What Exactly Went Down?
On Thursday, November 12, 2026, users worldwide began reporting widespread issues with Microsoft's services. Core applications essential for daily business, including Outlook, Microsoft Teams, SharePoint, and even the Copilot AI assistant, became inaccessible
or unreliable. The problems, which started in the Americas, quickly cascaded into a global event, affecting users and organisations in at least 74 countries. The outage was severe, stretching for more than 17 hours for many, crippling communication and productivity for government agencies, hospitals, schools, and countless businesses.
The Key Reason: A Server Capacity Collapse
Initially, Microsoft reported that its infrastructure was experiencing an 'elevated load'. This led many to assume a massive surge in traffic was the culprit. However, the company later clarified the true source of the problem. The core issue wasn't an increase in demand, but a sudden decrease in supply. A large number of Microsoft's own servers had become 'unexpectedly unavailable'. This sudden loss of server capacity placed an immense strain on the remaining infrastructure, which buckled under the pressure of trying to handle the entire global workload. A core process responsible for setting up secure connections became overwhelmed, causing the remaining active servers to also stop responding to requests, creating a vicious cycle of failures.
The Domino Effect of a Single Failure
This event starkly illustrates the interconnected nature of modern cloud services. The outage was centered on Microsoft's Exchange Online service, the backbone for its email and calendar functions. Because so many other Microsoft 365 applications—from Teams to third-party AI tools from OpenAI and Anthropic that connect to Microsoft's ecosystem—rely on Exchange Online for authentication and data, they failed as well. The impact was immediate and widespread, with organisations unable to send emails, trading firms experiencing delays, and at least one US company having to postpone its quarterly earnings call. Some services in the UK and Japan even resorted to using phones and fax machines.
Microsoft's Scramble to Restore Service
Microsoft quickly ruled out malicious activity or a cyberattack as the cause. Engineers scrambled to identify the root of the server failures. Early attempts to simply restart servers or redirect traffic proved insufficient, as the restored servers would fail again once they began receiving traffic. The company's statement that its fixes were 'not taking as expected' highlighted the complexity of the problem. Eventually, Microsoft identified a recent network configuration change that correlated with the start of the outage and began rolling it back. It also applied a patch to an authentication and routing component, which led to 'encouraging recovery progress' as services were slowly and carefully brought back online region by region.
