According to GitHub and Hacker News, the incident occurred at around 12:20 on September 10 (US time). A few minutes later, Anthropic issued an official notice confirming the shutdown. The company said it has deployed some versions of the bug and is monitoring the results.
We saw our API had a brief interruption at 9:30 a.m. Pacific time. The service has been quickly restored, an Anthropic representative told TechCrunch.
This is not the first time Anthropic has encountered a technical problem. Over the past few months, the company has continuously recorded incidents related to Claude and other AI models.
Incidents like this show the growing dependence of programming on AI platforms. Just a few dozen minutes of service outage is enough to cause many disruptions to work. For a company that is valued at tens of billions of USD like Anthropic, ensuring the stability of the system is not only a technical responsibility, but also directly related to the trust of the global user community.
