ChatGPT, Claude, Gemini and Grok went down together. What happened?

For a brief window on Thursday, four of the world's most used AI assistants stumbled at roughly the same time.

ChatGPT, Anthropic’s Claude, Gemini and xAI’s Grok all faced outages on September 3, 2026, at nearly the same time. The overlap immediately raised a larger question: did the AI industry just expose a shared weak link, or was this an unusual coincidence?

The answer, at least for now, is not fully clear. OpenAI and Anthropic did not publicly point to a common third-party provider, while xAI’s parent company SpaceX said Grok’s problem was linked to an outage at its Memphis compute centre.

What users saw during the outage

The incidents began during Thursday morning in the US, which meant late Thursday evening for many users in India.

OpenAI said a routing error that began around 7:43 am PT left some users unable to access ChatGPT and Codex across platforms. A fix was implemented by around 8:17 am PT, after which the company continued monitoring the service.

Anthropic’s issue started earlier. The company’s status updates described a partial outage from around 6:23 am PT, with elevated errors affecting requests to Claude Mythos 5.1, Claude Fable 5.1 and Claude Opus 5.

Anthropic later said it had identified the cause and deployed a fix, marking the issue resolved by 9:16 am PT.

xAI reported trouble with Grok from around 6:30 am PT across its platforms and services.

Whereas for Grok, its status page said the chatbot was experiencing issues and that xAI was working to restore service. By 10:05 am PT, the AI firm said the issue had been resolved, and traffic was healthy again.

Google's Gemini also experienced problems, with user reports showing disruptions during the same period. However, Google did not record a formal outage on its status page, making the exact nature of the Gemini incident less clear.

Why a shared cause was suspected

When several major online services fail at the same time, the first suspicion often falls on shared internet infrastructure. That could include a cloud provider, a content delivery network, or another vendor that quietly powers multiple platforms behind the scenes.

In this case, the timing looked suspicious. ChatGPT, Claude and Grok are competing products, but large AI systems often depend on overlapping layers of data centres, networking services and specialised computing capacity.

However, WIRED reported that major infrastructure players such as Cloudflare, Amazon Web Services and Microsoft Azure did not report outages on Thursday. That makes the picture more complicated.

OpenAI attributed its problem specifically to a routing error, while xAI pointed to its Memphis compute centre. Anthropic's latest explanation was an infrastructure issue.

What this says about AI reliability

The outage was short, but its impact was noticeable because AI tools are now part of daily work for developers, students, writers, customer support teams and businesses. Even a brief disruption can slow coding, research, documentation and internal workflows.

The episode also shows how dependent modern AI products are on complex infrastructure. A chatbot may look like a simple webpage or app, but behind it sits a large stack of servers, models, routing systems and compute partnerships.

For now, there is no confirmed single cause linking all four failures. What happened appears to be a rare cluster of outages, with separate public explanations and some unanswered questions.