Amazon invests up to an additional $25 billion in Anthropic.

      The agreement marks Amazon’s second consecutive substantial investment in a frontier AI lab, coming after a similar $50 billion commitment to OpenAI two months ago. Anthropic’s annual revenue has reached $30 billion. The deal secures Anthropic's use of AWS’s Trainium chips via Trainium4 and allocates up to 5 gigawatts of compute capacity for its Claude models.

      Amazon intends to invest up to $25 billion in Anthropic, the AI safety firm behind the Claude model family, as part of a broader agreement that also guarantees over $100 billion of Anthropic's cloud expenditure on Amazon Web Services over the next decade.

      The company will make an initial investment of $5 billion, with up to an additional $20 billion contingent on meeting commercial goals. This comes in addition to the $8 billion Amazon had previously committed to Anthropic since 2023, raising the total potential investment from Amazon to approximately $33 billion.

      The arrangement's structure is noteworthy. Amazon is not providing the full $25 billion in one go; the majority of the funding depends on the company's performance. Concurrently, Anthropic’s $100 billion commitment to AWS is not a direct payment but a promise to procure cloud computing, custom AI chips, and infrastructure services from Amazon over the next ten years.

      Thus, Amazon funds Anthropic, which then uses that investment to enhance Amazon’s infrastructure. This structure was also employed two months ago during Amazon's $50 billion investment in OpenAI, which included a similar $100 billion cloud commitment.

      The agreement primarily involves Amazon’s Trainium chip series, its in-house alternative to Nvidia GPUs used for AI training and inference. The deal encompasses the current Trainium2, the upcoming Trainium3, and Trainium4, which is yet to be released, in addition to tens of millions of Graviton CPU cores.

      Anthropic is assured up to 5 gigawatts of compute capacity for training and deploying its Claude models under this agreement. Amazon indicated that substantial Trainium2 capacity will be operational in the second quarter of 2026, with nearly one gigawatt of combined Trainium2 and Trainium3 expected by the end of the year.

      This compute commitment addresses what Anthropic refers to as rapidly increasing demand: the company’s annualized revenue has surged to $30 billion, an increase from around $9 billion at the close of 2025.

      The partnership also enhances the distribution of Anthropic’s models. AWS clients can now access the complete Claude Platform, Anthropic’s native product interface, directly through their existing AWS accounts, eliminating the need for separate credentials or billing.

      This marks an advancement beyond making Claude available via Amazon Bedrock, Amazon’s managed AI marketplace. Claude remains the sole frontier model accessible on all three leading cloud platforms: AWS Bedrock, Google Cloud Vertex AI, and Microsoft Azure Foundry.

      The geopolitical context of this deal is significant. Anthropic is currently restricted from Department of Defense contracts due to a supply-chain risk designation, which it is contesting in court.

      This situation exists alongside a broader AI policy landscape in which the Trump administration has been simultaneously urging AI firms to ease safety regulations while courting them for applications related to national security.

      Dario Amodei, Anthropic’s CEO, has been publicly navigating this tension. The commercial momentum generated by the AWS deal, along with the implicit endorsement of Anthropic’s scalable infrastructure, serves as a response to critics, including OpenAI, who recently suggested that Anthropic had erred by not securing adequate compute resources.

      In November 2025, Microsoft invested up to $5 billion in Anthropic as part of a separate agreement that mandated Anthropic to purchase $30 billion worth of Azure compute capacity. The AWS deal significantly surpasses this arrangement, firmly positioning Amazon as Anthropic’s main infrastructure partner.

      The initial investment in this round reflects Anthropic’s latest valuation of $380 billion. Venture capital firms have reportedly been proposing funding at valuations of $800 billion or more in anticipation of a potential IPO, although no such funding round has been announced.