OpenAI’s ChatGPT Faces Ongoing Outages, Users Frustrated

San Francisco, CA — OpenAI‘s ChatGPT experienced another significant service outage today, marking the latest in a series of disruptions that have frustrated users throughout the week. The issues, which began around 9 a.m. ET, were reported widely across the United States and the United Kingdom, suggesting that the problems stem from high user demand during peak hours.

According to reports collected by Downdetector, users encountered various error messages, including “something went wrong” on the desktop version and “upstream connect error” on the mobile app. OpenAI’s status page confirmed there were “increased error rates” and “degraded performance,” stating, “we are investigating the issue.”

OpenAI CEO Sam Altman addressed the situation on X (formerly Twitter) yesterday, noting, “we are getting things under control, but you should expect new releases from OpenAI to be delayed, stuff to break, and for service to sometimes be slow as we deal with capacity challenges.” Altman’s comments reflect the increasing pressures on the service as user interaction spikes.

The outage came on the heels of growing excitement surrounding ChatGPT’s new features, including a widely popular image generator that has driven record user engagement. The surge has led to capacity issues, requiring OpenAI to rethink its current infrastructure. “Biblical demand, I have never seen anything like it,” Altman reportedly stated, referencing the overwhelming response to the new capabilities.

As of this afternoon, the number of outage reports has decreased significantly, suggesting that some users may have begun to regain access to the platform. At one point, however, reports surged again, causing users to panic amid uncertainty about service reliability. “Reports seemed to peak at around 2,200 during the height of the outages,” remarked a source familiar with the situation.

Despite the recent troubles, OpenAI’s status page now indicates that the service is expected to recover continuously. By late morning, reports confirmed reduced error messages across the board. A recent update at 11:12 a.m. ET stated, “all impacted services have now fully recovered