Amazon Confirms Anthropic's Claude Remains Available on AWS Outside Defense Work

Summary

Amazon Web Services announced Friday that it will continue offering Anthropic's Claude AI models to its cloud customers, with the exception of Department of Defense-related work. The decision follows the Pentagon's designation of Anthropic as a "supply chain risk" on Thursday, after the AI startup refused to grant the DOD unlimited operational control over its technology. Amazon joins Microsoft and Google in clarifying that Claude remains accessible for non-defense applications.

The supply chain risk label requires defense contractors to certify they aren't using Anthropic's models in Pentagon work, prompting AWS to support affected customers in transitioning to alternative AI solutions. Amazon stated that "AWS customers and partners can continue to use Claude for all their workloads not associated with the Department of War," while offering transition assistance for defense-related workloads.

Amazon has invested $8 billion in Anthropic since 2023 and maintains a deep commercial partnership with the AI startup. AWS serves as Anthropic's primary cloud and training partner, with Anthropic committed to using 500,000 of Amazon's custom Trainium 2 chips as part of an $11 billion data center campus called Project Rainier. Claude models are distributed through AWS Bedrock, Amazon's platform for accessing multiple AI providers, and are available in AWS GovCloud for handling sensitive government data outside defense applications.

• The decision aligns AWS with Microsoft and Google, which made similar announcements about Claude's continued availability outside defense contexts