Eight major service disruptions in a single month. That is where Anthropic's Claude stands as of March 25, 2026, after yet another round of "Elevated Errors on claude.ai" knocked users offline for roughly two hours this afternoon.
The latest incident started at 13:45 UTC, with users hitting the familiar "This isn't working right now" error screen. Down Detector reports spiked from 150 to over 5,300 within ten minutes. Anthropic's team identified the root cause by 14:33 UTC and declared the incident resolved by 15:43 UTC, but offered no explanation of what actually went wrong.
This pattern has repeated all month. According to Claude's own status page, every non-government Claude service is now running below two nines of uptime (99%) for March. Over the past 90 days, claude.ai sits at 99.23% uptime and the API at 99.28%. For context, 99% uptime means roughly 7 hours of downtime per month. Two nines is the kind of number that makes enterprise buyers nervous.
What Is Driving the Instability
The outages appear to follow a whack-a-mole pattern: Anthropic stabilizes one service, and another breaks. The March 2 incident was the most severe, cutting off thousands of users globally. Subsequent disruptions have hit different parts of the stack, from login systems to specific models like Opus 4.6 and Haiku 4.5.
Part of the problem may be success. Claude recently climbed to the top of the App Store charts after a long stretch outside the top 20. More users means more load, and the infrastructure is clearly straining under it.
The Bigger Problem for Paying Customers
If you are building workflows around Claude's API, or paying $20/month for Pro, this month has been rough. Eight outages in 25 days means you are losing access roughly every three days on average. The API numbers matter even more for developers integrating Claude into production apps - 99.28% over 90 days is below what most SLA-conscious teams would accept.
Anthropic has not published a post-mortem or made any public statement about the pattern of March outages beyond individual incident updates on their status page. For a company positioning itself as the safety-focused, enterprise-ready AI lab, the silence is conspicuous.