At 9 a.m. last Thursday, Exchange Online servers in Switzerland began falling offline in what became one of Microsoft’s worst recorded service disruptions, The Verge reported.
The failure quickly spread to Mexico and other countries as automated failover systems attempted to shift email traffic.
Within hours, more than 80 percent of Exchange servers were unresponsive, leaving businesses, governments, and hospitals without access to Outlook for nearly 24 hours. Microsoft marked the incident as “sev0,” its highest internal tracking category for catastrophic outages, according to The Verge.
More details on the root cause can be found in the Microsoft 365 Admin center, and an abridged version can be found below. We are deeply sorry for the impact to all our customers and recognize the impact that this has on their normal business & operations. The incident on… https://t.co/p8mUHxDe67
Three Factors Causing Cascade
The office software giant issued an initial report identifying three specific triggers that created the global blackout, The Guardian reported. Those factors included a months-old software defect, malfunctioning internal connections, and reduced server capacity.
In a post on X, Charles Lamanna, executive vice-president of Copilot, Agents and Platform at Microsoft, explained that a latent bug from July caused servers to become overwhelmed under heavy loads. That defect interacted with a network gateway failure when Microsoft Teams failed over to the primary gateway, compounding the traffic strain.
Much of the modern world runs on Microsoft. Hospitals, airlines, banks, courts, governments — they all depend on its software and cloud servers. Last week a hidden bug knocked out Microsoft’s Outlook, Teams, and Copilot software around the world for roughly a day. It was so bad… pic.twitter.com/XD6tggKKua
“A latent defect caused servers to become overwhelmed and unresponsive under certain load conditions,” wrote Charles Lamanna, executive vice-president of Copilot, Agents and Platform at Microsoft, in a post on X.
Adding to the pressure, about 14 percent of the company’s server capacity was unavailable due to an unrelated issue from the previous day, according to The Guardian. Testing labs had missed the vulnerability because they focused on older, weaker hardware with fewer cores, whereas newer multi-core machines experienced worse lock contention, Computing UK reported.
Inside Microsoft's massive Outlook outage. In this week's (slightly earlier) Notepad newsletter, I dig into one of Microsoft's biggest cloud outages in history, what went wrong, and why it took so long to fix. Live now for subscribers (gift link) 👇https://t.co/nFNWBnbq0M
Estimating Global Insured Losses
Global cyber insured losses from the November 12-14 outage are estimated by various industry observers to range from $100 million to $700 million, Insurance Journal reported. CyberCube estimated the losses at under $100 million with an upper limit of $350 million, while Morningstar DBRS estimated global insured losses between $300 million and $500 million, and broker Acrisure placed the estimate at $700 million.
The disruption extended beyond email, affecting Teams calendars, meeting scheduling, Copilot, SharePoint, and Intune, Digital Trends reported. Artificial intelligence providers OpenAI and Anthropic were also impacted by integration tool failures, while air carriers including Air India, IndiGo, and Akasa Air reported support issues, and a Taiwanese firm temporarily resorted to fax machines to confirm trades.

Microsoft posted on X on Saturday, November 14, to confirm that services had been restored for the remaining affected customers and regions, Digital Trends reported. The company said that it is deeply sorry for the disruption and plans to publish a final post-incident review later this week.
“Our normally automated recovery processes were not designed to restore service from this broad unresponsive state. Additionally, since the main code defect was a latent bug from several months previous, our normal rollback, flighting, and feature flags were unable to quickly revert the change,” Lamanna wrote.
Last week's global Microsoft Outlook outage exposed a serious vulnerability. Businesses, hospitals, schools, and government agencies lost access to critical communications. Now imagine the consequences if malicious actors had been behind the disruption. That's why I introduced…
Common Outage Questions
Why did Exchange Online take longer to recover than other services?
Exchange servers boot up and run an operating system over a network using Preboot Execution Environment, and the sudden burst of rebooting traffic overwhelmed monitoring systems, creating a reboot loop, according to The Verge. Automated recovery processes were not designed to restore service from such a broad unresponsive state.
Was the global outage caused by a cyberattack?
Crash reports and logs quickly ruled out the possibility of an attack, and Microsoft did not attribute the blackout to a cybersecurity breach, The Guardian reported.
What hardware differences contributed to the missed bug?
The contention issue worsened on machines with more cores and newer hardware, whereas Microsoft’s load testing labs tended to focus on weaker hardware with fewer cores, The Verge reported.
“The contention issue is worsened on machines with more cores and newer hardware; whereas our load testing tended to focus on weaker hardware with fewer cores,” Lamanna wrote.