Amazon has quietly reversed an internal policy that barred its software developers from using rival AI coding assistants, opting instead to embrace two of the most popular tools on the market. The change, announced to employees this week, will give Amazon's engineering teams access to Anthropic's Claude Code and OpenAI's Codex, both of which will run on Amazon's Bedrock platform.
The decision marks a sharp turnaround from November, when Amazon leaders sent a memo urging employees to use Kiro, the company's in-house code-generating tool, and stated that no additional third-party AI development tools would be supported. At the time, the memo, quoted by Reuters, said: “While we continue to support existing tools in use today, we do not plan to support additional third party, AI development tools.”
That stance drew criticism from within the company. Some developers complained that it was awkward to promote Claude Code to customers through AWS Bedrock while being barred from using it internally. One employee wrote in a comment obtained by Business Insider: “Customers will ask why they should trust or use a tool that we did not approve for internal use.”
The reversal also comes amid a broader industry race. Amazon has invested tens of billions of dollars in Anthropic and OpenAI, both of which are locked in a heated competition to dominate the rapidly expanding AI coding sector. Meanwhile, Kiro has struggled to keep pace, according to internal discussions.
In a note to staff obtained by Business Insider, Jim Haughwout, Amazon's vice president of software builder experience, announced that Claude Code would be made available immediately, with Codex to follow next week. “To help you invent more for customers, we are expanding the agentic AI tools available to you,” Haughwout wrote.
Despite the shift, Amazon is not fully abandoning Kiro. A spokesperson told Business Insider that teams are still “primarily using” Kiro, claiming that 83 percent of engineers at the company rely on it. The spokesperson added: “At Amazon, we’ve long held there’s no one-size-fits-all approach to how our teams innovate. Our builders are using Kiro for agentic coding, and now with both Claude Code and Codex running on AWS, we are making additional tools available as well.”
Why the Change Matters
The policy reversal is notable for several reasons. It signals that Amazon, despite its massive investments in AI, is willing to let its developers use tools that may outperform its own. It also highlights the growing influence of employee feedback in shaping corporate AI strategies, especially when internal restrictions conflict with customer-facing recommendations.
The move also comes at a time when AI tools have faced scrutiny over reliability. Earlier reports suggested that some AI tools had caused downtime at Amazon, though the company disputed those claims. By integrating rival tools through Bedrock, Amazon aims to provide secure access to frontier models while maintaining some control over the environment.
For developers, the change means greater flexibility and access to tools that are widely regarded as industry leaders. For Amazon, it represents a pragmatic acknowledgment that no single tool can meet all needs, even as it continues to push Kiro as a primary option.