Widespread issues arose almost simultaneously across some of the largest artificial intelligence services, with users around the world reporting an inability to access OpenAI, Claude, Grok, Google Gemini, and Cursor. The timing of these outages suggests that the cause lies not with the individual AI models themselves, but with the infrastructure on which their operation relies.
Down Detector recorded a sharp increase in reports of problems at the same time. This data points to a malfunction that is not related to the code or algorithms of the AI models themselves, but possibly with a broader segment of the online infrastructure.
ΣΧΕΤΙΚΑ ΑΡΘΡΑ
One of the possible causes being examined involves the infrastructure of so-called hyperscalers—companies that operate massive data centers, such as AWS and Microsoft Azure. Many digital services host data on these infrastructures or route their traffic through peering points.
The impact, however, was not limited to AI services themselves. Software development tools such as Cursor also experienced issues almost immediately, as their operation relies on APIs from OpenAI and Anthropic.
The sheer number of reports also indicates that the issue was not limited exclusively to the field of artificial intelligence. During the same period, problems were also reported in other, unrelated services.
Based on the available evidence, the problem may be linked to a broader disruption in the international routing of internet traffic, including Tier-1 transit and DNS infrastructure.
Source: Proto Thema
