Microsoft 365 and Outlook Service Outages

Nov 13 · Microsoft 365 outage enters second day Updated 07:15

On Thursday, November 12, 2026, an outage affected Microsoft 365 business and enterprise services, including SharePoint Online, Microsoft Defender XDR, Word, Excel, and OneDrive. Users experienced general connection delays and an inability to access mailboxes or send and receive messages; the Washington Examiner reported that the outage impacted both the mobile and website versions of Outlook. OpenAI and Anthropic reported that the outage affected connectors linking their AI tools to Microsoft services. Microsoft attributed connectivity errors to elevated load across multiple regions and an increase in unavailable servers, though the company stated the outage was not due to increased traffic and there is currently no indication of a cyberattack. Microsoft identified a specific component responsible for processing server health as the likely cause of failed health checks. In response, Microsoft paused deployments, redistributed traffic, and initiated a restart of impacted servers. Microsoft stated that restarting servers and validating alternative health check methods had not yet provided the anticipated relief, and the company is isolating portions of its infrastructure and analyzing traffic patterns to restore stability. NPR reported that Microsoft stated it was testing a fix and rolling out alternative health check methods to restore servers. Microsoft later reported 'marginal success' in its recovery efforts. Engineering professor Mark Gregory described the event as a cascading problem likely caused by a network layer misconfiguration. The Washington Examiner reported that reports of issues peaked just before 1:30 p.m., while WIRED reported that Downdetector reports began to fall by 1:15 p.m. ET. The Guardian reported that the outage impacted financial traders in Singapore, Indian airlines, and Taiwanese securities firms. Several Australian government departments issued notices advising clients to use alternative communication methods. The Guardian reported that experts in Australia warned the economic cost of the outage could reach billions of dollars, particularly for small businesses without backup plans, and that the Australian Computer Society called for a government review of Australia's reliance on single foreign technology providers. CNN reported the outage disrupted tools across 74 countries and territories, affecting banks, city governments in the US, and courts in the Philippines. In the UK, CNN reported the outage hindered the National Health Service's ability to process drug prescriptions, and banks issued warnings regarding potential phishing scams. The NY Post reported the outage lasted approximately 12 hours, affected thousands of users in the Asia-Pacific region, and impacted government agencies in Japan and the US. The NY Post further reported that electric truck maker Workhorse Group rescheduled its third-quarter results report from Nov. 12 to Nov. 16 and that this event is the second major disruption to Microsoft 365 this year, following an outage in January. Newsweek reported that the Arapahoe County Sheriff’s Office in the U.S. required in-person bond payments, the Tenth Circuit Court of Appeal reported email delays, and the disruption affected airlines in South Africa, visa processing in New Zealand, and security agencies in Australia. Newsweek also reported operational delays in government systems in the Philippines and financial institutions in Hong Kong. Australian Home Affairs Minister Tony Burke stated that Microsoft likened the event to the 2024 CrowdStrike update failure. The outage entered its second day, causing disruptions to the US House of Representatives. Anthropic reported separate authentication issues with Claude Code.

The cause of the outage was not a lack of capacity, but a failure of the system designed to monitor server health. When the health check component failed, healthy servers were likely marked as unavailable, triggering a chain reaction that overloaded the remaining infrastructure.

The Guardian provided necessary technical depth by reporting that the outage affected connectors linking AI tools from OpenAI and Anthropic to Microsoft services. This reveals that the failure extended beyond internal productivity tools to disrupt the broader AI ecosystem relying on Microsoft's infrastructure.

CNN and The Guardian engaged with the systemic risk of single-provider reliance, while Euronews and Washington Examiner remained superficial. These outlets focused on repeating Downdetector figures and general service interruptions, treating the event as a temporary glitch rather than an architectural vulnerability.

NY Post provided a concrete example of corporate disruption by reporting that Workhorse Group rescheduled its third-quarter results report specifically because of the outage. This detail moves the conversation from general "inconvenience" to measurable business impact.