No One Explains the Reasons Behind Today’s Outages at OpenAI and Anthropic

On Thursday morning, the frontier models from Anthropic, OpenAI, and xAI faced unusual outages, resulting in downtime for their respective AI chatbots. SpaceX, the parent company of xAI, explained on Thursday afternoon that the issues with Grok were due to “an outage at our Memphis compute center this morning.”

These outages initially seemed connected, likely due to a shared third-party service provider, but neither OpenAI nor Anthropic mentioned any external factors in their comments to WIRED on Thursday. SpaceX, which did not respond to WIRED’s request for further comments, expressed publicly: “We’d also like to apologize to our impacted compute partners.” In May, Anthropic and xAI formed a “compute partnership” with SpaceX.

OpenAI spokesperson Kathleen Chaykowski informed WIRED: “A routing error starting at approximately 7:43 am PT on Thursday, September 3, caused ChatGPT and Codex to be unavailable for some users across platforms. By around 8:17 am PT on Thursday, an effective solution was put in place and continues to be monitored.”

Anthropic chose not to comment on the incident. The company started notifying users about a “partial outage” at 6:23 am PT on Thursday, indicating “elevated errors on requests to Claude Mythos 5.1, Claude Fable 5.1, and Claude Opus 5.” Shortly thereafter, the company confirmed it had “identified the cause” and that “a fix has been deployed.” The matter was declared resolved by 9:16 am PT, although Claude Sonnet 5 encountered similar issues briefly after 9 am PT.

At 6:30 am PT, xAI acknowledged Grok outages across all platforms and services, posting “investigating outage” on its service status page. The notice indicated, “Grok is experiencing issues. We are working to restore service as quickly as possible.” By 10:05 am PT, the incident was marked as resolved, with the company stating, “We have resolved the situation, and traffic is healthy again.”

There were also scattered reports of a possible Google Gemini outage on Thursday morning, although the company did not validate these claims or record any incidents on its service status dashboard. Google did not provide a comment to WIRED prior to publication.

Generally, simultaneous outages in a sector would suggest a third-party vendor, such as a cloud provider or content delivery network, is experiencing issues affecting multiple clients. However, OpenAI and Anthropic did not identify a common source, and significant players in the internet infrastructure sector—including Cloudflare, Amazon Web Services, and Microsoft Azure—did not report any outages on Thursday.