Three leading AI platforms, ChatGPT, Claude, and Grok, faced a collective service disruption lasting roughly three hours and 40 minutes. All systems were fully restored by 1:10 AM Beijing time on September 4th. Notably, Gemini reported no service interruptions throughout the incident.

Earlier reports from Downdetector user submissions and official status pages indicated ongoing outages across multiple AI providers. Anthropic acknowledged it was investigating errors affecting Claude and working on fixes, while OpenAI confirmed issues with both ChatGPT and Codex. Meanwhile, SpaceXAI noted that Grok was experiencing a failure and that the outage was under investigation.

According to Downdetector data, more than 12,000 user reports flagged problems with OpenAI, alongside approximately 1,200 reports for Claude and around 1,000 for Grok. While individual AI service outages are not uncommon, the simultaneous failure of three major platforms is highly unusual. Additionally, the disruption cascaded to the AI coding tool Cursor, which confirmed its own malfunction due to the downtime of Grok and Claude.

In contrast, Google's Gemini remained unaffected. Although user reports for Google services did increase during the same window, the company has not confirmed any service degradation.

The root cause may be linked to Microsoft Azure, which saw a sharp spike in reported faults as well. All three AI platforms rely on Azure for cloud computing services, though their computational architectures also incorporate other providers, including Google.