Particle.news

Anthropic Investigates Login Failures That Disrupted Claude.ai and Claude Code

It highlights how fragile sign‑in systems make widely used AI tools vulnerable to failures in external services.

Overview

  • Anthropic logged elevated authentication errors on its status page starting around 09:58 UTC on Friday, November 13, 2026, and the company says engineers are actively investigating.
  • The impact varies by component with claude.ai marked as a major outage, Claude Code and Claude Console listed as partial outages, and the Claude API reported as operational.
  • Anthropic identified a simultaneous Microsoft 365 service incident as the cause of errors in its Microsoft 365 connector used for mail, calendar, and search integrations.
  • User reports from the United States, India, and the United Kingdom describe failed sign‑ins, unexpected logouts, and app‑loading problems while outage monitors showed mixed signals about the scope.
  • The episode follows earlier 2026 sign‑in problems and separate session‑token thefts that year, raising continued questions about the platform’s authentication resilience and third‑party dependencies.