Update! Claude experienced a global outage of over four hours, and Anthropic states the issue has been resolved.

On the evening of March 2, 2026, Anthropic’s Claude experienced a global service outage, affecting website login and multiple functionalities. According to the official status page, the issue began at 11:49 UTC and has now been resolved and is under monitoring. The incident impacted multiple regions, with many users reporting login failures and server errors, disrupting users who rely on Claude for development, content generation, and workflows.

Event Timeline: Approximately 4 hours from investigation to resolution

Based on official information from status.claude.com, the development is as follows (UTC):

11:49: System anomaly begins, users experience login failures and error messages

12:21: Official confirms the issue is related to Claude.ai login/logout pathways; API functions are normal

13:22: Engineering team identifies the problem and begins implementing a fix

14:05: Some API methods start experiencing errors; ongoing troubleshooting

14:42: Confirmed elevated errors in Claude Opus 4.6

15:25: Fix completed, entering monitoring phase

The overall outage lasted about 3 to 4 hours, marking one of the largest service disruptions since 2026.

Claude website and Claude Code were affected by the outage, while the API remained relatively stable.

The incident mainly impacted the Claude website and Claude Code, with some models (including Opus 4.6) experiencing increased error rates. Many users reported login loops, HTTP 500/529 errors, connection drops, or request failures. However, the Claude API remained relatively stable, and platform.claude.com was functioning normally. The official statement emphasized that the issue was at the platform and authentication process level, not a fault with the models themselves.

Global user reports: workflow disruptions, developer impacts

On X (formerly Twitter), many users reported being unable to log in, halting development work, inability to use Claude Code, and interruptions to content and research tasks. Some users stated that Claude has become a core tool in their daily work, so the outage’s impact was much greater than typical website failures.

The cause has not been disclosed yet, but community speculation includes infrastructure issues and attacks, cloud infrastructure anomalies (Claude mainly runs on AWS), problems with Middle Eastern data centers, traffic surges, or external attacks.

This article has been updated! Claude experienced a major global outage lasting over four hours, and Anthropic states the issue has been resolved. The earliest report appeared on Chain News ABMedia.