Amazon.com, Inc. (AMZN) is telling its engineers to stop relying on outside AI coding assistants and switch to Kiro, the company's homegrown alternative. It's a classic case of "eat your own dog food," except the stakes involve billions in cloud revenue and Amazon's reputation in the AI arms race.

The Internal Push for Kiro

An internal memo sent to Amazon's engineering teams makes the directive clear: stop adding new third-party AI development tools, according to a Monday report from Reuters. While employees can keep using whatever external tools they already have, the company won't be supporting new ones going forward.

The memo was signed by senior executives Peter DeSantis and Dave Treadwell, who oversee AWS and e-commerce operations. Their pitch frames the move as essential for improving Kiro through real-world testing. Employees were told their feedback would help the company "aggressively improve" the tool.

Amazon confirmed the memo's authenticity to the publication.

The Irony of Billion-Dollar Partnerships

Here's where it gets interesting: Amazon is discouraging use of tools like OpenAI's Codex, Anthropic's Claude Code, and Cursor, even though the company has invested billions in Anthropic and recently expanded a multiyear cloud partnership with OpenAI. It's not exactly a vote of confidence when you're bankrolling someone but won't let your own team use their products.

The directive reflects Amazon's broader effort to shake off the perception that it's trailing behind OpenAI and Alphabet Inc.'s (GOOGL) (GOOG) Google in the race to develop cutting-edge AI tools. When you're Amazon Web Services, being seen as a fast follower rather than a leader stings.

For context, Amazon reported third-quarter net sales of $180.2 billion in October, marking a 13% year-over-year increase and beating analyst expectations of $177.8 billion. The company remains in solid financial shape, with long-term momentum intact even as short- and mid-term indicators show some cooling.