According to press reports from Axios and Wireline, Beijing time, between the night of 3 September and the early hours of 4 September ChatGPT With Codex, Anthropic Claude And under xAI Grok Services were disrupted almost at the same time, causing a large number of user barriers around the globe。
According to the web-based monitoring platform DowndetectorFaultMore than 37,000 high-end problem reports were received for OpenAI, and Claude and Grok received thousands of feedback on failures, respectively. As a result of the failure of the upstream basic model, the AI code editor Cursor also experienced automation and cloud downgrading during the same period。
Companies subsequently updated their status pages and disclosed the progress of the repairs. OpenAI spokesmen confirmed to the media that the anomalies in ChatGPT and Codex were mainly caused by the error of route configuration and that the team had returned to normal after the deployment of the rehabilitation programme
Anthropic claims that multiple Claude models such as Opus 4.8 and Opus 5 have been repaired and returned to service; the xAI status page shows that the Grok service has been restored to its health, and the reports indicate that the malfunction originated from its anomaly at the Memphis computing centre。
TO DATE, ALL OF THE ABOVE-MENTIONED AFFECTED AI SERVICES AND RELATED DOWNSTREAM DEVELOPMENT TOOLS HAVE RETURNED TO NORMAL. THIS TIME, THE MULTIPLE HEAD AI MODEL HAS RARELY BEEN ABLE TO FOCUS ON SERVICE INTERRUPTIONS AND HAS ALSO GIVEN RISE TO FURTHER INDUSTRY ATTENTION TO THE CURRENT GENERATION OF AI BOTTOM COMPUTING NETWORKS AND INFRASTRUCTURE REDUNDANCY。