In one of the most severe digital infrastructure failures of the decade, OpenAI’s flagship conversational chatbot, ChatGPT, suffered a massive global blackout on Thursday, locking out millions of professional, academic, and casual users worldwide.
The disruption, which began escalating rapidly during the evening hours, initially appeared to be an isolated failure for OpenAI. According to real-time telemetry from web monitoring platform Downdetector, user-submitted grievance reports skyrocketed past 30,000 within a single hour. Impacted individuals reported an absolute systemic breakdown, noting an inability to authenticate accounts, transmit textual prompts, or even access saved chat histories.
READ: Pentagon rolls out ChatGPT, Grok to millions of US military personnel (September 1, 2026)
However, what began as an OpenAI crisis quickly transformed into an unprecedented industry-wide emergency.
As displaced users scrambled toward alternative platforms to salvage their workdays, they discovered that the contagion had spread. Industry giants Anthropic (Claude), Google (Gemini), and xAI (Grok) simultaneously buckled under pressure.
UPDATE:
OpenAI acknowledged the issue and addressed it. The company initially said it was aware of the issue and had identified a possible fix. OpenAI was monitoring the platform to see whether the fix had worked, but later acknowledged that users were still facing degraded performance.
(This story has been updated with the latest developments.)