meta's deployment of millions of AWS Graviton cores for agentic AI, signaling a broader industry shift toward custom cloud silicon. This infrastructure will also support scaling production-grade AI agents.
Anthropic
@AnthropicAI
102retweets1.3klikes
View on XStill wondering? A few quick answers below.
Anthropic and Amazon have expanded their partnership to secure up to 5 gigawatts of compute capacity, or raw processing power, for training and deploying Claude models. The deal includes a $100 billion commitment from Anthropic to use AWS technologies over the next decade and an immediate $5 billion investment from Amazon.
The full Claude Platform is now available directly within AWS, allowing organizations to use the same account, billing, and security controls they already have. This integration removes the need for additional contracts, making it easier for enterprises to meet governance and compliance requirements while using Anthropic's frontier AI models on Amazon Bedrock.
New capacity is coming online starting in the second quarter of 2026. Anthropic expects nearly 1 gigawatt of total capacity from Amazon's custom Trainium2 and Trainium3 AI chips to be available by the end of 2026. This expansion is designed to address performance and reliability issues caused by the company's rapid growth and high user demand.





