Anthropic's popular AI chatbot Claude faced widespread disruptions Monday, March 2, 2026, with thousands of users reporting inability to access the service, login failures and error messages across platforms including claude.ai, the console and Claude Code.

The outage began surfacing around 11:49 UTC (6:49 a.m. ET), when Anthropic posted on its official status page (status.claude.com) that it was "currently investigating" elevated errors. Follow-up updates confirmed the team had identified issues tied specifically to consumer-facing surfaces — particularly the Claude.ai website, login/logout paths and related tools — while the core Claude API remained operational as intended.

Anthropic CEO Dario Amodei

By early afternoon UTC, Anthropic reported it had pinpointed the root cause and was actively implementing a fix, though no estimated time for full resolution was provided. Status updates through the day included notes that investigation continued, with some API methods initially affected before clarification that backend models and API endpoints were largely unaffected.

Downdetector, a service tracking user-reported issues, showed a sharp spike in complaints peaking around 6:40 a.m. ET, with nearly 2,000 reports at the height of the disruption. Complaints tapered to about one-third that level by mid-morning but remained elevated compared to normal. Users most frequently cited problems with the chat interface (around 39%), the mobile app (35%) and the website (15-20%), including HTTP 500 and 529 errors, timeouts, "connection terminated" messages and generic "This isn't working right now" prompts.

Anthropic described the incident as a "partial outage" in some communications, emphasizing it primarily impacted web-based access and authentication rather than the underlying AI models like Claude Opus 4.6 or enterprise integrations. Businesses using Claude via API or embedded systems appeared largely unaffected, allowing continued operations in those environments.

The precise technical trigger remained unclear in public updates, with Anthropic avoiding speculation on causes like server overload, configuration errors or external factors. Some reports suggested strain from "unprecedented demand" in recent weeks, as Claude gained traction amid competition with ChatGPT and rose in app store rankings thanks to features like advanced coding tools and improved reliability perception.

Social media platforms, including X (formerly Twitter), lit up with user frustration and memes, with developers joking about the irony of relying on Claude Code to potentially troubleshoot Claude outages. Others noted the timing amid broader AI reliability discussions.

The disruption came against a backdrop of heightened scrutiny for Anthropic. Recent news highlighted tensions with U.S. government contracts, including a reported Department of Defense termination over "safety alignment conflicts" and Claude's restrictions on certain military applications. Speculative posts linked the outage to unrelated geopolitical events, such as alleged strikes on AWS infrastructure in the Middle East — Anthropic's primary cloud provider — though no official sources connected the dots, and Anthropic attributed issues to internal web/login paths rather than cloud-wide failures.

Anthropic has not issued a broader statement beyond status page updates and brief confirmations to media. The company, known for its focus on AI safety and constitutional alignment principles, has seen rapid growth but faces ongoing challenges scaling infrastructure to match surging usage.

As of late March 2, 2026 (evening KST), partial recovery appeared underway for some users, though intermittent issues persisted in login flows and web access. Anthropic encouraged affected users to monitor the status page for real-time updates and advised trying again later or using API routes where applicable.

The incident underscores persistent challenges in AI infrastructure: even as models advance, front-end reliability, authentication systems and traffic handling remain critical pain points during peak adoption phases. Similar outages have hit competitors like OpenAI in the past, often tied to login surges or scaling hiccups.

Users worldwide expressed frustration over workflow interruptions, particularly those depending on Claude for coding, research or daily productivity. With no confirmed ETA, many shifted to alternatives like Grok, ChatGPT or local tools while awaiting restoration.

Anthropic's team continued working on resolution into the evening, with hopes the fix would fully restore service soon. The company has maintained high uptime historically, making Monday's event notable in its scope and visibility.