1. Microsoft outage stretches into Friday
A Microsoft outage centered on Exchange Online, the service that stores Microsoft 365 mailboxes and calendars, carried into Friday, November 13, after starting on Thursday. Users worldwide were locked out of Outlook email and could not schedule Teams meetings. Microsoft said on Friday that a patch was steadying some servers and that it was rolling the fix out more widely, but it gave no date for full recovery.
The disruption was not limited to a single app. Because Exchange Online sits underneath so much of the Microsoft 365 suite, the failure spread to features in Microsoft Copilot, Microsoft Defender for Office 365 and the Teams calendar, and to outside software that reads work email and calendars through Microsoft's systems. AI assistants from OpenAI and Anthropic, both of which offer Microsoft 365 connectors, reported problems as well.
Government offices were among those affected, and some could not send or receive email while the incident lasted. For a platform that many large organizations treat as basic infrastructure, an outage running past one business day is unusual, and administrators were left waiting on Microsoft's status updates for signs of progress.
2. What caused the Microsoft outage?
Microsoft traced the failure to an overloaded process on the affected servers, the component responsible for setting up secure connections. Once that process was overwhelmed, the servers stopped handling requests, so users could not reach their Exchange Online mailboxes through Outlook or other clients.
The explanation points to a capacity and connection-handling problem rather than lost data. Every time a mail client, a phone, a browser session or a third-party app talks to Exchange Online, it first has to open an encrypted connection with the server. If the software doing that work cannot keep up, the server behind it may be healthy but is effectively unreachable. To the people trying to use it, the result looks the same as the service being switched off.
Microsoft also said it found no indication that a security breach was behind the outage. That distinction matters to the organizations involved. A failure caused by an attack raises questions about whether data was accessed or changed, and it can trigger disclosure duties. A failure caused by infrastructure strain is disruptive, but it does not on its own put mailbox contents at risk.
Microsoft has not yet published a full root-cause analysis. It typically releases a post-incident report once an incident of this size is resolved, and that review should explain why the connection process became overloaded in the first place and why the problem spread across regions.
3. How Microsoft has tried to restore service
Microsoft's engineers tried several fixes as the incident went on. They restarted affected infrastructure, rerouted network traffic away from struggling servers and added server capacity to absorb the load. Each of these is a standard step for a service under strain, but none of them fully solved the problem, which is why the outage carried over into a second day.
By Friday the company's focus had moved to a software patch. Microsoft said the change had made some servers more stable, and it began deploying the fix across more of its infrastructure. Rolling out gradually is normal for a service at this scale: pushing an untested change everywhere at once risks making things worse, so providers usually confirm that a fix works on a slice of servers before widening it.
The downside is that recovery is uneven. Some organizations may see mail come back while others, served by different parts of Microsoft's infrastructure, stay offline. Until the patch reaches every affected server and Microsoft declares the incident mitigated, administrators have little choice but to track the official status updates and keep users informed.
4. What is the Microsoft outage affecting?
Outlook email, delivered through Exchange Online, took the biggest hit, with some users unable to open their mailboxes at all. Teams was affected more narrowly. Users had trouble scheduling meetings and sending invitations, because those features depend on the Exchange calendar, while Microsoft said most Teams calls and chats kept working.
The outage also reached Microsoft Copilot, the company's AI assistant built into Microsoft 365. Copilot features that draw on a user's email, calendar or files rely on the same back-end services, so when Exchange Online went down, those features lost the data they need. Microsoft Defender for Office 365, the security product that scans email and links for threats, was also affected.
That spread shows how tightly the suite is linked. Microsoft 365 is sold as a bundle of separate apps, but many of them depend on the same identity, mail and calendar services underneath. A failure in one of those shared layers can surface as problems in products that, from the user's side, look unrelated.
5. Why so many organizations felt the impact
Microsoft 365 is one of the most widely used workplace platforms in the world. Microsoft has reported more than 450 million paid commercial seats for the suite, covering businesses, schools, nonprofits and government agencies that rely on it for email, scheduling, documents and internal communication.
