On Thursday morning, the AI models developed by Anthropic, OpenAI, and xAI experienced a series of unusual interruptions, resulting in downtime for their respective chatbots. SpaceX, the parent company of xAI, later confirmed that the issues affecting Grok stemmed from “a disruption in our Memphis data centre this morning.”
Unusual Coincidence or Common Cause?
Initially, it appeared that the disruptions might be interconnected, potentially due to a shared external service provider. However, neither OpenAI nor Anthropic referenced any external source when providing statements on the situation. SpaceX did not respond to requests for further comments but publicly expressed: “We also wish to apologise to our affected computing partners.” Notably, Anthropic and xAI had announced a “computing collaboration” with SpaceX back in May.
Kathleen Chaykowski, a spokesperson for OpenAI, explained that “a routing error occurring at approximately 7:43 AM (Pacific Time) on Thursday, September 3, caused ChatGPT and Codex to become unavailable for some users across all platforms.” By 8:17 AM on the same day, a solution had been successfully implemented, and the situation was being monitored closely.
Anthropic’s Response to the Incident
Anthropic opted not to comment on the incident but began notifying users of a “partial disruption” at 6:23 AM (Pacific Time) on the same day. This disruption involved “an increase in errors for requests directed at Claude Mythos 5.1, Claude Fable 5.1, and Claude Opus 5.” Shortly thereafter, the company stated that it had “identified the cause” and that a solution had been applied. The issue was marked as resolved by 9:16 AM (Pacific Time). Claude Sonnet 5 also appeared to experience similar issues briefly after 9 AM.
xAI reported interruptions to Grok across all its platforms and services beginning at 6:30 AM (Pacific Time), at which point it noted “investigating the interruption” on its service status page. The status page indicated, “Grok is experiencing issues. We are working to restore service as quickly as possible.” By 10:05 AM (Pacific Time), the incident was marked as resolved, with the company stating, “We have rectified the situation, and traffic is returning to normal.”
Other Services and the Impact of Interconnected Systems
There were also isolated reports of a potential service disruption affecting Google Gemini on Thursday morning; however, the company did not confirm any incidents and recorded none on its service status panel. Google did not respond to requests for comments prior to publication.
Typically, multiple outages occurring simultaneously within the same sector might indicate that a cloud service provider, content delivery network, or another external vendor is facing issues impacting several clients. However, neither OpenAI nor Anthropic indicated a potential common cause. Major players in the internet infrastructure sector, including Cloudflare, Amazon Web Services, and Microsoft Azure, did not report any service interruptions on Thursday.
Ultimately, it remains uncertain whether these disruptions were mere coincidences. However, as AI models increasingly share services and become interconnected, the likelihood of such events occurring may rise.