Amazon has expanded its strategic partnership with Anthropic by investing an additional $5 billion in the AI company and securing a commitment for Anthropic to spend more than $100 billion on Amazon Web Services over the next 10 years. The agreement deepens one of the largest infrastructure-and-equity alliances in the artificial intelligence sector and reinforces AWS as Anthropic’s primary platform for large-scale model training and mission-critical cloud workloads.
The latest capital injection lifts Amazon’s total committed investment in Anthropic since 2023 to $13 billion. Amazon also indicated that future funding tied to commercial milestones could add as much as $20 billion more, bringing the broader potential relationship to nearly $33 billion. Even so, Amazon remains a minority investor in the company.
A long-term infrastructure pact built around custom silicon
At the center of the deal is Anthropic’s pledge to use AWS infrastructure at enormous scale over the next decade. The spending commitment includes current and future generations of Amazon’s Trainium and Graviton chips, along with tens of millions of Graviton cores. In practical terms, the arrangement gives Anthropic access to as much as 5 gigawatts of new compute capacity to train and serve its Claude family of models.
Amazon said significant Trainium2 capacity is expected to come online in the second quarter of 2026, with nearly 1 gigawatt of combined Trainium2 and Trainium3 capacity projected by the end of the year. The agreement highlights how leading AI labs are increasingly relying on dedicated compute reservations and vertically integrated chip roadmaps rather than spot market infrastructure.
Amazon CEO Andy Jassy framed the commitment as validation of the company’s progress in custom silicon. He said Anthropic’s decision to run its large language models on AWS Trainium over the next decade reflects the advances both companies have made together on chip performance and cost efficiency.
Anthropic’s growth is forcing a new infrastructure scale
The urgency behind the deal is closely tied to Anthropic’s rapid business expansion. According to the announcement, Claude’s run-rate revenue has climbed to more than $30 billion in 2026, up sharply from about $9 billion at the end of 2025. That growth has been fueled by rising adoption across enterprise, developer, and consumer segments, including free, Pro, Max, and Team offerings.
Anthropic CEO and co-founder Dario Amodei said customer demand is changing how the company operates. He noted that users increasingly see Claude as essential to their work, making it necessary for Anthropic to build infrastructure fast enough to keep pace with the growth in usage. He added that the collaboration with Amazon will support continued AI research while helping the company deliver Claude to customers, including the more than 100,000 already building on AWS.
The infrastructure expansion is not only about training larger models. Anthropic and AWS are also scaling inference capacity across Asia and Europe to address growing international demand. This is increasingly important as global enterprises adopt AI tools in production environments that require low latency, high reliability, and regional availability.
Claude becomes more tightly integrated into AWS
One immediate operational change is the direct availability of the full Claude Platform console inside AWS. Customers can now access it through existing AWS accounts, controls, and billing systems, eliminating the need for separate credentials or standalone commercial contracts. That tighter integration lowers friction for enterprise deployment and further embeds Claude within Amazon’s cloud ecosystem.
Anthropic’s model lineup also retains a distinctive market position. Claude remains the only frontier AI model available across all three major cloud platforms: AWS Bedrock, Google Cloud Vertex AI, and Microsoft Azure Foundry. Even so, AWS appears to be strengthening its role as the company’s primary infrastructure base through this latest agreement.
Amazon said that more than 100,000 customers are already running Claude models through Amazon Bedrock. Among the examples cited, Lyft reported an 87% improvement in customer service resolution speed, while Pfizer said it achieved a 55% reduction in infrastructure costs and saved 16,000 search hours annually. These examples are being used to underscore the growing commercial case for large-scale deployment of Claude in enterprise settings.
The broader AI capital cycle keeps accelerating
The Amazon-Anthropic agreement arrives during a period of exceptional private capital formation in AI. From mid-February to mid-April 2026, OpenAI and Anthropic together reportedly completed funding rounds totaling more than $150 billion combined. The scale of those financings reflects how rising GPU costs, data center expansion, and energy requirements are driving unprecedented capital needs for top AI labs.
This latest transaction also fits a broader pattern in the market: major technology companies are exchanging long-term compute guarantees, preferred platform access, and infrastructure commitments for strategic equity stakes in leading AI developers. Rather than simply acting as vendors, cloud providers are becoming financing partners and ecosystem anchors.
Anthropic and Amazon’s Annapurna Labs will continue to collaborate on custom silicon development as part of the expanded relationship. The companies are also moving forward with Project Rainier, a large-scale AI compute cluster built around nearly 500,000 Trainium2 chips. Its planned expansion signals Amazon’s intent to compete more aggressively in AI infrastructure not only through cloud services, but through its own chip architecture and deeply integrated compute supply.
For Amazon, the message is clear: strategic investment in AI labs can drive massive long-term cloud demand. For Anthropic, the partnership offers both capital and a pathway to secure compute at a scale increasingly necessary to train and serve frontier models. In a market where access to infrastructure may be just as important as access to funding, the deal marks another step toward tighter consolidation between AI model builders and hyperscale cloud platforms.