Notion restores Anthropic AI access after 12-hour infrastructure disruption

Notion has restored access to Anthropic’s artificial intelligence models following a temporary service disruption that resulted in degraded performance and increased failure rates for users. The incident, which affected the Opus 4.7 and 4.8 models within Notion AI, lasted approximately 12 hours before service was fully reinstated by early Sunday morning.

The disruption began early Sunday morning when Notion notified users that the specific Anthropic models were experiencing performance issues, leading to a higher rate of failures. In response, the company temporarily disabled the use of all Anthropic models in its automated productivity tool to manage the elevated error rates.

Notion’s head of product, Max Schoening, addressed the incident on social media, noting he was astonished by the volume of engagement on the initial outage notification. Public statistics on X indicated the post was reposted approximately 1,200 times. Schoening suggested that many users were eager to frame the technical hiccup as a reflection of poor model quality rather than a standard operational issue.

Schoening characterised the event as a brief infrastructure problem common to major technology platforms, drawing comparisons to incidents at GitHub, AWS, and other services. He emphasised that such disruptions are routine in the sector and did not indicate a fundamental flaw in the underlying AI models provided by Anthropic.

An Anthropic spokesperson confirmed the nature of the outage, stating that a short-lived infrastructure issue had caused elevated errors on multiple Claude models. The spokesperson noted the issue has since been resolved and thanked users for their patience while the company worked to restore service.

The broader technology sector frequently experiences similar infrastructure-related service interruptions, highlighting the complexities of maintaining reliability in cloud-based AI integrations. Notion’s integration with Anthropic’s capabilities remains a core feature of its productivity suite, with this incident representing a temporary setback rather than a long-term structural change.

As of early Sunday morning, access to the affected models has been fully restored, allowing users to resume normal operations within Notion AI. The company has not provided further technical details regarding the specific root cause of the infrastructure issue beyond the general description provided by Anthropic.