Was there one shared cause behind the ChatGPT, Claude and Grok outages?

No shared cause was confirmed. The Register reported no relevant problems at AWS, Google Cloud or Microsoft Azure, and Cloudflare said it was seeing no significant service disruption. 9to5Google noted that all three services lean on Azure and raised it as a possible link, but Microsoft denied responsibility. Each company gave its own separate explanation.

How long were ChatGPT, Claude and Grok unavailable on September 3?

ChatGPT's routing error started at 7:43 a.m. PT and a fix was in place by 8:17 a.m. PT. Anthropic logged a partial Claude outage lasting three hours and six minutes, with service restored at 16:16 UTC. Grok's operator began investigating at 6:30 a.m. PT. 9to5Google reported everything back to normal by 12:38 p.m. PT.

Did the Claude outage hit the API and Claude Code as well as the chat app?

Yes. The Register listed claude.ai, Claude Code, Claude Cowork and the Claude API among the affected services, with elevated error rates across Sonnet 5 and other models. 9to5Google reported that most Claude models recovered first while Opus 4.8 and Opus 5 lagged behind, so agent and API traffic was hit alongside the consumer app.

Which large AI assistant stayed online during the September 3 outages?

Google Gemini kept working. 9to5Google reported that Gemini saw its own spike in user reports at the same time, most likely from people arriving after the other services failed, but the service itself was not affected. That made Gemini the one major assistant a team could fall back to during the window.

OpenAI, Anthropic, xAI · 2026-09-03 · major

ChatGPT, Claude and Grok go down together — three faults, one morning

ChatGPT, Claude and Grok all broke on the morning of September 3, 2026. OpenAI blamed a routing error, Anthropic an infrastructure issue, and Grok's operator a failure at its Memphis compute center. Google Gemini stayed up.

Illustration for The Register's report on the simultaneous ChatGPT, Claude and Grok outages — The Register

Three of the biggest AI services failed within hours of each other on September 3, 2026, for three separate reasons.

Quick facts

Date	September 3, 2026
Affected	ChatGPT, Codex, Claude, Claude Code, Claude API, Grok
ChatGPT	Routing error, 7:43–8:17 a.m. PT
Claude	3 hours 6 minutes, restored 16:16 UTC
Grok	Fault at the Memphis compute center
Unaffected	Google Gemini
Shared cause	None confirmed

What is it?

ChatGPT, Claude and Grok all stopped working on the morning of September 3, 2026, within a few hours of each other. OpenAI's outage covered ChatGPT and Codex. Anthropic's covered claude.ai, Claude Code, Claude Cowork and the Claude API. Grok went down on its own timeline. Google Gemini kept running throughout.

How does it work?

Each company reported a different fault. A routing error made ChatGPT unavailable for some users from 7:43 a.m. PT until a fix landed at 8:17 a.m. PT. Anthropic described an infrastructure issue that caused a partial outage, with elevated error rates across Sonnet 5 and other models. The company behind Grok apologised for a failure at its Memphis compute center that also reached its compute partners.

Why does it matter?

The timing, not the individual faults, is what this outage exposed. Downdetector spikes hit three vendors at once, so teams that had deliberately spread work across providers found their fallback was down too. The Register found no matching incident at AWS, Google Cloud or Azure and Cloudflare reported no significant disruption, so a single shared dependency was never established — the coincidence is the uncomfortable part.

Who is it for?

teams that depend on hosted AI APIs

Frequently asked questions

Was there one shared cause behind the ChatGPT, Claude and Grok outages?: No shared cause was confirmed. The Register reported no relevant problems at AWS, Google Cloud or Microsoft Azure, and Cloudflare said it was seeing no significant service disruption. 9to5Google noted that all three services lean on Azure and raised it as a possible link, but Microsoft denied responsibility. Each company gave its own separate explanation.
How long were ChatGPT, Claude and Grok unavailable on September 3?: ChatGPT's routing error started at 7:43 a.m. PT and a fix was in place by 8:17 a.m. PT. Anthropic logged a partial Claude outage lasting three hours and six minutes, with service restored at 16:16 UTC. Grok's operator began investigating at 6:30 a.m. PT. 9to5Google reported everything back to normal by 12:38 p.m. PT.
Did the Claude outage hit the API and Claude Code as well as the chat app?: Yes. The Register listed claude.ai, Claude Code, Claude Cowork and the Claude API among the affected services, with elevated error rates across Sonnet 5 and other models. 9to5Google reported that most Claude models recovered first while Opus 4.8 and Opus 5 lagged behind, so agent and API traffic was hit alongside the consumer app.
Which large AI assistant stayed online during the September 3 outages?: Google Gemini kept working. 9to5Google reported that Gemini saw its own spike in user reports at the same time, most likely from people arriving after the other services failed, but the service itself was not affected. That made Gemini the one major assistant a team could fall back to during the window.