Frontier AI chatbots from Anthropic, OpenAI and xAI all suffered rare, roughly simultaneous outages on Thursday morning, temporarily knocking out services including ChatGPT, Codex, Claude and Grok. The timing initially suggested a shared underlying cause, such as a common cloud or infrastructure provider, but the companies gave differing explanations and none pointed to a common external culprit, leaving the coincidence largely unexplained.

OpenAI attributed its disruption to a routing error beginning around 7:43am PT, resolved by roughly 8:17am PT. Anthropic reported a partial outage from 6:23am PT affecting Claude Mythos 5.1, Claude Fable 5.1 and Claude Opus 5, deploying a fix and marking it resolved by 9:16am PT, though Claude Sonnet 5 briefly showed similar issues shortly after. xAI's Grok was down across all platforms from around 6:30am PT until 10:05am PT, with SpaceX later citing an outage at its Memphis compute centre and apologising to "impacted compute partners" — notable given Anthropic and xAI's compute partnership with SpaceX announced in May. Major infrastructure providers such as Cloudflare, AWS and Microsoft Azure reported no issues, and Google did not confirm reports of a possible Gemini outage.