Microsoft's Outlook outage knocked out email in 74 countries for up to a day
A months-old bug took down Outlook and parts of Teams on November 12. It's fixed now, but scammers are using it to steal passwords, and businesses can claim money back.
Developing story · America's week of outagesAction neededIgnore emails offering to restore your lost Outlook mail- On November 12, Outlook email stopped working for millions of Microsoft 365 users in 74 countries, and Teams and Copilot were hit too2,4.
- Microsoft says it wasn't a hack: a bug from a July fix overloaded the servers that route traffic to its email system5,8.
- Service was fully back by November 14, but fake "restore your messages" emails are now trying to steal passwords3,15.
- Businesses on Microsoft 365 can claim a service credit until December 31; personal plans aren't covered5,13.
Should you care?
3/5 Over now, but watch for scams
How hot right now
WarmHeavy coverage November 12 and 13; Microsoft's final report is due November 20
Fire meter checked Nov 19, 2026
Hype check
We'll check back on Feb 10, 2027: did it live up to the hype?
Who is this for?
- ✓People who use Outlook or Teams for work or schoolFor youWork and school mailboxes on Microsoft 365 were the ones hit hardest, in some cases for more than 18 hours4.
- ✓Small business owners on Microsoft 365For youYou can ask Microsoft for a partial refund on your bill, but only if you claim it by December 315,13.
- ~Personal Outlook.com and Hotmail usersMaybeSome individual users reported login trouble, though Microsoft's consumer Outlook service was listed as working, so reports conflict17,18.
- ~Microsoft 365 Personal and Family subscribersMaybeYou aren't owed a credit, because consumer plans don't come with an uptime guarantee13.
- ✓Anyone with an email addressFor youOutage-themed phishing emails are going out, offering to "restore" messages so they can grab your password15,21.
Outlook email went dark for a day. Here's what happened and the scam to watch for
A bug at Microsoft stopped Outlook email working for many people from November 12 until the next day. It's fixed, but watch out for fake emails about it.
On Thursday, November 12, a lot of people opened Outlook and found nothing. No new email, or no inbox at all. Microsoft's work and school email service went down around the world, with problem reports from 74 countries2. Teams, the video-call and chat app many offices use, had trouble too6.
How bad was it?
For some people, email was gone for more than 18 hours4. In the UK, some National Health Service staff went back to paper records, and banks in Taiwan were hit hard3. One research firm guessed the lost work could cost at least $7 billion, though that is an estimate, not a measured number16.
Microsoft says everything was fully working again by the morning of Saturday, November 143.
What went wrong?
It wasn't hackers4. Think of Microsoft's email system as a big post office with a few front doors. In July, Microsoft installed a small fix on those doors that had a hidden flaw. On November 12, extra traffic got sent through the same doors, and the flaw made them jam5,10. Most of the doors stopped opening at once, and fixing them took about a day11.
Microsoft apologized. One of its top executives wrote, "We are deeply sorry for the impact to all our customers"5.
What it means for you
If your email is working now, you are most likely fine: delayed messages trickled in as service came back4.
The thing to watch is scams. Crooks are sending fake emails that say your messages got "stuck" in the outage and will be deleted in 24 hours unless you click a link15. The link leads to a fake Microsoft login page that steals your password15. Don't click. If you're worried, open Outlook yourself instead of using the link.
If you run a small business that pays for Microsoft 365Jargon busterMicrosoft's subscription bundle of Outlook, Word, Excel, Teams and online storage, sold to businesses, schools and households., you can ask Microsoft for a partial refund on your bill before December 315. People on personal or family plans aren't covered13.
What should you do?
Three things went wrong at once, and who gets money back
Microsoft says a months-old bug, a Teams traffic shift and servers already offline combined to cause it. Office Watch reckons business customers can get back at most half a month's Exchange fee.
The first warning came shortly before 1 p.m. ET, when Microsoft flagged Exchange OnlineJargon busterThe Microsoft service that actually stores and delivers work and school email behind the Outlook app. mailbox problems in the Americas7,19. More than 12 hours later, Microsoft said it had only "marginal success" restoring service19. DowndetectorJargon busterA website that counts how many people report that a service isn't working. logged more than 130,000 Outlook reports at the peak6, and SharePoint, OneDrive and Copilot had problems alongside Teams6,19.
