Unexpected Outage Sends Ripples Across AI Users Worldwide
On June 10, 2025, regular users of ChatGPT found themselves stuck, waiting for responses that never came. Starting around 2:45 PM Indian Standard Time, thousands of people from India, the US, and the UK faced either sluggish performance or outright failures just trying to chat with the now-ubiquitous AI assistant. The trouble peaked by 3 PM, with dashboards lighting up in red—reports flooded Downdetector: nearly 800 frustrated users in India, close to 1,100 in the US, and more than 1,450 in the UK, all seeking answers for the same puzzling errors.
People hoping for a quick fix instead received cryptic messages. 'Hmm…something seems to have gone wrong' appeared on countless screens, pushing users to refresh endlessly. Others ran smack into 'A network error occurred. Please check your connection and try again,' despite strong Wi-Fi and no other internet problems. "Is it just me, or is ChatGPT broken for everyone?" became a trending question online as confusion grew.
Problems Spread to Sora, API, and Beyond
The outage wasn’t limited to the chatbot. OpenAI’s Sora video creation tool—a favorite for designers and content creators—also froze up, alongside the GPT API that powers hundreds of apps and bots across the globe. OpenAI moved quickly to confirm there were 'elevated error rates,' but officials stayed vague on how soon systems would be back to normal. Their official status page quietly marked the incident while users grew more anxious.
Social media saw a firestorm of complaints and curiosity. Some users were baffled to see 'too many concurrent requests' messages with barely any recent activity. Others saw flashes of humor, like Richard Pathray’s screenshot paired with a witty jab at ChatGPT’s sudden shyness when being put to the test. Software engineer Gergely Orosz, who tracks developer-facing incidents, noticed this outage hit at the same time as hiccups on unrelated tech platforms like Heroku and NVIDIA’s developer docs. He speculated these could have a shared root cause—maybe some big cloud hiccup, or cascading issues with internet infrastructure backing up many web-based tools at once.
By the morning of June 11, the situation was only partly resolved. The API and Sora tools managed to right themselves, but ChatGPT limped out of the storm, still dishing out rare but stubborn errors. For businesses, students, and teams who depend on the chatbot for everything from homework help to brainstorming ideas, it felt like a stark reminder: even the smartest machines aren’t immune to breakdowns.
Throughout all this, OpenAI did share updates, but frustrated users wanted more concrete explanations. With so much of daily work—and even social life—relying on these AI tools, vague promises ring hollow. If anything, this incident pulled back the curtain on the vulnerability of digital infrastructure supporting the entire AI ecosystem, prompting many to ask how stable even the most trusted tools really are when it matters most.
