Microsoft 365: “And half the world administration stood still for twelve hours”
A configuration error in Microsoft 365 led to massive outages for companies, governments and financial institutions worldwide.
Quick Look
- A massive disruption to Microsoft 365 services impacted businesses, banks and government agencies worldwide.
- Microsoft denied a hacker attack and cited server overload following a configuration error as the cause.
AI-generated summary
Why It Matters
Microsoft 365 is a widely used office software suite used by millions of companies and governments worldwide. The outage affected core features such as email, Teams and AI assistants.
Which services are disrupted
Above all, it is about programs from Microsoft 365, the subscription package for the group's office software: the email program Outlook - both in the app and in the browser, the chat and telephony program Teams, the AI assistant Co-Pilot, the security protection Defender. Users can see which Microsoft services are currently affected and how here.
Who is affected
It affected a good part of the economic world. Microsoft 365 sold more than 450 million licenses this year (one license does not exactly correspond to one user, but the number of users is also likely to be in the high millions). The software is used particularly by companies, but also by administrations and other organizations.
According to Bloomberg, stock market traders and banks from Frankfurt to Singapore had problems completing deals without their email and chat access. At the European stock exchange operator Euronext, employees had to pass on important company information over the phone instead of by email as usual. Euronext operates the stock exchanges in Paris, Brussels, Lisbon, Amsterdam, Oslo, Athens and Milan, among others.
Open AI and Anthropic, the leading developers of artificial intelligence, examined problems with their interfaces to Microsoft services. The email support of several Indian airlines was affected.
The US truck manufacturer Workhorse postponed the publication of its quarterly figures from Thursday evening US Eastern Time to Monday. The luxury goods group Richemont, which includes Cartier and Piaget, also postponed the announcement of its figures on Friday morning.
The disruption also affected governments. Japan's Ministry of Finance no longer had access to email. The government IT system was down. Security authorities are also affected. In Australia, the secret service and federal police reported failures. The government there even activated an agreement with Microsoft to ensure that its systems would be restored as a priority.
The federal administration in Germany is “capable of working and essentially not affected by the failure,” said the Federal Office for Information Security. The office generally recommended reducing dependence on Microsoft: “As BSI, we recommend that business-critical services be provided with alternatives if possible.”
What caused the problem?
Microsoft denies that this is a hacker attack: “We are currently not aware of any malicious activity that triggered this incident.” Traffic was not higher than usual, but some servers were “unexpectedly unavailable.” The workload that was actually intended for these servers was redirected to other servers - which were then overwhelmed and apparently collapsed.
In a series of messages on Platform The first measures were unsuccessful, the affected servers were overwhelmed: “We are isolating part of the infrastructure and restarting the affected components.” From then on it would take hours until improvement occurred. It was only on Friday afternoon German time that it was said that progress was now being made in key applications, such as access to emails. Now it's about reducing the queue of backed up emails. Microsoft had previously extended the “email timeout” window to 48 hours. E-mails sent during the disruption but not delivered should still be delivered after the time window and not be lost.
Administrators who take care of company IT should follow Microsoft's statements closely if the company provides new information.
The incident shows the problem of Microsoft's power, says Harald Wehnes. He is a computer science professor at the University of Würzburg and spokesman for the "Digital Sovereignty" working group of the Society for Computer Science, and says: "The disruption was not a security incident. It was a sovereignty incident." A configuration error was enough – “and half the world administration stood still for twelve hours”. He points out that the federal government spent almost half a billion euros on Microsoft licenses last year. The money could also have gone to German companies. “The dependence on business and administration is worrying.”
In addition to the stress of technical recovery, Microsoft mocks. Many users on the Internet point out that just two weeks ago the company announced that it wanted to get rid of 60,000 employees with severance pay - more than a quarter of its employees. Tech companies are cutting jobs across the board and at the same time spending more and more money to build AI infrastructure. However, there is no evidence that a reduction in security professionals caused the disruption on Thursday.
In the main thread in which Microsoft posted news about the disruption on In any case, the company lost one day.
What to Watch
AI outlook — possibilities, not facts
Microsoft will continue to optimize infrastructure stability.
Likely · Within months
Open Questions
- How high is the overall economic damage?
- What specific technical measures prevent a recurrence?