How long it lasted
It depends on who you ask. Microsoft's own incident record for Exchange Online runs from 16:20 UTC on November 12 to 16:10 UTC on November 13, just under 24 hours20. Microsoft 365Jargon busterMicrosoft's subscription bundle of Outlook, Word, Excel, Teams and online storage, sold to businesses, schools and households. expert Tony Redmond says his own tenant was down for nearly 18 hours12. Other outlets have called it a 30-hour outage, so expect the numbers to keep moving13.
What Microsoft says broke
The "front doors" above are Microsoft 365's networking gateway servers. Charles Lamanna, Microsoft's executive vice president of Copilot, Agents and Platform, laid out three factors on November 1610:
- A latent defect in a fix rolled out to those gateways in July5.
- A Teams failover about 15 hours earlier, which moved Teams traffic onto the same primary gateway Exchange Online uses10.
- About 14% of gateway capacity already offline from an unrelated issue the day before10.
At the worst point, more than 80% of the gateway servers stopped responding11. Because the bug had been live for months, Microsoft couldn't simply undo the change8. It also said the outage wasn't related to AI agents4, and that ordinary Windows PCs aren't affected by the bug5.
Money back, for some
The business refund comes as a service creditJargon busterA discount on a future bill that a company gives when it breaks its uptime promise., claimed from the Microsoft 365 admin center5,13. Credits run 25%, 50% or 100% of the affected service's monthly fee depending on uptime, and Office Watch reckons 50% of the Exchange fee is the most this incident can earn13. Personal and Family plans are left out because they don't have an uptime guarantee13.
Our take
The bigger story is concentration. When one company runs email for the NHS, Taiwanese banks and rail ticketing at once, one bad update has global reach3. Expect more businesses to ask about backup email and offline copies of critical documents.
What to watch
Microsoft's comprehensive list of fixes is due around November 209, and outage-themed phishingJargon busterFake emails or texts that pretend to be from a real company to trick you into giving away passwords or money. is still going14.
By the numbers
How we got here
- Jul 2026Microsoft deploys a fix to its Microsoft 365 gateway servers that contains a hidden defect5,8
- Nov 11, 2026About 14% of gateway capacity goes offline from an unrelated issue10
- Nov 12, 2026Teams traffic fails over onto the gateway Exchange uses; hours later Outlook mailboxes start failing worldwide10,19
- Nov 13, 2026Email returns for most users and Microsoft apologizes4
- Nov 14, 2026Microsoft declares full recovery3
- Nov 15, 2026Preliminary post-incident report blames the July defect8
- Nov 16, 2026Executive Charles Lamanna explains the three causes and says sorry10
- Nov 20, 2026 · coming upTarget date for Microsoft's full list of fixes9
- Dec 31, 2026 · coming upDeadline for business customers to claim service credits5,13
The bug, the slow recovery and the SLA math
Per Microsoft's preliminary report, a lock-contention bug exhausted CPU and memory under load, worse on high-core-count hardware that the load tests missed. Our math puts the outage in the 50% credit tier.
Microsoft's preliminary post-incident reportJargon busterA company's write-up after an outage explaining what went wrong and what it will change. came out on November 158, and Lamanna's three-factor summary on X the next day was an abridged version of it10.
Timeline (UTC)
| When | What |
|---|---|
| Jul 2026 | Fix deployed to M365 networking gateway servers, containing a latent defectJargon busterA bug that sits unnoticed in software until the right conditions trigger it.5,8 |
| Nov 11 | About 14% of gateway capacity out from an unrelated issue10 |
| Nov 12, ~01:20 | Teams fails over from its alternative gateway to the primary gateway shared with Exchange OnlineJargon busterThe Microsoft service that actually stores and delivers work and school email behind the Outlook app.10 |
| Nov 12, 16:20 | EX1550543 impact begins20 |
| Nov 13, 16:10 | EX1550543 impact ends, 23h50m total20 |
The 01:20 figure is our back-calculation from "about 15 hours before" the outage10.
The bug
Per The Register's reading of the PIR, the July change introduced lock contention inside an internal Windows process on the gateway hosts. Under certain load patterns, threads piled up waiting on the lock until CPU and memory were exhausted8. The contention scales with core count, so it was worse on newer, higher-core hardware. The update passed load testing because those tests ran mostly on older machines with fewer cores8.
