Microsoft outage takes Outlook down across Europe as the NHS, Bank of England and Euronext scramble

For most of Thursday night and Friday morning, a large part of Europe's working world had no email.

A failure inside Microsoft 365 began at around 16:00 GMT on Thursday 12 November, when people first found they could not open Outlook. Within roughly an hour, the outage tracker Downdetector had complaints from 74 countries and territories. By Friday morning the problem was still there, and servers that Microsoft had brought back were dropping out again. On Friday Microsoft said it was seeing encouraging recovery in Europe and improvement in North America, while warning that some users could still run into problems on and off.

The company has not said how many customers were hit. Around 430 million people use its 365 services, and Downdetector logged close to 130,000 reports from the US alone at the worst point on Thursday. Those numbers only count people who bothered to report a fault, so the real figure is almost certainly much larger.

Email went dark, Teams mostly kept talking

The damage was uneven, and the pattern says a lot about how Microsoft's products fit together. Exchange Online, the mail system behind Outlook on desktop, web and phone, was at the centre of it. Many people could not reach their inbox at all. Others saw mail arrive hours late.

Teams calls, meetings and group chats largely kept running. What broke was everything in Teams and elsewhere that leans on the mailbox layer: calendar invites, moving or editing meetings, contact syncing and the Copilot assistant's answers about mail, calendars and files. The trouble spread past Microsoft's own apps too. OpenAI and Anthropic both said the connectors their AI tools use to reach Microsoft 365 data were affected.

To stop mail from being lost while servers came back, Microsoft stretched the window that incoming messages wait before they are rejected to 48 hours.

What Microsoft says went wrong

Microsoft has ruled out an attack. Charles Lamanna, an executive vice president at the company, wrote late on Thursday that "there is no indication this is the result of malicious or cyber activity."

The explanation so far is a technical one. Microsoft said a group of servers became unavailable without warning and failed their automated health checks. Their share of the work moved to the servers still standing, which then could not cope either. The company stressed that the strain was not caused by more people using its services. It pointed to a process that sets up secure connections, which became swamped and stopped the servers from handling requests, and it began undoing a recent change to its network settings whose timing lines up with the start of the trouble. It has not yet said that this change was the cause.

That chain helps explain why the first fixes did not hold. Each time engineers brought servers back, the backlog of waiting connections rushed at them and pushed them over again. Overnight, Microsoft admitted that changes it had applied were not working as planned. Early on Friday it rolled out a patch to its authentication and routing setup and reported better stability, then widened it to regions starting their working day. It is now draining, restarting and fixing affected machines in stages, and it has promised a full report once the incident is over.

Hospitals, markets and trains on the phone

In the UK the effects were felt in health care first. NHS mail runs on Microsoft, and several GPs said they could not see prescription requests or send referrals. Staff in parts of the health service went back to paper. The GMB union said its members in emergency call handling were struggling.

"Those who work in 111 and 999 services are unable to refer patient details in the normal way which is creating a massive backlog and has the potential for chaos," said Rachel Harrison, the union's national secretary. Emergency call-outs themselves were not hit, she added.

The financial system had its own problems. The Bank of England and the Financial Conduct Authority were both affected, and Euronext asked listed companies to phone in their corporate information instead of emailing it. In Germany, Deutsche Bahn said it was hard to send out timetables for trains on diverted routes. Schools across Britain asked parents to call about absences, several police forces told the public to ring rather than use online forms, and the UK Parliament put other ways of communicating in place so the day's sitting could go ahead. Heathrow, Manchester and Stansted airports said their operations carried on as normal.

Microsoft shares were down 2.6% in trading before the US market opened on Friday.

A third failure this year, and a sharper sovereignty debate

This is not a one-off for Microsoft in 2026. In January its 365 services went down across North America for close to ten hours, which it put down to heavy load during maintenance. In late August, thousands of Outlook users around the world were locked out of their mailboxes.

For Europe, the timing is awkward. Governments have spent the year talking about relying less on US cloud and software providers. France's prime minister, Sébastien Lecornu, said earlier this year that officials would drop Zoom and other American video tools in favour of French software. A day without email in hospitals, central banks and stock exchanges will give that camp a fresh example.

Not everyone thinks it changes much in the short term. Customers are locked into licences, and moving thousands of staff off Outlook is slow and costly. Walter Pasquarelli, an AI researcher affiliated with the University of Cambridge, doubted that any rival could step in.

"I don't think that there's any firm, at least not in Europe, that because of this outage could provide a credible and solid alternative," he said.

For founders and IT leads, the more practical lesson is about single points of failure. When mail, calendars and an AI assistant all sit on one service, losing that service for a day takes all three. A staff phone list, a second chat channel and a plan for a working day without email cost little to set up, and this week showed why they are worth having.