Amazon is one of Anthropic’s largest investors and a key partner in bringing the startup’s AI models and products to customers.
Inside Amazon, however, employees aren’t allowed to use Anthropic’s Claude Code for production work or for live products without formal approval.
That disconnect became more visible late last year, when Amazon rolled out internal guidance steering teams toward Kiro, its in-house AI coding assistant, according to internal messages seen by Business Insider. The guidance encourages using Kiro over non-approved third-party tools, including Claude Code, for production code.
The policy prompted criticism across internal Amazon forums. In one discussion thread, roughly 1,500 Amazon employees endorsed the formal adoption of Claude Code, according to the messages.
The restrictions have been particularly awkward for engineers who help sell AWS Bedrock, Amazon’s platform that offers customers access to third-party AI services, including…
