What caused Claude's multi-hour outage?

Widespread errors disrupted Anthropic's Claude services

Anthropic experienced a significant service disruption that left many users unable to access Claude products for more than two hours. The company reported elevated error rates across public-facing services — including the web app and developer console — and users reported login failures and broken session handling. Engineers later said they had resolved the immediate errors and restored access.

Affected areas included:

• The consumer chat interface and Claude.ai
• Developer tooling such as Claude Code and the management console
• Some users reported interruptions to integrations that rely on Claude’s API

The company’s status messages and third‑party coverage indicate that API endpoints continued to function in some capacity even while the web front ends showed problems; Anthropic described the API as operating as intended even as consumers reported degraded experiences. No definitive root cause was publicly disclosed, and reporting suggests engineers were focused on fixing service stability rather than offering a technical post‑mortem immediately.

Why this matters

Service outages at leading AI providers have outsized effects because businesses increasingly embed models into customer‑facing workflows. Interruptions can halt developer builds, delay deployments, and erode confidence in a provider’s reliability—particularly for companies using AI for production tasks.

What we still don’t know

It remains unclear whether the disruption stemmed from software bugs, deployment errors, capacity strain, or an external dependency. Anthropic resolved the outage and restored normal operations, but a full technical explanation has not been released publicly.