On the morning of 3 September 2026, three widely used artificial intelligence platforms experienced overlapping service disruptions. OpenAI's ChatGPT, Anthropic's Claude, and xAI's Grok all reported problems on Thursday morning.
While nearly 38,000 users logged complaints about ChatGPT on Downdetector, Claude and Grok each recorded roughly 1,300 reports at their peaks around 11:00 a.m. ET. But what came as a surprise was that Google's Gemini reported no official outage.
OpenAI acknowledged the problem directly on its status page, confirming 'elevated errors across ChatGPT and Codex,' a reference to its developer-facing code-generation tool. The root cause, however, was not immediately reported.
Microsoft Azure, which provides cloud infrastructure for AI services from OpenAI, Anthropic, and xAI, also recorded a minor spike in outage reports during the same period. However, it was not immediately clear whether the Azure issues were connected to the three AI disruptions.
The overlapping outages coincided with reports of problems affecting some Microsoft Azure services, prompting speculation about whether shared infrastructure played a role. However, neither Microsoft nor the affected AI companies had established a common cause, and there was no public evidence showing that the consumer versions of ChatGPT, Claude, and Grok were all affected by the same Azure incident.
Interestingly, Google did not issue a corresponding official Gemini outage notice during the period, although Downdetector also recorded some user reports involving Gemini. Gemini runs on Google Cloud, Google's cloud infrastructure platform.
The contrast nevertheless highlighted the infrastructure question: whether greater concentration among cloud providers could create common points of failure for competing AI services.
How Each Platform Was Affected
Anthropic initially reported elevated errors affecting several Claude model families. As service recovered, Opus 4.8 and Opus 5 remained affected after other models returned to normal error rates.
The affected models include some of Anthropic's most capable and most commercially deployed systems. Claude's API is integrated into third-party developer applications, meaning an outage can extend beyond users accessing Claude directly through a browser.
Grok, the AI chatbot operated by Elon Musk's xAI, reported an outage affecting its web interface and related services. The incident overlapped with the disruptions affecting ChatGPT and Claude.
The disruption came only days after Grok for Government was added to GenAI.mil, the US military's secure generative-AI platform, alongside Google's Gemini and OpenAI's ChatGPT Mil. The Pentagon announced the expansion on 1 September, although there was no immediate evidence that the 3 September Grok outage disrupted the government deployment itself.
Google had not posted a corresponding Gemini incident on its public Google Cloud Service Health dashboard at the time, despite some user reports.
Why AI Reliability Matters
The concentration of critical AI workloads among a small number of major cloud providers has long been recognised as an enterprise technology risk. The latest outage was a real-world stress test at the consumer scale.
Microsoft did not immediately attribute the AI platform disruptions to Azure. The connection between the Azure outage reports and the AI platform failures could not be independently confirmed.
The disruptions underscored the extent to which AI platforms are becoming operational dependencies in business workflows, developer tools, and government systems.
Businesses and developers that rely on third-party AI APIs can face interruptions when those services experience elevated error rates, while government deployments such as GenAI.mil highlight the growing importance of reliability for institutional AI systems. There was no immediate evidence, however, that the Grok outage disrupted the Pentagon's GenAI.mil deployment.
Debate around AI models has largely focused on capability, safety, and price. The overlapping outages added another concern: resilience. When the infrastructure underneath a platform fails, who stays standing?
