Widespread AI outage hits ChatGPT, Claude and Grok at the same time

Three of AI’s biggest rivals suddenly had something in common Thursday morning: they stopped working.

OpenAI’s ChatGPT, Anthropic’s Claude and xAI’s Grok all suffered service disruptions on September 3, sending users to outage trackers and company status pages as errors spread across some of the most widely used AI tools. The overlap was unusual enough to raise an immediate question: did three separate systems fail by coincidence, or was something deeper happening underneath them?

“OpenAI’s ChatGPT, Anthropic’s Claude, and xAI’s Grok are all down Thursday morning,” Axios reported, citing the outage company Downdetector.

OpenAI reported higher error rates across ChatGPT and Codex and said it was investigating. The company later applied a mitigation and began monitoring the recovery. Anthropic reported higher error rates across multiple Claude models, with Claude.ai, Claude API, Claude Code and Claude Cowork showing partial outages on its status page. Anthropic said it had identified the cause and was working on a fix. xAI’s status system recorded a model outage affecting Grok, with problems reported across its apps.

ChatGPT, Claude and Grok Suffer Simultaneous Outages as Microsoft Azure Reports Problems

The timing is what turned a routine technology headache into a much bigger story. Major online services have outages every day. Three competing AI platforms running into trouble during the same window is far less common, especially at a moment when millions of people and businesses increasingly treat AI assistants as everyday work tools.

Microsoft Azure reported a network infrastructure incident in its East US region during the same period. Microsoft said the issue affected a subset of infrastructure and could degrade dependent Azure services.

“The intrigue: Microsoft Azure, which provides cloud services for all three models, is also experiencing outages, and some tech outlets have suggested Microsoft’s problem may be contributing to the AI model outages,” Axios said.

Axios reported that Microsoft provides cloud services used by all three AI companies, fueling speculation that the Azure incident might be connected to the broader disruption. No company has confirmed a link, and overlapping timing is not proof of a shared cause.

Microsoft Azure problems add a bigger question

That distinction matters. AI services sit on layers of infrastructure that users rarely see, including cloud regions, networks, databases, authentication systems, model-serving clusters, and outside services. A problem in a shared layer can hit several products at once. Separate failures can look nearly identical from the outside.

Google’s Gemini also entered the picture. Users in several countries reported problems with Gemini through outage trackers, including errors, slow responses and failed generations. Google had not confirmed a broad consumer outage at the time of reporting, making Gemini’s situation less clear than the confirmed disruptions at OpenAI, Anthropic and xAI.

Meanwhile, ChatGPT has suffered major outages before. On January 23, 2025, OpenAI reported an “Elevated Error Rate for ChatGPT and API,” with users experiencing higher error rates across both services. A separate ChatGPT incident later that day caused elevated errors for several hours. OpenAI said the disruption stemmed from problems in its database layer and later outlined work to make that infrastructure more resilient.

Thursday’s incident shows how much has changed in a short period. A chatbot outage once meant people could not ask questions or generate text for a few hours. AI systems now sit inside software development, customer support, research, content production, internal operations, and automated workflows. A failure can interrupt real business processes.

The outage could still be several unrelated technical failures arriving at almost the same moment. A shared infrastructure issue could emerge as part of the explanation. For now, the companies have confirmed service problems and recovery efforts, but no common root cause.

That may be the bigger lesson from Thursday morning. The AI race has focused heavily on which company can build the smartest model. The simultaneous outages raise another question: how dependable is the infrastructure beneath the AI economy when several of its biggest services stumble at once?