Claude’s instability disrupts key services ahead of Nasdaq debut

2–3 minutes

Current situation

On Friday, November 13, Claude, an AI assistant developed by Anthropic, experienced performance issues. The claude.ai website and Claude Cowork showed degraded performance, while the API, Claude Code, Console, and versions hosted on partner clouds like AWS, Google Cloud, and Microsoft Foundry continued to operate normally.

Microsoft 365 connector issue

Since the afternoon of Thursday, November 12, Anthropic has been monitoring issues with the Microsoft 365 connector, which integrates Claude with email, calendar, SharePoint, and Teams. The investigation began at 2:52 PM Brasília time when users started reporting errors and slow responses in these integrations.

  • 2:52 PM: Investigation launched following reports of slowdowns.
  • 3:34 PM: Problem traced to an external vendor.
  • 7:23 PM: Anthropic attributed the failure to an incident in Microsoft 365 services, rendering email, calendar, and search tools inoperable for most requests.

The company advised users to monitor Microsoft’s status page and pledged to provide updates once the service was restored. Claude for Outlook was also affected.

Recent history of outages

Two other incidents were reported in recent weeks:

  • Tuesday, November 10: Fable 5.2 and Mythos 5.2 models returned overload errors for approximately four minutes.
  • Sunday, November 8: Claude Fable 5.1 experienced high error rates between 5:18 and 5:32 PM Brasília time, simultaneously affecting claude.ai, the API, Claude Code, and Claude Cowork. A fix was quickly applied, and the incident was resolved the same morning.

How to monitor status

Anthropic provides real-time updates on status.claude.com, where users can check the status of each service and view incident history. User reports can also be tracked on Downdetector.

Frequently Asked Questions

Is Claude down? On Friday, November 13, degraded performance was reported for claude.ai and Cowork, which may result in slowdowns or errors. The API, Code, and Console appeared to be operational.

When will the service return to normal? At the time of publication, Anthropic had not provided a timeline for normalization or opened a specific incident report regarding the slowdown.

Is Microsoft’s failure related to Claude’s slowdown? Anthropic attributed the errors in the Microsoft 365 connector and Claude for Outlook to an incident in Microsoft’s services. The cause of the slowdown in claude.ai and Cowork remains unclear.

How can I monitor the status? The status.claude.com page provides up-to-date information on operational, unstable services, and incidents under investigation.