Claude Went Dark for Thousands, Behind the Outage and How It’s Being Fixed Now

Many users of Anthropic’s AI assistant Claude experienced significant service disruptions, causing frustration and widespread reports of errors. Starting from around 10:29 AM ET, Down Detector recorded over 1,400 reports of issues related to Claude, signifying a notable outage affecting both desktop and mobile access.

Users encountered delays and error messages such as "Something went wrong," with some unable to log in or receive timely responses. Anthropic promptly acknowledged the issue and shared updates on Claude’s official status page, confirming an ongoing investigation into performance problems affecting Claude.ai and Claude Code, while the Claude API remained unaffected.

Chronology of the Outage and User Impact

1. Users reported sluggish or failed responses on both free and paid accounts, including the Max 5x plan, with frequent errors on submitting queries.
2. Signing into Claude was also disrupted, with some accounts receiving warnings about delivery issues linked to certain email providers.
3. Down Detector metrics peaked with over 1,400 incident reports shortly after the outage began, reflecting the wide impact across the user base.
4. Later, reports began to decline gradually but remained well above normal levels, indicating a partial recovery across various platforms.

Anthropic’s transparent communication was evident through frequent status updates. As of one key update posted at 11:44 AM ET, the company stated: "We’re currently investigating issues with Claude Code and Claude.ai. Some users may be unable to log in, and others may experience slower than usual performance. The Claude API is not affected." This helped clarify that the technical difficulties were localized to particular service components rather than the entire system.

Service Recovery and Resolution Efforts

Subsequent observations showed Claude working again on mobile devices, specifically iOS, with fewer disruptions reported. Desktop web access experienced a delayed recovery but gradually resumed normal functionality. Down Detector reports dropped below 1,000 incidents as reliability improved.

Anthropic then updated their status page confirming that the root cause had been identified and a fix was actively being implemented. This statement reassured users that the problem was addressed and that service stability was being restored.

This outage underscores the challenges of scaling AI-powered services to accommodate rapid user growth. Despite the disruption, Anthropic’s swift acknowledgment and real-time updates helped manage user expectations during the downtime.

Ongoing monitoring is crucial given Claude’s increasing integration into workflows that depend on consistent AI assistance. Users are advised to check the official status page for updates and report any lingering issues to support channels as Anthropic continues to stabilize the system.

The incident provides valuable insight into how AI platform providers handle infrastructure disruptions and maintain transparency, which are critical for building trust among a growing user community.