ChatGPT hit by global outage, users in India among the worst affected

OpenAI's ChatGPT went down for users across the world on Saturday afternoon, with the outage beginning around 2.30 pm IST and disrupting the chatbot's web platform, mobile apps, voice features, and developer API simultaneously.

ChatGPT not working? Global outage hits millions of users

Complaints started flooding outage-tracking site Downdetector from roughly 2.15–2.20 pm IST, as users found themselves unable to send prompts, load their chat history, or start new conversations. Many reported being stuck in endless loading loops, while others ran into repeated 'Application Error' and 'too many concurrent requests' messages. The sidebar containing saved chats reportedly failed to load for a large number of users, leaving them staring at empty windows.

Simple conversation wasn't the only source of trouble. Failures were also reported by users of OpenAI's coding tool, Codex, and developers working on its API, indicating that the problem was not just with the consumer-facing app but rather with the company's underlying infrastructure.

Heatmaps from Downdetector confirmed that this was not a regional anomaly by displaying clusters of complaints dispersed over North America, Europe, and Asia. The number of complaints in India surpassed 1,000 and continued to rise throughout the afternoon. The majority of these complaints were related to ChatGPT's general performance rather than the mobile app in particular. Around 2:55 PM IST, the UAE had a rise in reports as well, with most of the problems again focusing on the main ChatGPT service rather than the app or API separately. Users in Japan and Australia also chimed in on social media with similar complaints.

OpenAI's response

The status page reportedly flipped back and forth between showing an active incident and normal operations while engineers worked on a fix. OpenAI had not given users a firm timeline for full restoration.

Not the first rocky week

In just a few days, ChatGPT has experienced two significant disruptions. Two days prior, on July 23, OpenAI struggled with a different outage that affected Codex, ChatGPT, and its API. It took nearly a full day to fully resolve. When combined, the consecutive occurrences have raised concerns about OpenAI's infrastructure's resiliency as it grows to support hundreds of millions of users each week.

As is now customary during any major AI outage, frustrated users flocked to social media to vent — and joke. Several pointed out how dependent their daily workflows, from coding to writing, have become on the chatbot being available. Some later reported service returning to normal in patches through the evening, though experiences appeared to vary by region and platform.

For more updates, join/follow our WhatsApp, Telegram and YouTube channels.