That's a classic test-fleet mismatch. It also explains why the bug sat dormant for months: it needed the right mix of hardware, traffic shape and volume to trigger. The Teams failoverJargon busterAutomatically moving traffic from one system to a backup when the first one has trouble. supplied extra load and a changed traffic pattern, and the missing 14% of capacity removed headroom10. Lamanna's account is that none of the three would have done it alone10.
Why recovery was slow
The usual safety nets, flighting, feature flags and rollbackJargon busterUndoing a software update by going back to the previous version., all assume a recent change, and this one was months old5,8. Early status updates described Exchange servers failing health checks and dropping out, overloading the survivors, a cascading failure that looks like a capacity problem until you find the trigger19. Microsoft adds that Windows Server deployments aren't affected either; the fix was built specifically for the M365 gateways5.
SLA math
Exchange Online's financially backed SLAJargon busterService level agreement: a company's written promise about how reliably a paid service will work, with credits if it doesn't. pays 25% below 99.9% monthly uptime, 50% below 99% and 100% below 95%13. Our calculation: 23h50m out of November's 720 hours is about 96.7% uptime, which lands in the 50% tier. That matches Office Watch's estimate that 50% is the ceiling for EX155054313. Claims need the incident ID and tenant ID5,13. Consumer plans sit under the Microsoft Services Agreement, which has no uptime commitment13.
Phishing follow-on
Microsoft says the scam emails carry subject lines of "Outage" plus a six-character code15. Responders report adversary-in-the-middle relays that can capture sessions after MFA, so a password reset alone isn't enough: revoke sessions and review sign-in logs15,21.
Open questions
- The final repair-item list, targeted for November 209.
- Why regional and component isolation failed to contain a gateway fault.
- Whether Microsoft changes its load-test fleet to match production core counts.
The facts
Hover or tap anything underlined to see what it means.
- Company
- Microsoft5
- Started
- Thursday, November 12, 202619
- Fully recovered
- Morning of November 14, 20263
- Hardest-hit service
- Outlook and Exchange OnlineJargon busterThe Microsoft service that actually stores and delivers work and school email behind the Outlook app. mailboxes12
- Exchange incident length
- 23h50m (16:20 UTC Nov 12 to 16:10 UTC Nov 13)20
- Reach
- Users in 74 countries2
- Peak Downdetector reports
- 130,000+ for Outlook6
- Cause
- Latent defectJargon busterA bug that sits unnoticed in software until the right conditions trigger it. in a July 2026 gateway serverJargon busterA computer that sits at the front door of a service and routes incoming traffic to the right place. fix, plus extra load and reduced capacity5,10
- Incident IDs
- EX1550543, MO155082113
- Consumer plans
- No SLAJargon busterService level agreement: a company's written promise about how reliably a paid service will work, with credits if it doesn't., no credit owed13
Jargon buster
- Microsoft 365
- Microsoft's subscription bundle of Outlook, Word, Excel, Teams and online storage, sold to businesses, schools and households.
- Exchange Online
- The Microsoft service that actually stores and delivers work and school email behind the Outlook app.
- Gateway server
- A computer that sits at the front door of a service and routes incoming traffic to the right place.
- Failover
- Automatically moving traffic from one system to a backup when the first one has trouble.
- Latent defect
- A bug that sits unnoticed in software until the right conditions trigger it.
- Lock contention
- When many parts of a program wait in line for the same shared resource, so work piles up and slows everything down.
- Rollback
- Undoing a software update by going back to the previous version.
- Downdetector
- A website that counts how many people report that a service isn't working.
- SLA
- Service level agreement: a company's written promise about how reliably a paid service will work, with credits if it doesn't.
- Service credit
- A discount on a future bill that a company gives when it breaks its uptime promise.
- Phishing
- Fake emails or texts that pretend to be from a real company to trick you into giving away passwords or money.
- Post-incident report
- A company's write-up after an outage explaining what went wrong and what it will change.
America's week of outages
Microsoft, the FAA, Chase, United, Apple and an NSA network all failed within a week. Officials say it wasn't an attack; Congress wants answers.
