Anthropic’s Claude AI services experienced four significant outages during March 2026, affecting Claude.ai, the mobile apps, API access, and Claude Code. The most recent incident on March 22 took down Claude.ai and Claude Login before being resolved at 18:51 UTC. The pattern of repeated disruptions has drawn attention from developers and enterprise customers who rely on Claude for production workloads.
The month began with a major outage on March 2, when approximately 2,000 users reported failures at 11:30 UTC. Affected services included Claude.ai, mobile applications, and the API, with symptoms ranging from failed requests and timeouts to inconsistent responses. Anthropic acknowledged the incident and restored service, but did not publish a detailed post-mortem explaining the root cause.
A second widespread outage on March 10 introduced login failures and OAuth timeouts, preventing users from accessing their accounts entirely. This incident was particularly disruptive for teams using Claude through enterprise single sign-on integrations, where the OAuth failure cascaded into workspace-wide lockouts. The March 17 incident was the largest by user reports — over 5,000 people flagged issues, primarily with Claude Code and the API, encountering HTTP 500 errors that indicated server-side failures rather than client configuration problems.
For developers building applications on Claude’s API, the reliability pattern is a commercial concern. Production systems that depend on Claude for real-time inference — customer support agents, code review pipelines, content generation workflows — require uptime guarantees that four outages in three weeks cannot deliver. Anthropic offers a status page but has not published service level agreements with financial penalties for downtime, a standard practice among cloud infrastructure providers.
The outages coincide with a period of rapid growth for Anthropic’s services. Claude Code, launched in late 2025, has driven significant new API traffic from developer tools and IDE integrations. Claude Opus 4.6, released in early 2026, attracted heavy usage from enterprise customers. The combination of new users, new products, and new model deployments creates scaling pressure that may explain the increased incident frequency — though Anthropic has not confirmed whether capacity constraints are a contributing factor.
