According to Sina Finance, Microsoft said its November 12 global outage was unrelated to cybersecurity, AI or AI agents, while X user @nik0p0l5 said a viral post claiming AI could not find a root password was a joke. The outage affected services including Outlook and Exchange for about 24 hours on November 12 and 13, impacting millions of accounts. On November 14, the user said the post was a prank; Microsoft executive Charles Lamanna also said the outage was neither a cyberattack nor caused by AI tools.
Microsoft’s preliminary report, released November 15, said a software defect, a surge in network traffic and reduced server capacity combined to cause the outage. A July code update to Microsoft 365 network gateway servers contained a latent flaw that triggered lock contention in a Windows process under certain workloads, consuming CPU and memory and leaving servers unresponsive. Differences between stress-test traffic and hardware configurations and those in production meant the flaw went undetected for more than three months.
The day before the outage, a separate incident put about 14% of Microsoft’s internal server capacity out of service. About 15 hours before the outage, Teams switched from a backup gateway to a primary gateway that also carried services including Exchange Online, sharply increasing traffic and changing its mix. As servers stopped responding, traffic shifted to other servers and added to their load; safeguards failed to contain the spread, which reached network gateway infrastructure worldwide. Microsoft said it would issue a final post-incident report this week and expected to provide a comprehensive list of remediation projects on November 20.