- Nov 12, 2026Microsoft's Outlook and Teams go down for up to a day You're reading this
- Nov 13, 2026A problem at the FAA grounds flights across the US The Register
- Nov 16, 2026A software bug disrupts Chase banking, and a tech issue delays United departures SecurityWeek (Associated Press)
- Nov 17, 2026Apple's iCloud, App Store and Apple Pay stumble after an internal update; an NSA network goes down TechCrunch
- Nov 18, 2026Officials tell Congress the outages weren't coordinated; two senators press AI chief Jay Clayton Our story →
- The NSA saying its internal network is fully restored
- Classified briefings for Congress on the outages
- Post-incident reports from Chase, Apple or United on what broke
- A White House reply to the Cantwell-Rounds letter
- The Super Intelligence Force's 120-day AI report, due early 2027
Outlook's rough November
Two separate problems hit Outlook within days: a security hole hackers were already using, then a day-long outage.
- Jul 2026Microsoft rolls out a server change with a hidden defect that later causes the outage ABC News
- Nov 10, 2026Microsoft's monthly update fixes an Outlook bug hackers were already using Our story →
- Nov 12, 2026Outlook goes down for up to a day, in 74 countries You're reading this
- Nov 14, 2026Microsoft declares a full recovery Al Jazeera
- Nov 15, 2026Microsoft's first report blames the July defect The Register
- Nov 20, 2026 · coming upMicrosoft's full list of fixes is due The Next Web
- Dec 31, 2026 · coming upLast day for business customers to claim service credits ABC News
- Microsoft's full list of fixes after the outage
- Any new Outlook security problems
Sources21 pages · Al Jazeera, ABC News, TechRepublic and moreShowHide
The small numbers in the text point to these pages. Everything is written in our own words. How we work · Spotted a mistake? Tell us
- 1Microsoft working 'urgently' to fix outages · Al Jazeera
- 2Microsoft hit by global outages: What happened and why? · Al Jazeera
- 3Microsoft says affected services fully recovered following worldwide outage · Al Jazeera
- 4Outlook outage updates: Microsoft apologizes for email disruption, says there was no hack · ABC News
- 5Microsoft identifies issues leading to global Outlook outage · ABC News
- 6Microsoft Outlook Down: Over 130,000 Outage Reports as Email, Teams Disrupted · TechRepublic
- 7Live updates: Microsoft outage hits Outlook, Teams across the world · NBC News
- 8Microsoft blames monster email outage on software defect introduced in July fix · The Register
- 9Microsoft blames its Microsoft 365 outage on a dormant July update defect · The Next Web
- 10Microsoft Apologises for Massive Global Outage and Reveals Three Factors Behind the Widespread Disruption · International Business Times UK
- 11Microsoft traces its 365 outage to a Windows bug hidden since July · Channel Dive
- 12The Great Exchange Online Outage of November 2026 · Office 365 for IT Pros
- 1330 hour Microsoft 365 outage: what you can claim and how · Office Watch
- 14Microsoft's Outlook outage is over. The scams aren't · PCWorld
- 15Microsoft Warns of Active Password Stealing Attack Exploiting the Microsoft 365 Service Disruption · Cyber Security News
- 16Microsoft outage could lead to $7B in losses · Security Magazine
- 17Outlook Outage: Microsoft's Email Service Goes Down Nationwide · Variety
- 18Is Outlook Email Down? Is Hotmail Down? News on the Microsoft Outlook Outage · Men's Journal
- 19Is Outlook Down? Microsoft Meltdown Leaves Millions Unable to Open Inboxes · International Business Times UK
- 20Microsoft 365 outage 12 November 2026: post-incident report · 365fanatic
- 21Microsoft 365 outage theme drives credential phishing and support scams · Field Effect
Read next
Internet & apps · ★ Saved
Chase, United, Apple and an NSA network all went down this week. Officials say it wasn't an attack
Chase, United, Apple and an NSA network all had outages on November 16 and 17. Each blamed a bug or an upgrade, and Chase, United and Apple are working again.
Internet & apps · ★ Saved
Hulu moves fully into Disney+ in December: what changes for subscribers
If you pay for the Disney+ and Hulu bundle, everything Hulu offers, live TV included, will work inside the Disney+ app from December. No new prices, and the Hulu app isn't going away yet.
Internet & apps · ★ Saved
Instagram now lets you fix a Story after you post it, for 15 minutes
Instagram finally lets you change a Story after posting it, as long as you do it within 15 minutes. The Story keeps the views and replies it already has.