OpenAI Confirms Global ChatGPT Outage Affecting Millions Worldwide
A nearly two-hour disruption knocked out ChatGPT, Codex, and APIs across multiple continents on July 25, 2026.
OpenAI confirmed a major worldwide outage on July 25, 2026, that knocked out ChatGPT's web platform, mobile applications, and developer APIs for nearly two hours, leaving millions of users unable to access the AI tools they depend on for work and daily tasks.
The disruption began around 5:00 AM ET (approximately 2:30 PM IST), with DownDetector recording significant spikes in user complaints starting at 4:47 AM ET. Reports flooded in from North America, Europe, Asia (particularly India and the UAE), Japan, and Australia.
What Went Wrong
Users globally experienced login failures, "too many concurrent requests" errors, and an inability to load chat histories or send prompts. The outage also affected OpenAI's Codex coding platform, indicating a failure in the underlying infrastructure rather than an isolated service issue.
In the UAE, DownDetector data showed 68% of reported issues were linked to core ChatGPT operations, 18% to the mobile app, and 9% to API and Codex services.
OpenAI acknowledged the problem on its official status page, stating: "We are investigating the issue for the listed services." According to official status page data, the incident lasted approximately 1 hour 50 minutes, from 09:17 AM UTC to 11:08 AM UTC, though some users reported disruptions lasting up to 5-7 hours depending on their region and when services were fully restored.
A Pattern of Instability
The July 25 outage occurred during a period of perceived instability for OpenAI's infrastructure. It followed another significant disruption on July 23, 2026, which also affected Codex and the API. Sources indicate the earlier incident lasted approximately 24 hours, though with fluctuating stability throughout the resolution period.
These consecutive failures have raised industry concerns regarding the resiliency of OpenAI's systems as they scale to support hundreds of millions of weekly users.
Why It Matters
The global nature of the outage highlights the critical dependency that millions of individuals and businesses have developed for AI tools in their daily workflows, including coding, writing, and customer service operations. Frequent high-profile downtimes challenge the reliability of AI as a production-grade utility and put pressure on OpenAI to improve its infrastructure stability.
As AI tools become embedded in professional workflows, the expectation for enterprise-grade reliability grows. For OpenAI, maintaining trust means ensuring that the infrastructure supporting hundreds of millions of users can withstand scale without catastrophic failures.