ChatGPT went down for a second time in two days on February 4, leaving thousands of users unable to access OpenAI's popular AI chatbot. Starting around 9:30 a.m. Pacific, ChatGPT began suffering an outage, with reports quickly hitting over 10,000 on Down Detector. The service has since recovered—OpenAI designated the main outage as resolved at precisely 5:14 PM ET—but it marks the second major disruption in as many days.
The issues started just 24 hours after a major ChatGPT outage on February 3 affected over 12,000 users, with reports surging from a handful to over 12,000 within thirty minutes. During both outages, affected users encountered persistent "Hmm... something seems to have gone wrong" error messages. Users couldn't access conversations, send messages, or use image generation features—essentially, the entire service was affected.
OpenAI hasn't officially confirmed what's causing the repeated issues, but timing suggests infrastructure stress. On Monday, OpenAI released a native Codex app for macOS, which CEO Sam Altman reported received over 200,000 downloads on its first day. Some analysts suggest the surge in new users and API calls from these integrations may have strained the existing infrastructure. The good news: OpenAI stated "We have applied the mitigation and are monitoring the recovering", meaning engineers are actively watching the system to prevent another crash.
If you're using ChatGPT for work or study, keep an eye on OpenAI's status page for the latest updates. For now, the service appears to be back to normal.