Multiple AI services go down simultaneously in rare coordinated outage

A wave of service interruptions struck the cloud-based AI sector Thursday morning, affecting four of the industry's largest players—OpenAI, Anthropic, Google, and xAI—in an unusual period of overlapping downtime.

Timeline of disruptions#

The outages began when Anthropic disclosed elevated error rates across multiple Claude models at 9:23 am Eastern. The company pinpointed the root cause within 15 minutes and deployed a fix by 12:16 pm, though a second incident affecting Claude Sonnet 5 briefly occurred around noon.

OpenAI reported degraded performance affecting ChatGPT and Codex starting at 10:43 am, roughly an hour after Anthropic's initial alert. Engineers implemented a mitigation within 30 minutes, fully resolving the issue by 12:55 pm.

xAI's Grok continues to display error messages at time of writing, with the company acknowledging it is "working on restoring service as quickly as possible."

Scope and user impact#

The disruptions affected a broad spectrum of AI services. Anthropic's incident touched Claude Mythos 5.1, Claude Fable 5.1, Claude Opus 5, and Claude Sonnet 5—some of its most widely used models. OpenAI's outage impacted both its flagship conversational AI and its code generation tool.

Reports from DownDetector tracking Grok outages surged dramatically, climbing from fewer than 10 submissions just before 9 am to 1,365 by 9:45 am, before subsiding to 273 at time of publication.