
OpenAI has suffered another major outage that has left thousands of users in the US and UK unable to use ChatGPT. The company has blamed the outage on an “increased level of errors” in its API.
According to Downdetector, more than 6,500 users in the UK reported the service was down at 1:47pm, the peak of the issue. More than 4,200 people in the US were experiencing similar issues, with 91% of the complaints specifically related to ChatGPT. The main issues were CAPTCHA failures and “Bad Gateway” errors;
OpenAI confirmed that the API experienced an outage between 3:33am and 4:23am PST due to “increased workload.” The company did not provide exact technical details, but experts speculate that a sudden spike in traffic, an internal server conflict, or a failed system update could have been the cause. This is not the first time OpenAI has faced a technical outage: in December 2024, ChatGPT, along with other services including the text-to-video tool Sora, was down for more than four hours due to server overload. Similar incidents in the past show that OpenAI is working to scale its resources to meet global demand.
ChatGPT’s large-scale outages highlight the importance of stability in AI operations. OpenAI should invest more in server optimization and load management to avoid similar incidents in the future.