What happened
OpenAI experienced a major system outage beginning November 19, 2025, causing ChatGPT and related services to freeze completely for many users by November 21. The outage persisted for over 48 hours, affecting multiple devices, browsers, and networks, despite the official status page showing no issues. The root cause was linked to a third-party service provider, and although a fix was implemented, the platform's stability remained compromised during the incident.
Why it matters
- The prolonged outage disrupted access to ChatGPT and other OpenAI services, impacting users globally.
- It exposed vulnerabilities in OpenAIâs server scaling and platform stability.
- The incident challenges OpenAIâs dominance and reliability in the AI service market.
Key facts
- 1. Minor disruptions began on November 19, 2025, escalating to a full freeze by November 21 around 14:00 UTC.
- 2. The outage lasted over 48 hours, affecting new and existing chats across all user devices and networks.
- 3. The official OpenAI status page inaccurately showed a âgreenâ status during the outage.
- 4. The incident was traced to issues with a third-party service provider.
- 5. A fix was deployed and services eventually recovered, with a detailed root cause analysis pending.
Key Voices
- Users reported widespread freezing of ChatGPT chats and called for prioritizing server stability over UI or filter changes.
- OpenAI acknowledged the incident was caused by a third-party service provider and confirmed service recovery after implementing a fix.
- The Help.OpenAI bot provided only generic responses and did not escalate issues during the outage.
- Not stated in the article: direct quotes from OpenAI executives or third-party providers.
