Amazon Denies AI Code Caused Major Shopping Outages Despite Internal Documents Suggesting Otherwise

Amazon Pushes Back on AI Blame

Amazon issued a sharp rebuttal this week to reports that AI-generated code was behind a string of crippling outages on its retail platform. The company called the claims "false," insisting that only one recent incident involved AI-assisted tooling in any way, and that the root cause was an engineer who followed inaccurate advice that an AI agent pulled from an outdated internal wiki.

The Scale of Disruption

The denial comes after a turbulent stretch for Amazon's ecommerce operations. A March second incident tied to Amazon's AI coding assistant resulted in roughly one hundred and twenty thousand lost orders and one point six million website errors. Days later, a separate outage caused a ninety-nine percent drop in orders across North American marketplaces, wiping out an estimated six point three million orders over nearly six hours.

Internal Tensions Surface

Internal documents paint a more nuanced picture than Amazon's public stance suggests. A briefing note described a "trend of incidents" with "high blast radius," partly attributed to "Gen-AI assisted changes." Amazon's senior vice president of eCommerce Foundation acknowledged that "the availability of the site and related infrastructure has not been good recently."

Industry-Wide Concerns

The incidents have reignited debate about AI-generated code in production environments. A new industry report found that among developers using AI coding tools multiple times daily, fifty-one percent report more code quality problems and fifty-three percent report more security incidents. Amazon has quietly implemented a ninety-day code safety reset, introducing additional review requirements and controlled friction for changes to critical retail infrastructure.