A hidden software bug was behind the outage
According to Microsoft’s preliminary incident report (PDF), the issues were caused by an unknown bug that had been present since July 2026, Cyber Security News reports.
The flaw was found in a Windows process on the company’s network servers, causing them to become overloaded under certain conditions. The processor and RAM were put under increasing strain until the servers became unresponsive.
Microsoft claims that the flaw was not detected during testing because those tests were primarily conducted on less powerful servers. The problem was exacerbated on modern hardware with multiple processor cores.
The bug was specific to Microsoft’s server infrastructure and should not affect standard Windows computers.
Two events triggered it
Approximately 15 hours before the outage, Teams traffic was automatically routed to the same network infrastructure that handles Exchange Online. At the same time, almost 14 percent of server capacity was unavailable. The combination of increased load and reduced capacity triggered the previously undetected software failure.
When the servers stopped responding, Microsoft’s automated system attempted to redirect traffic to working servers. However, these too became overloaded, creating a chain reaction.