Claude AI Services Recover Following Brief Disruption Across Global Networks
San Francisco, Tuesday, 29 September 2026.
Anthropic restored its Claude AI platform following a widespread disruption on September 29, 2026, highlighting systemic operational risks for businesses increasingly reliant on automated digital workflows.
Service Disruption Confirmed on September 29
Anthropic, the San Francisco-based artificial intelligence developer, confirmed a significant service outage affecting its Claude AI platform on Tuesday, September 29, 2026 [1]. The disruption began shortly before 10:00 a.m. ET, preventing users and developers from authenticating or accessing automated workflows across multiple interfaces [3]. While the company initially listed the status as investigating, it quickly acknowledged elevated error rates affecting both consumer and enterprise-facing tools [5]. This incident underscores the operational dependencies corporations are forming around generative AI systems for daily business functions [1]. The outage was marked as a partial outage on the official status dashboard, indicating specific components were degraded rather than a total system failure [6].
Timeline of Mitigation and Recovery
Anthropic initiated a mitigation strategy approximately 36 minutes after the errors began, with improvements noted by 10:36 a.m. ET [4]. By 10:59 a.m. ET, the company stated that all services had been operating normally, although some users continued to experience secondary issues with sign-ins [1]. The rapid response time highlights the pressure on infrastructure providers to maintain uptime amidst surging demand [3]. Despite the mitigation, some users reported persistent difficulties with single sign-on (SSO) and Sign in with Apple services during the recovery window [6]. The company advised users who were still signed in not to log out to avoid authentication loops [6].
Impact Metrics and User Reports
Data from DownDetector showed a sharp spike in user reports, peaking at over 11,800 incidents at 10:17 a.m. ET on September 29 [2]. Following the mitigation efforts, reported issues dropped to 3,371 by 10:59 a.m. ET, representing a significant reduction in error volume [2]. The percentage decrease in reported incidents during this 42-minute window can be calculated as 71.432 [2]. This volume of reports indicates a widespread impact across global networks, with users encountering failed requests and errors loading conversations [5]. The scale of the disruption was sufficient to trend on search engines and social media platforms within the first hour of the outage [2].
Scope of Affected Services and Context
The outage impacted Claude.ai, desktop and mobile applications, Claude Code, Claude Cowork, and the Claude API [5]. Claude for Government was notably exempt from the disruption, maintaining operational status throughout the incident [3]. This event follows a pattern of instability earlier in the month, with previous incidents reported on September 22, September 15, and September 11, 2026 [6]. A major outage involving multiple AI providers occurred earlier in September, suggesting broader industry-wide infrastructure challenges [4]. Anthropic continues to monitor the situation to ensure full stability across all product lines [3].