- Anthropic’s Claude experienced at least five service disruptions during March 2026, affecting Claude.ai, Claude Code, and mobile apps while the API remained largely functional.
- The March 2 outage began at 11:30 UTC with approximately 2,000 users reporting issues, followed by recurring elevated errors on Claude Haiku 4.5 and Opus 4.6 models throughout the day.
- The March 11 incident triggered over 1,400 reports on Down Detector and lasted approximately two to three hours, with login failures and “Something went wrong” errors.
- Anthropic attributed the disruptions to infrastructure scaling challenges as user volume surged following a wave of migrations from competing platforms.
What Happened
Anthropic’s Claude AI assistant suffered at least five significant service disruptions during March 2026, with major incidents on March 2, March 11, March 20-22, and March 25-27. The outages primarily affected Claude.ai, Claude Code, and mobile applications, while the API generally continued operating normally. The cluster of disruptions prompted calls from enterprise customers for greater transparency and proactive capacity planning.
The March 2 incident, confirmed by BleepingComputer, began at 11:30 UTC with approximately 2,000 users reporting failed requests, timeouts, and inconsistent responses. The issue affected both Claude Haiku 4.5 and Claude Opus 4.6 models. Anthropic implemented a fix by 18:07 UTC, but the problem recurred, requiring a second fix at 18:54 UTC.
Why It Matters
The March outages occurred during a period of rapid user growth for Claude. According to analysis by industry observers, a wave of users migrated to Claude following disputes between competing AI providers and regulatory authorities, pushing the Claude app to the top of App Store charts. Infrastructure built for a certain user volume could not automatically scale to handle two to three times that volume within a single week.
For enterprise customers who depend on Claude for production workflows, the repeated disruptions raised questions about Anthropic’s infrastructure resilience. The pattern of the API remaining functional while consumer-facing products failed suggests the bottleneck was in the web application and authentication layers rather than the underlying model serving infrastructure.
Technical Details
The March 2 outage manifested as login failures and elevated error rates across web and mobile clients. Anthropic’s status page updates confirmed the company was “investigating” at 11:49 UTC and continued troubleshooting through 12:06 UTC before implementing fixes later in the afternoon. The issues specifically impacted the login and logout authentication paths.
On March 11, Down Detector registered over 1,400 user reports starting at 10:29 AM ET. Anthropic issued a status update at 11:44 AM ET confirming issues with both Claude Code and Claude.ai. The company identified the problem by approximately 2:22 PM ET, with total disruption lasting roughly two to three hours. Users reported “Something went wrong” errors and email delivery issues affecting login verification. The iOS mobile app continued functioning during this incident.
The March 25-27 incident affected Opus 4.6 and Sonnet 4.6 models specifically. Anthropic traced the root cause to networking performance degradation within its infrastructure and resolved it by migrating affected workloads to healthy infrastructure. A separate issue on March 20-22 caused responses to hang for approximately five seconds after text finished streaming.
Who’s Affected
Both free and paid Claude users experienced disruptions, including subscribers on the Max 5x tier. Enterprise API users were largely unaffected, as the API continued processing requests during most incidents. Developers using Claude Code, Anthropic’s CLI tool for software development, lost productivity during the March 11 outage specifically.
Businesses evaluating Claude for production deployment may factor these reliability incidents into their vendor selection process. Competitors including OpenAI and Google have experienced similar scaling-related outages during their own growth phases, but the concentration of five incidents within a single calendar month is notable and raises questions about infrastructure headroom.
What’s Next
Anthropic has not published a formal postmortem covering the full scope of March disruptions. The company’s status page updates during each incident were limited to confirming investigation and resolution without disclosing root causes in detail. As Claude’s user base continues to grow, Anthropic faces pressure to invest in infrastructure redundancy and auto-scaling capabilities that can absorb sudden demand spikes without service degradation. The March incidents serve as a stress test that exposed specific weaknesses in the authentication and web application layers that the company will need to address ahead of further growth.