For most of those customers, Exchange Online is where their mailboxes and calendars live. Organizations that once ran their own email servers have largely moved to Microsoft's cloud service, handing over maintenance, security updates and capacity planning. The trade-off is that when Microsoft's service has a problem, thousands of organizations lose email at the same moment, and there is little any one customer can do to route around it.
That concentration has been debated for years in IT and policy circles. Supporters of large cloud platforms say they are generally more reliable and better defended than systems individual organizations could run on their own. Critics say that placing so much of the economy's communication on a handful of providers creates single points of failure, and incidents like this one feed that argument.
6. OpenAI and Anthropic connectors disrupted
The outage also showed how much newer AI tools now lean on Microsoft's infrastructure. OpenAI reported errors with the Microsoft 365 connector that lets ChatGPT read a user's work email and calendar. Anthropic likewise reported problems with Claude's Microsoft 365 integration as requests to Microsoft's systems failed.
These connectors are a growing part of how businesses use AI assistants. Rather than pasting information into a chat window, a user can link a work account so the assistant can search their inbox, summarize threads, check schedules or draft replies using real context. The assistant does not store that mail. It retrieves it from Exchange Online when asked, through Microsoft's programming interfaces.
That design means an assistant can only be as available as the services it connects to. ChatGPT and Claude kept running, but any task that needed live data from Microsoft 365 failed while Exchange Online was down. For companies building AI-assisted workflows around their email and calendars, the incident is a reminder that those workflows inherit the reliability of every system underneath them.
7. Government email caught in the disruption
Public sector bodies were among those affected. Government agencies in several countries use Microsoft 365 for their day-to-day communications, and some reported that their email systems were hit, including government systems in Japan that could not receive messages on Friday.
For government offices, losing email for even part of a working day can delay filings, correspondence with the public and coordination between departments. Many agencies have backup procedures for urgent communications, but routine work that depends on email and shared calendars usually has to wait until service returns.
The public sector impact will probably add to an ongoing debate over how much governments should rely on a single commercial provider for core communications. Several governments have looked at ways to diversify their technology suppliers or keep sovereign alternatives, and outages that cross borders tend to bring those discussions back.
8. Why the outage matters for crypto and finance teams
Firms in digital assets and traditional finance run on the same productivity tools as everyone else. Trading desks, compliance teams, fund administrators and exchange support staff depend on email and calendars for client communication, counterparty coordination and internal approvals. A day without reliable access to Outlook or Teams scheduling slows those processes even when trading systems themselves are unaffected.
There is also a security angle. Outages create openings for phishing, as attackers may send messages posing as IT departments or Microsoft support, telling users to sign in again or install fixes. Microsoft said it saw no sign of a breach behind the incident, but security teams generally advise staff to treat unexpected password-reset or login requests with extra suspicion during and right after a major service disruption.
The incident's effect on AI connectors is relevant to the industry as well. Crypto firms have been quick to adopt AI assistants for research, reporting and customer operations, and many connect those tools to Microsoft 365. Firms that do may want to check which of their processes would stop if a single upstream provider went offline, and whether those processes have a manual fallback.
9. What happens next
The immediate priority is finishing the patch rollout. Microsoft said the fix was improving stability on some servers and that it was extending it more widely, but it did not say when it expected every affected mailbox to be back. Administrators should keep watching the Microsoft 365 service health dashboard for confirmation that the incident is fully mitigated.
Once service is restored, the next step is Microsoft's post-incident review. That document should cover why the secure-connection process was overwhelmed, why restarts, traffic rerouting and extra capacity did not fix it, and what Microsoft plans to change to prevent a repeat. Large enterprise and government customers are likely to press for those answers, and some may raise the incident in conversations about service-level commitments.
For OpenAI, Anthropic and other companies whose products depend on Microsoft 365 data, recovery should follow automatically once Exchange Online is stable, since their connectors draw directly on Microsoft's service. Until then, users of those integrations may keep seeing errors on tasks that involve work email, calendars or search.