Amazon Web Services (AWS) faced a significant disruption in December 2025, reportedly linked to its AI system, Kiro, according to a report from the Financial Times. The incident lasted for 13 hours and was described as an outage affecting “one of our two Regions in Mainland China,” as stated by an Amazon spokesperson to Mashable. However, multiple sources cited by the Financial Times indicated that the disruption was triggered when engineers permitted Kiro to perform certain tasks, leading the AI to “delete and recreate the environment.”
The scale of this event was not comparable to the major AWS outage in October 2025, which had widespread repercussions. In response to the Financial Times reporting, AWS published a blog post aimed at correcting what it deemed inaccuracies. An AWS spokesperson asserted that the December incident was a “brief event” attributed to “user error,” distancing the company from any direct blame on the AI system. They emphasized that the engineers’ actions, rather than Kiro itself, were responsible for the disruption. Furthermore, they clarified that the issues did not affect major infrastructural services as the October outage had.
In a detailed email to Mashable, an Amazon spokesperson explained that the incident was confined to the AWS Cost Explorer service, which assists customers in managing AWS costs and usage. This limited in-scope interruption did not impact other critical services—such as compute, storage, database, or AI technologies—and the company reported no customer complaints associated with the event. According to the spokesperson, “In both instances referenced, the root cause was user error — specifically misconfigured access controls — not AI error.” Kiro’s framework requires user configuration for the actions it can undertake, and the system is designed to request authorization before proceeding with any tasks. Following the disruption, AWS implemented several safeguards, including mandatory peer reviews for production access and enhanced training on AI-assisted troubleshooting.
Disruptions in service have become a recurring theme across the internet, with recent outages affecting platforms such as YouTube, Verizon, Cloudflare, Microsoft 365, Google Cloud Platform, Microsoft Azure, and TikTok. Experts remain divided on whether such outages are becoming more frequent. However, the increasing reliance on a limited number of cloud service providers, including AWS, underscores the potential for a single outage to have cascading effects throughout the internet.
The incident raises important questions about the integration of AI into critical infrastructure and the associated risks. As companies worldwide continue to adopt AI technologies across various operational facets, stakeholders must consider the implications of AI-driven decision-making processes. While AWS has moved to distance itself from the notion that its internal AI can contribute to infrastructural failures, the situation serves as a cautionary tale about the complexities of human and machine collaboration in high-stakes environments.
UPDATE: Feb. 20, 2026, 8:24 p.m. EST This story has been updated with additional statements from Amazon Web Services, clarifying the affected region of the disruption.
UPDATE: Feb. 20, 2026, 12:36 p.m. EST The article has been revised to clarify that Amazon has attributed the outages to human error, not AI.
See also
Sarvam Launches Indus AI Chatbot with 105B Model and Multilingual Support in India
Isomorphic Labs Reveals IsoDDE, a Major Advance in AI Drug Discovery
AI Transforms Health Care Workflows, Elevating Patient Care and Outcomes
Tamil Nadu’s Anbil Mahesh Seeks Exemption for In-Service Teachers from TET Requirements
Top AI Note-Taking Apps of 2026: Boost Productivity with 95% Accurate Transcriptions
