Find a service or website

Type at least 2 characters to search.

ChatGPT and Claude outages hit AI users as reports spread across Gemini, Copilot and Grok

September 3, 2026
Close-up of a smartphone showing Gemini, Grok, ChatGPT and Claude app icons.

AI users encountered overlapping service problems on Thursday, September 3, as OpenAI and Anthropic confirmed separate incidents affecting ChatGPT, Codex and Claude products. Independent reports also described problems reaching Gemini, Microsoft Copilot and Grok. The timing prompted speculation about a shared infrastructure failure, but no provider published evidence connecting the incidents to Cloudflare or Amazon Web Services.

OpenAI and Anthropic confirm separate incidents

OpenAI’s incident record displayed an investigation into elevated errors across ChatGPT and Codex at 14:43 UTC. The company displayed a mitigation and recovery-monitoring update at 15:17 UTC, then marked the issue resolved at 16:55:49 UTC. OpenAI did not publish a cause, so the record supports a confirmed service incident but not a conclusion about the underlying infrastructure.

Anthropic opened its multiple-model incident at 13:26 UTC. The affected components included claude.ai, the Claude API, Claude Code and Claude Cowork, while its updates named several Mythos/Fable and Opus models. Anthropic said it identified a cause, deployed a fix at 16:06 UTC and ended the impact at 16:16 UTC, but did not identify the responsible system or vendor publicly. A spokesperson described the event to NewsNation as an infrastructure issue.

Gemini, Copilot and Grok evidence was less definitive

PCMag and The Hour reported concurrent user-facing problems across several AI platforms. NewsNation also included Copilot among the disrupted services, while stating that Google had not confirmed a Gemini incident. By early afternoon, the reports indicated that most services were recovering.

The public status evidence did not assign the same level of confirmation to every product. Microsoft’s public service-health page later showed all products operational without exposing a detailed historical Copilot incident, Google Cloud showed no matching broad severe event, and xAI’s status page displayed no declared incident when checked. Those gaps do not disprove individual user problems, but they mean the article should distinguish official incidents from reported symptoms.

Cloudflare and AWS were not confirmed as the cause

Cloudflare directly told PCMag that it was not experiencing any significant service disruption and that its services were operating normally. Its public status record showed only minor product-specific or regional events, including R2 and WARP issues plus earlier short-lived errors in Hong Kong, Seattle and Western North America. Cloudflare’s AI Gateway was operational. None of those notices established a link to the simultaneous AI-service problems.

The AWS Health Dashboard likewise showed no matching global incident during the AI outage window. Its visible multi-service disruption was the long-running, geographically limited impairment in the Middle East (UAE) region after physical infrastructure damage during regional conflict. Treating that separate event as the cause of worldwide chatbot problems would go beyond the available evidence.

Concurrent timing does not establish a shared cause. The confirmed record is narrower: OpenAI and Anthropic each resolved documented incidents, while several other AI products drew user reports without equivalent public detail. Readers can compare current signals on the ChatGPT, Claude, Gemini and Microsoft Copilot service pages and the Status Is Down real-time dashboard. Creating an account and selecting a monitoring plan can help users follow new disruption signals while official status updates establish scope and cause.

Sources

Image credit: Editorial image. ID 446605540 © PhotoPawel | Dreamstime.com — https://www.dreamstime.com/warsaw-poland-feb-close-up-smartphone-screen-displaying-various-app-icons-including-gemini-grok-cha-warsaw-poland-feb-close-image446605540

Recent Outage News and Features

Loading...

🌐 This article is available in other languages:

Français·Español·Português·中文·العربية·हिन्दी·日本語·Русский·Deutsch·Bahasa Indonesia·Türkçe·한국어·Kiswahili·Italiano·Nederlands·Polski·Tiếng Việt·Filipino·বাংলা