ChatGPT faces worldwide outage, OpenAI says recovery in progress

ChatGPT users globally reported access issues on June 10, with error messages and disruptions lasting a few hours before services appeared to stabilise

OpenAI's ChatGPT experienced an unexpected global outage on Tuesday, June 10, leaving many users unable to access the service via its app or website. According to Downdetector, over 860 reports flooded in between 12:30 pm and 3:30 pm (IST), with users encountering technical difficulties.  ''We are observing elevated error rates and latency across ChatGPT and the API. Our engineers have identified the root cause and are working as fast as possible to fix the issue.'' OpenAI has issued a statement on social media site X.

 

The disruption triggered a flood of posts on X, with frustrated users sharing screenshots of error messages. Many encountered a "Network error occurred" response when attempting to interact with the chatbot. It is still uncertain whether the outage is universal or limited to certain regions. Some users even made fun of the situation, stating that the "AI robot" was "taking a nap", while others acknowledged how dependent they had grown on the platform for the simplest tasks.

 

 

We are observing elevated error rates and latency across ChatGPT and the API.  Our engineers have identified the root cause and are working as fast as possible to fix the issue.  For updates see our status page: https://t.co/oUGSSyltRU

— OpenAI (@OpenAI) June 10, 2025

ChatGPT has experienced several outages throughout 2025, highlighting how disruptions remain a recurring issue with widely used AI services. In January, a global outage left millions unable to access the platform, with most users encountering “Bad Gateway” errors. Further service interruptions followed in March, including two significant incidents that affected prompt generation and caused elevated error rates across multiple models.

 

Earlier in the year, CEO Sam Altman acknowledged that ChatGPT’s servers were under immense pressure, even “starting to heat up” due to usage spikes. The company has also pointed out how even seemingly minor user behaviour can contribute to system load, remarking that adding “please” and “thank you” to prompts, while polite, still consumes computing resources.

 

By 3:45 pm (IST), reports on Downdetector began declining, suggesting service restoration. Since then, user complaints appear to have subsided.

 

OpenAI has not made any public comment on this outage.