Amazon Refutes Financial Times Report on AI-Caused Service Outage

Amazon clarifies the timeline and cause of its recent outage, clearing its AI bot of any wrongdoing.

On February 19, 2026, the Financial Times published a report that Amazon Web Services (AWS), the backbone of much of the modern internet, had suffered outages caused by its own AI coding tools — including an “agentic” assistant called Kiro that allegedly decided the best way to fix a problem was to delete an entire environment.

The story quickly went viral. It tapped into deep-seated fears about handing over critical infrastructure to autonomous AI. Yet less than 24 hours later, Amazon fired back with an unusually detailed public correction, insisting the entire narrative was wrong — and that the real culprit was classic human error, not rogue AI.

What really happened?

The Financial Times Report

According to the Financial Times, citing four people familiar with the matter (including current Amazon employees), AWS experienced a 13-hour interruption in mid-December 2025 to AWS Cost Explorer — the customer-facing dashboard that helps users visualize, understand, and manage their cloud spending.