Anthropic's Claude, Perplexity, and OpenAI's ChatGPT all suffered major outages in a string of incidents that affected users all over the world. According to OpenAI's service status tracker, ChatGPG went offline many times, first in the early hours of June 4 and then again later that same day. The problem was resolved by 04:19 AM PDT, having begun at 12:21 AM PDT. But at 07:33 AM PDT, there was another disruption, which ended at 10:17 AM PDT. During this time, the service indicated it was at capacity, humorously updating its status in pirate speak. It is reported that the outage did not impact its API or other OpenAI platforms, such as platform.openai.com. With over 100 million weekly users, ChatGPT had encountered similar interruptions due to third-party API dependencies in the previous month.
Claude and Perplexity both reported service outages at the same time. A little after 12:10 PM ET, Claude's service resumed regular operations after displaying an error warning and advising users to try again later. Perplexity similarly informed users of reaching capacity due to high traffic, with intermittent availability following the resolution of Claude's issues.
The parallel occurrence of these outages on several different AI providers points to a possible internet-wide issue or underlying infrastructure issue, similar to the disruptions that have affected multiple social media sites at the same time. Although it was up and running at the time of reporting, Google's Gemini was also the subject of brief reports of outages.
In addition to being inconvenient, the simultaneous outage raised questions about possible outside influences and how resilient the current AI infrastructures are during peak loads. As these platforms recover, the tech community is keenly observing the response strategies and preventative measures to reduce future interruptions.
Turn data into intelligence with Machine Learning at Code Labs Academy.