ChatGPT experienced a widespread outage on Monday, with thousands of users across multiple regions reporting access issues and degraded performance.
Quick Summary – TLDR:
- OpenAI confirmed degraded performance across ChatGPT and Codex, affecting login, conversations, and voice mode on Monday.
- Reports on Downdetector surged past 5,000 users within hours, with numbers climbing steadily from under 1,000 to over 5,000 in a short window.
- The outage affected multiple regions globally, with the UK recording over 8,000 reports on Downdetector compared to around 1,875 in the US.
- ChatGPT Business users were also impacted, with OpenAI noting issues after account upgrades or addition of new seats.
What Happened?
OpenAI’s ChatGPT went down for a significant number of users on Monday, with disruptions first registering around 10:05am ET. Downdetector, which tracks outages by aggregating status reports from multiple sources, began recording a sharp spike in complaints almost immediately after that window.
The outage was not limited to casual users. OpenAI’s own status page acknowledged the problem, confirming that ChatGPT and Codex were both experiencing degraded performance. The company stated it was actively investigating the issue, though no cause was identified publicly at the time of reporting.
ALERTE 🚨 Ongoing outage in France on ChatGPT everything is down. pic.twitter.com/mCHfekG9F1
— Cold Structure (@ColdStructure) April 20, 2026
Reports Climbed Rapidly as the Outage Widened
What made this outage particularly notable was the speed at which user complaints escalated. According to Downdetector data, reports went from under 1,000 to above 1,500 within minutes of the first spike. By 8:01am PT, nearly 4,000 users had filed reports. That number crossed 5,000 by 8:14am PT, with the platform status checker confirming users were unable to load both ChatGPT and Codex.
The majority of complaints focused on the core ChatGPT platform itself, though issues spanned nearly every functional area including login failures, broken conversations, and a non-responsive voice mode.
Implications for the Future of AI Platform Reliability
Monday’s outage arrives at a moment when AI tools are no longer optional accessories for many businesses. They are embedded into daily workflows, customer service pipelines, and coding environments. An hour of downtime is no longer just an inconvenience; for some teams it represents real operational disruption.
OpenAI’s infrastructure is under sustained pressure as ChatGPT’s user base continues to grow. The platform reportedly serves hundreds of millions of users, and scaling that infrastructure without reliability tradeoffs is a genuinely difficult engineering challenge. Incidents like this one may accelerate enterprise demand for service level agreements with guaranteed uptime clauses.
SQ Magazine’s Takeaway
In my experience covering tech outages, the speed at which this one escalated is what stands out most. Going from under 1,000 reports to over 5,000 in roughly 30 minutes is a steep curve, and it signals just how many people now depend on ChatGPT as a daily work tool rather than a novelty.
I think the regional imbalance is worth watching closely. The UK seeing more than 4 times the US report volume at peak raises real questions about how OpenAI distributes its infrastructure load across regions, and whether European users are consistently getting a lower tier of service reliability.
What to watch next is how OpenAI communicates post-incident. A transparent root cause analysis would go a long way toward reassuring the enterprise customers it is actively courting. Without that, every outage chips quietly at the trust it is working hard to build.
