Anthropic and Amazon Expand Partnership, Securing Up to 5GW of Compute for Claude

Anthropic announced on April 20, 2026, a major expansion of its collaboration with Amazon that will secure up to 5 gigawatts (GW) of new compute capacity to train and deploy Claude. The agreement accelerates capacity coming online this year and ties together deeper infrastructure, platform integration, and additional capital investment — steps Anthropic says are needed to meet surging customer demand and keep Claude reliable at scale.

What the agreement covers

The deal spans three core areas: infrastructure at scale, platform integration, and continued investment. Anthropic committed to more than $100 billion in AWS technologies over the next ten years, with options to purchase future generations of Amazon’s custom silicon. Significant Trainium2 capacity is expected in the first half of 2026, and scaled Trainium3 capacity later in the year. In total, Anthropic expects nearly 1GW of Trainium2 and Trainium3 capacity to be added by year’s end. The partnership also expands inference capacity in Asia and Europe and keeps AWS as Anthropic’s primary cloud provider for mission-critical workloads.

Why this matters for reliability and growth

Demand for Claude has accelerated sharply in 2026 across enterprise, developer, and consumer tiers. Anthropic reports that growing usage has strained infrastructure — affecting reliability and performance for free, Pro, Max, and Team customers during peak hours. The new capacity commitment aims to relieve that pressure quickly, with meaningful compute available within months and a steady ramp to the larger 5GW scope. In practical terms, more purpose-built AI silicon and a diversified hardware strategy should reduce bottlenecks, lower cost per operation, and improve latency for real-world applications.

The technical footprint: Trainium, Graviton, and Project Rainier

Anthropic and Amazon’s work together has already produced large-scale infrastructure, including Project Rainier — described as one of the largest compute clusters in the world. Anthropic currently uses over one million Trainium2 chips for training and serving Claude. The expanded agreement covers Graviton and Trainium families (Trainium2 through Trainium4), so Anthropic can scale across multiple generations of Amazon’s custom silicon as it becomes available. That diversity helps spread workloads and reduces single-point dependencies in their hardware stack.

Platform integration and customer experience

A notable component of the announcement is the deeper integration of the full Claude Platform into AWS. Claude Platform on AWS will operate within customers’ existing AWS accounts, keeping the same controls and billing, which simplifies governance, compliance, and procurement for organizations already invested in AWS. Anthropic notes that Claude remains uniquely available across all three major cloud ecosystems — AWS (Bedrock), Google Cloud (Vertex AI), and Microsoft Azure (Foundry) — while this new integration brings more of Claude’s capabilities directly into the AWS ecosystem. Claude Platform on AWS is entering private beta, with access via account teams.

Financial commitments and strategic investment

Amazon is investing $5 billion in Anthropic immediately as part of this expanded collaboration, with the potential for up to an additional $20 billion in the future. This builds on prior investments totaling $8 billion. Anthropic also announced a long-term spend commitment exceeding $100 billion on AWS technologies over the next decade. For context, Anthropic reported a run-rate revenue surpassing $30 billion in early 2026, up from roughly $9 billion at the end of 2025 — growth that helps explain the aggressive infrastructure and investment posture.

What this means for users and the market

For enterprises and developers, the expansion promises faster access to Claude capabilities with fewer integration hurdles on AWS, improved regional availability, and better performance during peak usage. For Anthropic, the deal reduces operational risk and positions the company to scale its research and product deployments. For the cloud market, it underscores the strategic value of custom AI silicon and long-term partnerships between major model builders and hyperscalers.

Looking ahead

Anthropic and Amazon’s expanded collaboration is squarely aimed at matching infrastructure to demand: short-term capacity to stabilize service and a long-term pipeline of custom silicon and cloud resources to sustain growth. As Trainium3 and further generations roll out, and as Claude Platform on AWS moves beyond private beta, the practical effects — lower latency, higher throughput, and wider geographic availability — will be the real test of this ambitious commitment.