OpenAI, Anthropic and xAI hit by rare simultaneous outages
Frontier AI services experienced concurrent downtime on Thursday morning, but the absence of reported issues from major infrastructure providers leaves the root cause unclear.
Rare simultaneous outages affected the frontier artificial intelligence models of OpenAI, Anthropic and xAI on Thursday morning, causing downtime for ChatGPT, Claude and Grok. The concurrent nature of the failures raised questions about a potential shared third-party cause, yet the specific reason for the overlap remains murky.
OpenAI spokesperson Kathleen Chaykowski attributed the issue to a routing error that began around 7:43 am PT. The error made ChatGPT and Codex unavailable for some users across platforms. A solution was successfully implemented by 8:17 am PT, with the company stating the fix was continuing to be monitored.
Anthropic reported a “partial outage” starting at 6:23 am PT, citing elevated errors on requests to Claude Mythos 5.1, Claude Fable 5.1 and Claude Opus 5. The company indicated it had identified the cause and deployed a fix, marking the issue as resolved by 9:16 am PT. Claude Sonnet 5 also appeared to experience similar issues briefly shortly after 9 am PT.
Meanwhile, xAI reported Grok outages across all platforms beginning at 6:30 am PT. The company’s service status page stated that Grok was experiencing issues and that the team was working to restore service as quickly as possible. The episode was marked complete at 10:05 am PT, with traffic described as healthy again.
SpaceX, the parent company of xAI, stated that the Grok issues resulted from an outage at its Memphis compute centre. The company apologised to impacted compute partners, a notable reference given that Anthropic and xAI announced a “compute partnership” with SpaceX in May.
Typically, multiple outages in the same sector at the same time point to a cloud provider, content delivery network or other third-party vendor having issues affecting multiple customers. However, major infrastructure providers, including Cloudflare, Amazon Web Services and Microsoft Azure, did not report outages on Thursday. Neither OpenAI nor Anthropic cited an external source in their comments, leaving the link between the distinct incidents unclear.
