Quick Answer: A Microsoft 365 outage starting Thursday, November 12, 2026 disrupted Outlook, Teams, and Copilot worldwide. Microsoft said a technical issue during a network capacity upgrade, following a recent network change, caused it, not a cyberattack. Services were fully restored by Saturday, November 14, after hours of gradual recovery.
A global Microsoft 365 outage took out email, chat, and AI tools for users in dozens of countries, and it stayed in the headlines well after the lights came back on. Downdetector logged reports from more than 70 countries and territories. Courts paused trials, police asked people to phone in crime reports, and some prescriptions were delayed. If you run operations on Microsoft's stack, this event is a useful stress test. Here's what actually happened and what you can do before the next one.
What Happened, in Order
Here's the sequence, as Microsoft's public updates and the Associated Press reported it:
- Thursday, midday U.S. time: Problems began appearing across Microsoft 365 services, including Outlook, Teams, and Copilot.
- Within the hour: Downdetector reported outage reports from 74 countries and territories.
- Friday: Microsoft apologized and said the incident was "not related to cyber" or AI agents. It traced the root cause to an issue during an earlier network capacity upgrade, following a recent network change.
- Friday into Saturday: Engineers restored services gradually, which took longer than a single fix would have. Microsoft said it had addressed the underlying issue and restored remaining regions.
The gap between step 3 and step 4 is where most of the pain lived. Recovery was not a switch flip. It was a region-by-region process, and users in some places stayed affected well after the headline "resolved" status appeared.
Why a Network Change Became a Global Outage
Microsoft's explanation points to a network change that interacted badly with a capacity upgrade. In a large cloud environment, network changes rarely affect one thing. Routing, authentication, and mailbox availability all depend on shared paths, so a fault in one layer can make many services look broken at once.
Microsoft also said that a failure caused a number of Exchange Online servers to become "unexpectedly unavailable" during automated checks. That detail matters. Automated health checks are designed to protect the system, but when they remove servers from service during a broader fault, they can deepen the impact. This is the cascade effect the Associated Press described: one provider, many dependent services, and few easy alternatives.
Here's the cause-and-effect chain in plain terms. When a network change degrades capacity, dependent services time out, mailboxes become unreachable, and clients retry. Those retries add load to the recovery path, so engineers can't simply restart everything. That's why the company's own posts suggested its team struggled to get the problem under control for some time.
Why the "Cyberattack" Assumption Misleads Response Teams
Here's the counter-intuitive part. When a major platform goes dark, the reflex in many security teams is to assume a breach. That reflex can be the wrong one, and acting on it can waste the first critical hours.
Microsoft was explicit that there was "no indication" of malicious activity. If your incident runbook jumps to isolating accounts, rotating credentials, or disabling integrations, you may make things worse. You also risk treating a vendor-side fault as an internal compromise.
The better first move is to check the provider's status channels, such as the Microsoft 365 Service Health dashboard and the Azure Service Health alerts, before treating the event as a security incident. Then run your security checks in parallel, not in sequence. You'll still catch a real breach, and you won't lose the morning chasing a ghost.
That said, there's a real caveat. Confirming a vendor outage doesn't rule out a separate incident happening at the same time. Verify, don't assume.
Who Got Hit, and How Far the Impact Reached
The disruption didn't stay inside software. The Associated Press reported effects across several sectors, and the table below summarizes what was documented in the source material:
| Sector | Reported impact | Practical exposure |
|---|---|---|
| Courts | A Manchester trial was halted; the judge sent jurors home | Scheduling and evidence handling depend on reliable email |
| Police and law enforcement | UK police could not receive online crime reports; Arapahoe County, Colorado, saw online bond payments affected | Public-facing digital intake fails silently without a phone fallback |
| Healthcare | The NHS faced threats to emailed prescriptions; pharmacy waits were extended | Clinical workflows often assume email is always up |
| Government | Japan, Singapore, Australia, Greece, and Norway reported partial operational impact | Ministry-level email becomes a single channel for official business |
Notice the pattern. In every row, the failed system was not the core service. It was the communication layer wrapped around it. That's where most organizations are least prepared.
The Single Point of Failure Trade-Off
Microsoft 365 reportedly serves more than 430 million users, and Microsoft said in January it had more than 450 million paid commercial subscriptions. That scale is exactly why a single fault can ripple so widely. It's also why many organizations think "just add a second provider" is the cure.
The trade-off is real, and it cuts both ways:
- Single-vendor advantages: One identity system, one support contract, and simpler compliance mapping.
- Single-vendor risks: Correlated failure. When the provider has a bad day, everything fails together.
- Multi-vendor advantages: Failure isolation and a fallback path for critical mail.
- Multi-vendor costs: Duplicate licensing, split identity, more complex security controls, and a larger attack surface.
For most mid-sized organizations, the right answer isn't a full migration. It's a targeted fallback for the few functions that cannot stop, such as urgent notifications, on-call escalation, and critical regulatory communications.
How to Prepare Before the Next Outage
Here's the sequence most practitioners use to reduce exposure without rebuilding their stack:
- Map your critical communications. List the five workflows that cannot pause, such as prescriptions, court filings, or payment intake.
- Create an out-of-band channel. Use a phone tree, SMS provider, or a secondary messaging tool that doesn't share the primary identity system.
- Publish a status fallback. Post a holding page or phone line that customers can reach during an email outage.
- Plan for delayed mail, not lost mail. Mail senders generally retry deferred messages over an extended period under SMTP, so many "lost" emails arrive late. Tell staff this before they resend, which can create duplicates.
- Rehearse the outage. Run a tabletop exercise once a year with a scenario in which your primary email and chat are down for four hours.
Each step is cheap compared with the cost of a missed prescription or a halted hearing. The goal is continuity of the few things that matter, not perfect redundancy across everything.
Frequently Asked Questions
What is a Microsoft 365 outage?
A Microsoft 365 outage is a period when Microsoft's cloud productivity services, such as Outlook, Teams, and Copilot, are unavailable or degraded for some or all users. Microsoft typically attributes the cause to internal infrastructure issues, and recovery is often gradual across regions. Status updates usually appear on the Microsoft 365 Service Health dashboard and on Microsoft's X account.
Why did Outlook stop working during the outage?
Outlook depends on Exchange Online mailboxes and authentication services, both of which were affected. Microsoft said a number of Exchange Online servers became unexpectedly unavailable during automated checks, so users couldn't send, receive, or load their inboxes. Mail that couldn't be delivered generally retries from the sending server, so some messages arrived late.
How to prepare for a Microsoft 365 outage?
Start by listing the workflows that cannot pause, then give each one a fallback that doesn't depend on Microsoft 365, such as phone lines or SMS. Rehearse that fallback with staff at least once a year. Subscribe to vendor status alerts so you can confirm a provider outage before triggering a security response.
Was the Microsoft 365 outage caused by a cyberattack?
Microsoft said the outage was "not related to cyber" or artificial intelligence agents and that there was "no indication" of malicious activity. It attributed the failure to an issue during a network capacity upgrade that followed a recent network change. Because a vendor outage and a security incident can look similar at first, verify the provider's status before assuming a breach.
Keep Your Plan Simple and Test It
The most useful takeaway is this: identify the few communications your operation cannot lose, then give each one a fallback that doesn't share a failure point with your primary platform. Try running a four-hour tabletop exercise this month where email and chat are down, and note which decisions stall first. Pass this to whoever owns your business continuity plan, especially if they're still treating vendor outages as a security problem.