TLDR: OpenAI successfully resolved a global outage affecting its ChatGPT AI services on September 3, 2025. The disruption, which began around 4 a.m. EST, prevented users worldwide from receiving responses to prompts, impacting various functionalities including web, mobile, and image generation. OpenAI identified a “Conversations” issue as the root cause and restored full service within approximately two hours.
OpenAI’s widely used artificial intelligence chatbot, ChatGPT, experienced a significant global outage on Wednesday, September 3, 2025, causing widespread disruption for its vast user base. The incident, which began around 4 a.m. EST (9 a.m. BST), saw users across the globe, including in major markets like India, the United States, and the United Kingdom, unable to receive responses to their prompts.
According to reports from Downdetector, a website that tracks online service statuses, hundreds to thousands of users reported issues. In India alone, over 500 users flagged problems, with 85 percent of all outage reports specifically citing issues with ChatGPT itself, while 13 percent reported problems with the broader OpenAI website. The outage affected a comprehensive list of 23 components of the ChatGPT service, including core functionalities such as login, web and desktop access, mobile applications (both iOS and Android), WhatsApp integration, search capabilities, file uploads, and image generation. Notably, other OpenAI products like the API, Sora, and Playground remained unaffected during this period.
OpenAI swiftly acknowledged the problem on its official status page, initially stating, “We are investigating the issue for the listed services.” Approximately 50 minutes after the initial reports, around 4:50 a.m. EST (9:50 a.m. BST), the company identified the root cause, attributing it to a problem with “Conversations.” OpenAI then confirmed they were actively working on a resolution. Within roughly two hours of the initial disruption, by 3:29 a.m. EST (8:29 a.m. BST), OpenAI announced that ChatGPT was “fully operational” again, with Downdetector reports significantly decreasing.
This recent outage is not an isolated incident for the popular AI chatbot, which reportedly serves around 700 million active users weekly. ChatGPT has experienced several service interruptions in the past, including a similar brief outage in August, a more prolonged 10-hour global outage in June, and other significant disruptions on January 23, 2025, December 26, 2024, and February 5, 2025. The recurring nature of these outages highlights the challenges in maintaining continuous service for complex AI infrastructure.
Also Read:
- ChatGPT’s Dominance Solidifies with Over 81% Generative AI Market Share in July 2025
- ChatGPT User Base Exceeds 20 Million in South Korea, Marking Significant AI Adoption
During the downtime, users took to social media platforms, particularly X, to voice their frustration and share memes, underscoring the world’s increasing reliance on AI tools for daily tasks and creative endeavors. While the service has been restored, these incidents serve as a reminder of the critical importance of robust infrastructure and rapid response mechanisms for essential digital services.
