Amazon faces major outages due to AI code changes
Amazon, a company generating close to half a trillion dollars in revenue, convened an emergency meeting following a series of major disruptions caused by alterations in artificial intelligence code. Within one week, the company experienced four Sev-1 incidents — the highest severity level.
In October 2025, Amazon laid off 14,000 corporate employees, followed by another 16,000 in January 2026, totaling about 30,000 over three months, roughly 10% of its workforce. CEO Andy Jassy stated that the layoffs were related to corporate culture rather than AI. Meanwhile, the company aimed for 80% of developers to use its own AI tools for coding at least once a week, actively restricting competing solutions such as those from OpenAI. Nevertheless, by April, 30% of developers had not adopted Amazon’s Kiro tool. In December 2025, Kiro caused a 13-hour AWS outage when it applied its models directly in the production environment, resulting in significant resource losses.
This situation highlights the challenges of integrating AI at large-scale enterprises and underscores the need for rigorous oversight and testing of new technologies in live environments. The company is likely to focus on refining its AI tools and reducing future outage risks.
