Claude AI Down Twice in 24 Hours as Demand Surge Strains Anthropic

The Story

Anthropic’s Claude AI went offline twice within roughly 24 hours in early March, affecting users worldwide. The company classified the issue as an “elevated error” impacting consumer-facing services and related tools including Claude Code, Claude Console, and the latest Claude Opus 4.6 model, while saying the Claude API remained unaffected during the main March 2 disruption. Anthropic later stated that Claude was back up across claude.ai and its apps. The period with Claude AI down came as downloads surged amid political controversy surrounding Anthropic’s stance on military uses of AI.

The Facts

• Two outages within roughly 24 hours

  Claude experienced two global disruptions in early March, preventing users from accessing the AI chatbot and associated services. One outage on March 2 lasted about three hours, while additional disruptions were reported the following day. Anthropic’s status page also shows an earlier “elevated errors” incident on February 28, 2026 affecting Claude Opus 4.6.

• Issue classified as “elevated error”

  Anthropic’s status page categorized the incident as an “elevated error.” At 11:49 a.m. GMT on March 2, the company said it was investigating, later adding that the issues were tied to Claude.ai, login/logout paths, and some API methods not working.

• Multiple tools impacted, but API stayed up

  The disruption affected the Claude chatbot and products including Claude Code, Claude Console, and Claude Opus 4.6. Anthropic said the Claude API that powers businesses remained unaffected during the main March 2 outage.

• User reports confirmed widespread service failure

  Downdetector recorded more than 2,000 user reports during the March 2 outage, offering a clearer picture of how widely Claude AI down was being reported by users.

• Fix deployed by Anthropic

  Anthropic said it had identified the issue, implemented a fix, and later stated on March 2 at 4:37 p.m. UTC that Claude was back up across claude.ai and its apps.

• Unprecedented demand cited as a major factor

  Anthropic said it was restoring service while experiencing “unprecedented demand” for Claude over the previous week and later said it was working to scale infrastructure as quickly as possible to keep up with the surge.

• App Store ranking spike

  Claude surged to the No. 1 position on Apple’s App Store for free downloads in the United States, overtaking ChatGPT. ChatGPT was listed second and Google Gemini fourth.

• Free-user growth accelerated sharply

  Anthropic said free Claude users had increased by more than 60 percent since January, helping explain the strain on consumer-facing services.

• Federal agencies moving away from Anthropic products

  Anthropic declined to remove safeguards restricting fully autonomous weapons and mass domestic surveillance, arguing such uses exceed what current AI systems can safely support.

  The US administration ordered agencies to phase out Anthropic tools, labeling the company a “supply chain risk.” Some agencies reportedly began switching internal systems to other models such as OpenAI and Google Gemini. Anthropic has also said it had not yet received formal notices of the designation.

• Anthropic expected to challenge the designation

  The company has described the designation as “legally unsound” and is expected to challenge the move.

• Usage surge linked to US government dispute

  Claude usage increased sharply after Anthropic refused to grant the US military unrestricted access to its AI systems.

Key Risk Factors

• Cloud infrastructure dependency

  Anthropic relies on cloud infrastructure to run Claude at scale, and the broader internet increasingly depends on a small number of providers such as Amazon and Microsoft, making outages more visible and potentially more far-reaching.

• Centralized systems create broad failure points

  New Scientist reported on March 4 that experts see the shift to cloud computing as a core vulnerability of the modern internet: when centralized infrastructure fails, many services can go down at once. The same reporting said large, wide-ranging incidents are often more consistent with systemic failure or human error than with a deliberate broad cyberattack.

Industry Reaction

The outages triggered widespread discussion across social media and developer communities. Reports of Claude AI down quickly spread as users shared login failures and service errors, highlighting how widely the chatbot is now used for research, coding, and everyday tasks.

Why This Matters

The incident shows how concentrated cloud dependence can affect the reliability of AI platforms. The stretch with Claude AI down was not just a chatbot failure; it was a visible example of how frontier AI services now sit on shared digital infrastructure that can become a bottleneck when demand jumps suddenly.

This article was drafted with the assistance of generative AI. All facts and details were reviewed and confirmed by an editor prior to publication.