Story Content



On April 20, 2026, ChatGPT crashed in various countries, with users looking at blank screens, infinite loading indicators, or error messages. At its activity peak, the number of requests to Downdetector increased to over 13,000, which is an indication of a serious issue on the side of OpenAI. 

The failure caused a multitude of issues, including blank pages, timeouts, and issues to log in. ChatGPT was being opened by some users who were unable to send or receive any messages, and voice mode, image generation, and Codex, the coding tool provided by OpenAI, also ceased functionality. 

It was reported that thousands of users were unable to access ChatGPT, Codex, or related services on the web and mobile platforms in the United States, United Kingdom, and India. In the UK, the number of reports totaled over 8,000, and in the United States, it was approximately 1,875. 

Mitigation was implemented at approximately 1 PM ET and appeared to fix all the issues in the next half-hour. The outage seemed to have an impact on the accessibility of the projects by the users, and in other situations, work that was already in progress was lost. 

Notably, other AI services such as Claude, Google Gemini, Microsoft Copilot, and Meta AI continued to operate as usual during the ChatGPT outage and were reportedly overwhelmed by a huge influx of so-called refugee traffic due to displaced users. 

The current outage is a reminder that there is an increasing issue with the tech industry, and that is our dependence on several AI platforms. When something goes amiss, the effect is experienced across the board and real-time.