Major AI Instability Leaves Users Without Access

On Thursday morning (September 3, 2026), major artificial intelligence tools experienced simultaneous failures, causing an outage that affected users in Brazil and various regions worldwide.

What Happened on the AI Platforms

According to Downdetector's monitoring, the first complaints emerged around 10:30 AM (Brasília time) for xAI's Grok and Anthropic's Claude. Around 12 PM, the problem extended to Google's Gemini and OpenAI's ChatGPT, affecting web browsers, mobile apps, and desktop programs.

How the Companies Responded

SpaceXAI confirmed that Grok faced server instability. Claude's status page indicated resolution around 1 PM, while ChatGPT's registered normalization at 1:55 PM and Grok's at 2:07 PM. No official statement detailed the technical causes of the outages on the other platforms.

Possible Motivations for the Timing

The reports emerged amidst rumors of an OpenAI announcement. Industry sources suggest that the company might disclose information today about its next language model, named Astra, which would represent an evolution from GPT-5.6 to GPT-6. The timing of the failures fueled speculation about infrastructure preparation for the launch.

Impact on Users

During the period of instability, users reported unavailability across web versions, mobile applications, and desktop clients for all four tools. The peak of complaints registered for ChatGPT exceeded 9,900 reports on the monitoring platform, indicating widespread repercussions among AI consumers.

Next Steps and Expectations

With the platforms now normalized, attention turns to OpenAI's scheduled announcements and recent competing releases, such as Claude Fable 5.1, Mythos 5.1, and Gemini 3.8. While companies have not clarified the technical details of the failures, users await the full resumption of services and potential new language model developments.

With information from UOL, Poder360.