Anthropic’s AI chatbot Claude has run into fresh trouble, going offline for the second time in less than a day. Thousands of users across the United States and India reported being unable to access the platform, raising fresh concerns about stability as demand for generative AI tools continues to surge.
Data from Downdetector shows that over 4,000 users in the US flagged problems with the service. In India, close to 500 complaints were recorded at the time of writing.
Claude Users Report Chat, App & Website Failures
The bulk of complaints centred around Claude’s core chat function, with nearly 39 per cent of affected users reporting issues in accessing conversations. Another 36 per cent said the mobile application was not functioning properly, while about 15 per cent struggled to open the website.
Anthropic’s official status page acknowledged ongoing disruptions, listing “Elevated Errors” across claude.ai, cowork, platform and claude code. The issue remains marked as unresolved.
The repeated outage comes at a time when AI assistants are increasingly embedded into workflows, making even short downtimes disruptive for developers, enterprises and individual users.
Second Disruption Within A Day
This marks the second service breakdown in under 24 hours. On March 2, Claude, along with its console and claude code services, also reported elevated error rates.
During the earlier disruption, users encountered two major technical responses. The first was an HTTP 500 error, a generic internal server issue indicating a malfunction on Anthropic’s servers that prevented request processing. The second was an HTTP 529 error, which typically appears when systems are overloaded and unable to manage the incoming volume of traffic.
Such errors often signal strain on backend infrastructure, particularly during periods of heavy demand.
Company Says Investigation Is Ongoing
As of the latest update, Anthropic has stated that it is still probing the root cause. The company’s status log reads: “Update – We are continuing to investigate this issue.Mar 03, 2026 – 04:43 UTC Update – We are continuing to investigate this issue.Mar 03, 2026 – 04:41 UTC Update – We are continuing to investigate this issue.Mar 03, 2026 – 04:39 UTC Investigating – We are currently investigating this issue.Mar 03, 2026 – 03:15 UTC”
There is no official timeline yet for the full restoration of services.
With back-to-back outages in less than a day, attention is now likely to turn toward infrastructure resilience and how AI providers are scaling systems to meet rapidly growing user demand.
