Microsoft Outlook outage continues for nearly 16 hours, users still unable to send or receive email
Microsoft 365 service disruptions have continued for more than 15 hours. Outlook (Exchange Online) users are still widely unable to access mailboxes or send and receive email, and services such as Teams, Copilot and SharePoint Online have been affected as well. The outage began at about 1 p.m. ET on November 12 (UTC 16:17) and quickly spread to at least 74 countries and regions worldwide.
Microsoft confirmed in a status update that the repair measures have only achieved partial, regional recovery, but have not led to sustained improvement. The company said that the process responsible for establishing secure connections on the relevant servers became overloaded, causing the servers to stop processing requests, and explicitly ruled out the possibility of a network attack. At its peak, Downdetector received nearly 140,000 reports of Outlook problems, affecting banks, government agencies, courts and airlines, among others.
Microsoft has not given a timeline for full recovery. Engineers have continued adjusting traffic routing, restarting affected infrastructure and expanding configuration changes, but multiple updates have said that the measures did not work as expected or did not achieve sustained recovery. Some Exchange-dependent functions, including calendars and meeting scheduling, remain disrupted, while Teams core messaging and online meetings are generally available.
Original source: cnn.com/2026/11/12/business/live-news/microsoft-global-outage-intl-hnk
ABAB AI Insight
The core of the outage lies in a capacity collapse on Exchange Online front-end servers. Overload of a single service can trigger cascading failures because Teams, Copilot, calendars and other services are highly dependent on the same mail and identity infrastructure, with pricing power and availability both locked into Microsoft’s cloud control layer.
The pattern of lasting more than 15 hours and repairs that cannot be sustained exposes a recovery bottleneck in centralized cloud architecture: restarting and rerouting traffic can only temporarily divert load and cannot quickly solve underlying process overload. For a historical comparison, the chain outages after the 2024 CrowdStrike incident can be referenced; key-person and key-process risks also exist at the infrastructure layer.
The simultaneous impact across 74 countries has forced enterprises and government agencies to interrupt email and collaboration channels. In terms of capital flows, this will push some organizations to accelerate multi-cloud or local backup strategies and create pressure to renegotiate Microsoft 365 SLAs and pricing; the short-term beneficiaries are alternative collaboration tool providers and local email solution providers.
Conclusion: this is a verified infrastructure-level outage, not a security incident. The duration and the statement that recovery could not be sustained have already entered the market’s pricing of cloud concentration risk, but the full root cause and final losses still await Microsoft’s formal postmortem.
ABAB News · Cognitive Law
Overload triggers cascading failure: the collapse of a single front-end process can lock up all downstream dependent services.
Partial recovery does not equal availability: if temporary improvement cannot be sustained, availability is still priced as an outage.
Concentration is a recovery bottleneck: the more concentrated the control layer, the more limited the repair path is by the same infrastructure.