Microsoft 365 Crawls Back to Life After a 24-Hour Outage
AI-assisted digest, checked against the original source before publishing. How we work
Microsoft says a botched network upgrade—not a cyberattack or AI—crashed 365 services across 70+ countries for roughly a day.
Microsoft 365 services are coming back online after an outage that lasted roughly 24 hours, hitting businesses, governments, and tens of millions of users worldwide. Reports of Outlook trouble began appearing on Downdetector just after noon Eastern on Thursday; within an hour Microsoft said it was investigating what it initially described as a North American problem. The failure quickly widened to more than 70 countries.
Microsoft has repeatedly ruled out a cyberattack, and has also ruled out AI or AI agents playing any role. A company spokesperson attributed the incident to an earlier network capacity upgrade that surfaced after a recent network change, calling it "a technical issue not related to cyber or AI/agents." Microsoft says it has addressed the underlying issue and is continuing to restore service to remaining customers and regions, with a full post-incident report promised once its review is complete.
The recovery itself promises to be messy for users. Jonathan Leitschuh, an independent researcher, warns that inboxes and notification feeds may turn into a mess as subsystems come back online at different times—possibly delivering "a flood of notifications a few hours down the pike." Microsoft said it raised the email queue time to 48 hours so that messages sent while systems were down would still be delivered rather than dropped.
The blast radius extended well beyond email clients. Governments from Japan to the United States reported disruptions to functions ranging across their operations. Airlines had trouble communicating with customers, law enforcement and emergency services were disrupted, and hospitals reported widespread interruptions. The total impact may never be fully measured, but the outage is already being treated as significant in scale.
For builders and operators, the episode is a case study in single-provider concentration risk. Leitschuh framed the underlying condition plainly: "We're very interdependent on these systems in modern life," adding that next time delayed emails may be the least of the worries. The event sits alongside a track record of large-scale cloud incidents that cascade across sectors none of them own.
What remains open is the mechanics of the failure: which part of the network capacity upgrade interacted badly with the recent change, why the initial blast radius appeared regional before going global, and what guardrails Microsoft will add before the next routine network change. The promised post-incident report is the document to watch—for its account of detection and rollback gaps, and for whether 48-hour queues and staged subsystem recovery become standard playbook for this class of failure.
Key facts
- • Outage began just after 12 pm ET Thursday on the US East Coast.
- • Spread from North America to more than 70 countries.
- • Microsoft pinned cause on a network capacity upgrade plus recent network change.
- • Ruled out cyberattack and AI or AI agents as causes.
- • Email queue time raised to 48 hours to preserve undelivered messages.
Why it matters
Teams, Outlook, Copilot, and email dependencies should be treated as a shared failure domain: a single routine network change can halt operations across 70+ countries. Builders should expect post-recovery notification floods, plan for staged subsystem restoration, and pressure vendors for post-incident detail on change-detection and rollback controls before the next routine maintenance window.
Spotted an error? Send a correction.
Get the day's AI news in one email
The Daily Epoch · 60-second read · unsubscribe anytime
