On a Thursday evening in November, a widespread failure in Microsoft's cloud services left people around the world unable to reach their email, calendars and shared work tools. It is the kind of quiet fragility modern life rests on: a few servers drop out, the systems that remain strain under the extra weight, and suddenly colleagues in distant countries cannot see each other's schedules or files. Microsoft describes an infrastructure failure rather than an attack, and its engineers were still trying to steady the platform as the day wore on, aware that a recovery can look complete before it t
Microsoft Struggles to Restore Global Outage Hitting Outlook, Teams and Microsoft 365 Users
On a Thursday evening in November, a widespread failure in Microsoft's cloud services left people around the world unable to reach their email, calendars and shared work tools. It is the kind of quiet fragility modern life rests on: a few servers drop out, the systems that remain strain under the extra weight, and suddenly colleagues in distant countries cannot see each other's schedules or files. Microsoft describes an infrastructure failure rather than an attack, and its engineers were still trying to steady the platform as the day wore on, aware that a recovery can look complete before it truly holds.
By around 18:00 GMT on Thursday, reports of trouble with Outlook had climbed to nearly 140,000 people worldwide, according to the monitoring site Downdetector. It was the high point of a global outage at Microsoft, and the company spent the evening working to bring its services back, still at it when the reports were filed.
The disruption spread across Microsoft 365, the subscription suite many businesses use for Outlook, Teams and OneDrive. For organisations working across borders, these tools are often the main way colleagues stay in touch. Microsoft warned that some people could not open certain files or view details in their calendars.
In a series of posts on X, the company blamed a number of servers being unexpectedly unavailable, which left the remaining systems overloaded. It reported positive results from its recovery work and said it was moving quickly, thanking customers for their patience.
The recovery has not been smooth. Some services that returned have become unresponsive again, and Microsoft has cautioned that users in any region could still be affected. Reports came in from the United States, Japan and Singapore, a sign of how far the failure reached.
The BBC understands there is no indication a cyber-attack caused the outage. The explanation points instead to infrastructure failure, with servers dropping out and remaining capacity struggling under the load.
The consequences extend past Microsoft's own customers. Companies that depend on its systems have been caught up too, including the AI developers Anthropic and OpenAI. OpenAI, which makes ChatGPT, said it was looking into a problem with its connections to Microsoft's services that had stopped some of its tools from working properly.
The figures are user reports, not confirmed counts from Microsoft, and they suggest tens of thousands affected. What comes next is whether Microsoft restores its services fully and stably, and whether OpenAI and Anthropic recover once Microsoft's systems settle.
By around 18:00 GMT on Thursday, reports of trouble with Outlook had climbed to nearly 140,000 worldwide, according to the monitoring site Downdetector. That was the peak of a global outage at Microsoft, which spent the evening trying to bring its services back and was still working on it.
The disruption reached across Microsoft 365, the subscription platform many businesses use for workplace tools including Outlook, Teams and OneDrive. For international organisations, it is a main channel for connecting and working with colleagues in other countries. Microsoft told users that among the problems, some people might be unable to view certain files or details in their calendar apps.
In a series of updates on X, Microsoft said the cause was a number of servers "being unexpectedly unavailable", which put strain on the systems that remained running. The company said it was seeing "positive results" from its recovery efforts and was rushing to fix the problem. A spokesperson said Microsoft appreciated customers' patience while it worked to restore services as soon as possible.
The recovery was not clean, though. Microsoft later warned that some services that had come back could become unresponsive again, and that customers in any region could be affected. Users in the United States, Japan and Singapore were among those reporting disruption, which gives a sense of how widely it was felt.
The BBC understands there is no indication that a cyber-attack caused the outage. The explanation Microsoft has given points to an infrastructure failure: servers dropping out unexpectedly and the remaining capacity struggling under the load.
The effects spread beyond Microsoft's own customers. Companies that work with its systems were hit as well, including the AI developers Anthropic and OpenAI. OpenAI, the maker of ChatGPT, said it was investigating a problem with its connections to Microsoft's services, which had caused some of its tools to stop working as expected.
The reported numbers are tens of thousands of affected users, with the Downdetector peak near 140,000 for Outlook alone. Those figures come from user reports rather than from Microsoft. What to watch now is whether Microsoft achieves full, stable restoration, given its own warning that recovered services may falter again, and whether the knock-on problems at OpenAI and Anthropic clear up once Microsoft's systems settle.
Want the full story?
Read the original at BBC ↗