Claude users worldwide experienced login failures and degraded web performance during a major service disruption that Anthropic investigated on its status page.
Anthropic first flagged elevated errors at 11:49 UTC and said engineers were investigating, the company reported on its status page.
Updates posted by Anthropic said some API methods were not working at 13:37 UTC while other messages noted that a fix was being implemented by 15:25 UTC, and that the incident was resolved by 15:50 UTC.
Downdetector data showed a sharp spike in user reports and nearly 2,000 active complaints at peak, with about 42 percent citing the web interface and 34 percent citing the mobile app, as reported by Downdetector.
Impact And Context
The outage affected claude.ai, the Claude Console, and Claude Code, with status updates describing degraded performance across front end and developer tooling.
Anthropic’s status notes also indicated conflicting signals about the API, saying the core Claude API appeared to be functioning for some developers while other status entries reported API method errors, suggesting a partial disruption of different service paths.
The interruption drew attention because Claude and Claude Code are central to many users’ workflows, and because Claude Code was released in 2025 as an agentic system that can autonomously execute high level programming tasks, according to AFROTECH.
Boris Cherny discussed Claude Code on Lenny’s Podcast on February 19, 2026, saying he has not manually edited a single line of code since November 2025 and predicting broad changes to software roles, as reported by AFROTECH.
The Independent noted the outage came amid scrutiny of Anthropic’s relationship with the US Department of Defense, reporting that the DoD pressed Anthropic for unrestricted access and that Anthropic refused on ethical grounds, while it reported OpenAI has provided AI help to the US military.
