On June 4, 2024, ChatGPT went down in one of the most significant outages the platform had seen, leaving millions of users staring at error screens for several hours. The disruption hit all ChatGPT plans — Free, Plus, and Team — and underscored just how dependent the world has become on OpenAI's flagship AI chatbot.
What Happened During the ChatGPT Outage
The ChatGPT outage began around 0700 UTC on June 4. Within roughly 20 minutes, OpenAI acknowledged on its status page that something was wrong and began investigating. The company identified a database failure as the primary root cause — a server-side problem that prevented the chatbot from responding to queries on both the web and mobile apps.
By 0800 UTC OpenAI said it was "mitigating" the issue, and the service was fully restored by approximately 1700 UTC. The total window of disruption lasted close to ten hours from first detection to complete resolution, though many users regained access earlier as fixes were rolled out.
Who Was Affected
The outage impacted ChatGPT's web interface, its mobile application, and the newer Mac desktop app. ChatGPT Plus subscribers experienced the same disruptions as free-tier users, with the chatbot either refusing to respond or taking several minutes to generate even basic answers.
At the peak of the incident, outage-tracking site DownDetector recorded more than 3,000 user reports. OpenAI's own status page listed the problem as affecting "all users on all plans of ChatGPT." The platform at platform.openai.com and the core API were not listed as affected.
DALL-E 3 and API Services Also Hit
Beyond the main chat interface, users reported that DALL-E 3 image generation was also failing during the outage. OpenAI separately noted disruptions to "non-completion API endpoints," meaning some services built on top of the OpenAI API experienced degraded performance even though the primary API itself remained operational.
OpenAI's Response
OpenAI deployed a fix and restored service without issuing a detailed public post-mortem on the root cause. The company said it was closely monitoring the service to prevent similar problems in the future. Despite the scale of the disruption, OpenAI did not publicly disclose the full technical details of what triggered the database failure.
Not the First Time
This was not an isolated event. ChatGPT had previously suffered an outage that prevented roughly 10% of users from sending messages to the model. The June 4 incident was notably larger in scope, impacting all users simultaneously rather than a subset of the base.
With more than 100 million weekly active users, ChatGPT outages carry outsized consequences — not only for individual users but for businesses and developers who rely on the platform for production workflows. OpenAI has not announced structural changes to its infrastructure following the incident, though the company said it was committed to improving reliability.
