Compute#Cloud#Email#Office suite

Microsoft 365 outage ends after roughly two days, blamed on a network capacity upgrade

Microsoft 365 broke from Exchange Online outward on November 12, hitting Outlook, Teams and Copilot; recovery came November 14, blamed on a capacity upgrade.

Data center aisle lined with server racks under blue light

Microsoft 365, the productivity suite most of the corporate world runs on, spent roughly two days degraded this week. Exchange Online failed first, at 11:17 a.m. ET on November 12, and the failure spread within hours to Outlook, Teams, Copilot, SharePoint and Intune. Microsoft declared recovery on November 14 through its @MSFT365Status account, tracked in the admin center as incident MO1550821. By then Downdetector had logged more than 7.4 million problem reports worldwide, per Reuters as cited by Digital Trends, across 74 countries and territories per NBC News, which also cites Downdetector.

Key points

  • Timeline: Exchange Online mailbox access broke at 11:17 a.m. ET on November 12; a multi-product incident followed by early afternoon; on November 14 Microsoft declared the underlying issue fixed and remaining customers restored. The recovery post itself could not be opened directly for this write-up (the X page showed only the November 12 investigation posts); its wording is quoted by Digital Trends.
  • Scale: more than 7.4 million Downdetector reports across 74 countries and territories, peaking at about 19,650 reports per minute against fewer than 15 on a typical Thursday.
  • Cause: Microsoft says an issue “during an earlier network capacity upgrade that surfaced following a recent network change”; it found no evidence of malicious activity and called it “a technical issue not related to cyber or AI / agents.”

Background

Exchange Online is the mail spine of Microsoft 365 — Teams calendars and Copilot’s access to work data both run through it, which is why the Microsoft 365 outage looked the way it did. Per Microsoft’s status updates as compiled by NBC News, Teams calls and chats mostly held, but calendars broke and private or shared channels could not be created or deleted; Copilot Chat would not load and could not reach mail, calendar or files; Intune device enrollment, check-in and compliance checks flapped. Azure showed no issues on Microsoft’s status page.

The nearest precedent is a month old: on October 20, a power event in Google Cloud’s London region knocked out 38 products and reached UK banks and courts. Microsoft’s own arc is worth noting too: on October 1 it reorganized Copilot into a front door for work — and this week that front door was among the first things users could not open.

The facts

Cross-checking Microsoft’s updates against NBC News, Digital Trends and AFP, the verifiable specifics of the Microsoft 365 outage are:

ItemDetail
Start2026-11-12, 11:17 a.m. ET: Exchange Online mailbox access failing across connection methods (Microsoft’s account, per NBC and Digital Trends)
EscalationReports climbed by early afternoon; Digital Trends dates the multi-product incident to 1:26 p.m. ET, NBC to about 12:50 p.m. from user reports
AffectedExchange Online (Outlook web, desktop, EAS, REST; delayed mail), Teams calendars and channel management, Copilot Chat, SharePoint, Intune enrollment and compliance; Azure unaffected
Scale7.4M+ Downdetector reports, 74 countries and territories, peak about 19,650 per minute
Incident IDsAdmin center MO1550821; November 12 X posts: Exchange Online EX1550543, Teams TM1550488
RecoveryPatch to an upstream authentication and routing configuration component early November 13 ET; email timeout window extended to 48 hours; traffic reinstated region by region; recovery declared November 14
AttributionMicrosoft: no evidence of malicious activity; not related to cyber or AI/agents; full post-incident report due after recovery completes

What could be read first-hand came from @MSFT365Status’s November 12 posts: one said Microsoft was “investigating reports of users experiencing issues accessing Exchange Online mailboxes through multiple connection methods in the Americas”, another pointed to Teams problems in North America. The recovery post was not directly readable; Digital Trends quotes it: “We have completed our efforts to address the underlying issue of the incident and restored service to the remaining impacted customers and regions.”

Recovery followed Microsoft’s updates step by step. Early on November 13 ET, engineers patched an upstream authentication and routing configuration component and saw “positive results for server stability”; the email timeout window was extended to 48 hours so queued mail would not be dropped; traffic came back in phases across Asia-Pacific, Europe and North America, with queues “significantly reduced and approaching normal levels” by that morning. Microsoft’s explanation of how it spread shifted along the way: first “unusually high traffic”, then backend servers becoming unexpectedly unavailable and overloading the rest, with restarted machines failing again once traffic returned. The conclusion that an earlier capacity upgrade was exposed by a later change came only as recovery neared completion.

What others say

Microsoft’s apology was the most direct statement of the day. AFP reporting carried by the Korea Times quotes its X post: “We are deeply sorry for the disruption this incident has caused our customers and partners.”

Affected institutions described the damage differently. Rachel Harrison, head of public services at the UK’s GMB union, said staff taking emergency calls were “unable to refer patient details in the normal way”; the National Pharmacy Association warned pharmacies might miss emailed prescriptions from GPs (AFP; NBC News). IndiGo and Akasa Air confirmed email support was down, while Air India said flight operations were unaffected; Immigration New Zealand halted all visa processing; Japan’s economy minister Ryosei Akazawa warned of a “significant impact”.

Outside observers put the Microsoft 365 outage in a larger frame. Australia’s Home Affairs minister relayed Microsoft’s view that this was no cyberattack, likening it to 2024’s CrowdStrike event; the former head of Britain’s NCSC said it “looks like being one of the biggest outages of all time” (both via NBC News). Prins Ralston, CEO of the Australian Computer Society, called for a national review of sovereign IT risk; in the ACS’s publication Information Age, Ralston wrote: “We need to know the systems that keep this country running won’t simply fall over without notice.” The ecosystem felt it too: OpenAI and Anthropic both reported their Microsoft 365 connectors were affected (NBC News).

One source difference is worth flagging: the AFP dispatch says the outage started on a Thursday without giving a date. NBC News and Digital Trends both put the start on November 12, a Thursday, and this entry follows Microsoft.

Our take

For teams that live in Microsoft 365, three lessons from this outage are actionable. First, restored does not mean delivered: with the mail timeout window stretched to 48 hours, queued messages kept trickling in after the all-clear, so a quiet inbox is not proof of a quiet week. Second, change management is the keyword in the root cause: an earlier capacity upgrade planted the flaw and a later change tripped it — the exact scenario change review and staged rollouts exist for. Third, Copilot is pitched as the front door to work, but its availability inherits the mail spine; before wiring an AI assistant into a critical path, map what it falls back on.

The honest caveats: the root-cause narrative is Microsoft’s alone for now, with no independent verification and the full post-incident report promised only after recovery completes — worth revisiting when it lands. The 7.4 million figure counts Downdetector reports, not users, and no reliable count of affected organizations exists. The CrowdStrike comparison is structurally apt — a single point taking down global business — but mechanically different: that was a content update in security software, this was a network change.