On March 2–3 local time, Anthropic’s AI assistant Claude suffered a global service outage, preventing a large number of users from logging in or using the product normally. The company’s status page indicated the disruption impacted multiple ecosystem components, including the Claude web interface, API, Claude Code, and Claude Cowork. Anthropic first confirmed the issue around 1:30 p.m. ET, then rolled out fixes and kept monitoring, though a few models such as Claude Opus 4.6 still showed anomalies.
According to monitoring data, the partial outage lasted nearly three hours and affected thousands of users. Anthropic thanked users for their patience in a statement and noted that the team was dealing with a recent surge in demand. Notably, the downtime came at a sensitive moment—Anthropic is currently in a standoff with the U.S. Department of Defense over military use restrictions, with the Pentagon even threatening to invoke the Defense Production Act to force open access to AI tools.
