Microsoft has moved to full recovery after a widespread disruption to Microsoft 365 that knocked services offline for more than a day, impacting common productivity tools including Outlook and Teams. The company emphasized that the incident was technical in nature and not linked to a cyberattack.

In updates shared via the Microsoft 365 status channel, the firm said service restoration continued in stages, with remaining affected customers and regions brought back online and routed to the relevant incident details in the Microsoft 365 admin center under MO1550821.

Key takeaways

Microsoft reported that Microsoft 365 service disruptions were restored progressively after more than 24 hours of outages.
The company traced the outage to an issue introduced during an earlier network capacity upgrade and triggered after a subsequent network change.
Microsoft said the disruption was not caused by cyber activity or AI/agent-related systems.
Third-party platform status pages indicate that Microsoft 365 integrations in ChatGPT and Claude for Outlook were later returned to normal operation.

A technical outage, not a security incident

Microsoft’s account of the disruption points to the failure beginning with changes to its network environment. The company said the outage stemmed from a problem introduced during a previous network capacity upgrade, which only surfaced after a more recent network change.

In its communications, Microsoft reiterated that the incident was a “technical issue” and explicitly not related to cyber threats or AI/agents. That clarification matters for organizations running mission-critical workflows on Microsoft 365, because it narrows the likely root cause and reduces the immediate risk calculus that typically follows security-linked incidents.

Thursday outage expands across Microsoft 365 services

The outage began on Thursday and affected access to multiple Microsoft 365 components. Reports noted disruption to Exchange Online mailboxes, alongside features connected to Outlook and Teams, as well as services associated with Microsoft Copilot.

Government agencies in Japan, Singapore, and the United States were among those reporting email issues. The impact was not confined to traditional email workflows; connections between Microsoft 365 and AI assistants were also affected, showing how deeply the productivity suite is interwoven with newer assistant features.

Microsoft said it worked to address the underlying issue and restored services gradually, with stability checks before moving to the next stage of recovery. The company also apologized for the disruption and indicated it would publish a full post-incident report once its review is complete.

Integrations with ChatGPT and Claude for Outlook also recovered

Beyond Microsoft’s own service restoration, the incident had knock-on effects for AI integrations that depend on Microsoft 365 connectors.

OpenAI later reported that the Microsoft 365 connector in ChatGPT was operating normally again. Similarly, Anthropic marked its Microsoft 365 connector and Claude for Outlook incident as resolved, according to its status updates.

For builders and enterprise users, this sequence highlights a practical point: even when core SaaS providers restore their systems, downstream applications tied to external connectors may require additional time to return to full functionality. Observability from third-party status pages can therefore be an important guide for teams monitoring integration health during and after major provider incidents.

What to watch as Microsoft prepares its post-incident report

Microsoft has said it will publish a fuller post-incident report once its review is complete. That follow-up will be key for understanding whether the network-related failure is tied to a specific configuration, what safeguards were in place during the capacity upgrade, and how Microsoft plans to prevent recurrence.

In the meantime, organizations relying on Microsoft 365—especially those using mail routing, calendar automation, Teams collaboration, or Copilot workflows—should continue verifying that business-critical features and AI-linked connector functionality are stable across regions.

The immediate recovery is a positive development, but the more detailed technical findings from Microsoft will determine how quickly enterprises can close the operational loop and adjust monitoring or runbooks for similar network-change disruptions.