Microsoft 365 and Outlook slowly recover after nearly two-day global outage

What happened

Microsoft 365 and Outlook are slowly returning to normal after an outage that has stretched to nearly two days and left millions of users around the world without reliable access to email and other workplace tools, The Verge reports. Some services remain in a degraded state, and even users whose mailboxes are reachable again may see delays while Microsoft works through a backlog of queued messages.

How wide the disruption reached

The incident went well beyond inbox access. According to The Verge, the problems affected Outlook and the parts of Microsoft 365 that depend on Exchange Online and its email plumbing, including SharePoint Online, Teams and Copilot, along with other 365 business services. For organisations that run daily operations on Microsoft's cloud, documents, meetings and AI features were hit at the same time as email.

Where recovery stands

In a status page update published at 7AM ET on Friday, Microsoft said recovery was moving in an encouraging direction in Europe, with availability and the speed of email delivery improving across more of the affected infrastructure. Japan remained comparatively stable, the company said, while North America was also improving, though users could still run into intermittent issues while fixes continued.

Microsoft added that draining the backlog of queued email is a priority, and that it has extended the window before undelivered messages time out to 48 hours, protecting delayed mail from being discarded while delivery catches up. Some people, including staff at The Verge, have already reported that access to their email has been restored. There was still no estimate for when full service would return.

What caused it

Microsoft said it has found no sign of malicious activity behind the incident, and committed to publishing a detailed post-incident report covering what went wrong and how it was fixed once the problems are fully resolved. In the meantime, The Verge's senior correspondent Tom Warren reports that the disruption is tied to configuration changes to Azure VNet, Microsoft's virtual networking service. In other words, a routine infrastructure change on the cloud side, rather than an external attack, appears to be the trigger.

Why it matters

An outage of this scale is a reminder of how much modern work is concentrated in a single vendor's cloud. When one configuration change can simultaneously take out email, chat, collaboration documents and AI tooling for millions of paying customers, the blast radius of even routine cloud operations becomes clear.

It also puts operational safeguards under a spotlight. The 48-hour email timeout extension shows Microsoft manually compensating to avoid losing queued messages, but the near two-day duration will raise questions for enterprise customers about redundancy, service-level commitments and how quickly a bad change can be isolated once it is live. The promised root-cause report will be closely read by IT teams weighing how much of their business they are willing to run on a single provider.