Anthropic has secured $25 billion in total investment from Amazon and committed over $100 billion to AWS infrastructure in a decade-long partnership that will provide up to 5 gigawatts of compute capacity for training and deploying Claude.

The AI safety lab announced a $5 billion immediate investment from Amazon, with up to $20 billion more available in future tranches. This builds on Amazon's previous $8 billion investment in the company.

The expanded partnership centers on three key areas. Anthropic will gain access to massive Trainium2 and Trainium3 chip capacity, with significant Trainium2 resources coming online in Q2 2026 and scaled Trainium3 capacity by year-end.

The full Claude Platform will integrate directly into AWS, allowing enterprise customers to access Claude through their existing AWS accounts without additional contracts or credentials. Claude remains available across all three major cloud platforms: AWS Bedrock, Google Cloud Vertex AI, and Microsoft Azure Foundry.

Anthropic currently uses over one million Trainium2 chips and operates Project Rainier, described as one of the world's largest compute clusters. The company serves more than 100,000 customers through Amazon Bedrock.

"Our custom AI silicon offers high performance at significantly lower cost for customers, which is why it's in such hot demand," said Andy Jassy, Amazon's CEO.

The partnership addresses Anthropic's infrastructure strain from rapid growth. The company's run-rate revenue has surged to over $30 billion, up from approximately $9 billion at the end of 2025.

Meeting unprecedented demand

Enterprise and consumer demand for Claude has accelerated sharply in 2026, creating reliability issues during peak hours for free, Pro, Max, and Team users. The infrastructure bottleneck has particularly impacted consumer tiers as usage spiked.

Dario Amodei, Anthropic's CEO and co-founder, said the collaboration will allow continued AI research advancement while serving growing customer demand. "Our users tell us Claude is increasingly essential to how they work, and we need to build the infrastructure to keep pace."

The agreement will deliver meaningful compute capacity within three months and nearly 1 gigawatt total before the end of 2026. Combined with additional capacity expansions and Anthropic's diversified hardware strategy across multiple chip types, the infrastructure build-out aims to maintain Claude's frontier performance while ensuring reliable service.

Amazon's Trainium chips are designed specifically for AI training and inference workloads, offering what the companies describe as high performance at significantly lower costs than traditional GPU alternatives.