Microsoft 365 Outage: How an Exchange Online Failure Reached Calendars, Copilot and Outside AI Tools [Japanese Reactions]

Table of Contents

Key Points

・A failure in the email backbone of Microsoft 365 disrupted Outlook and spread to calendars, Copilot and other companies’ AI connectors.

・Microsoft said the cause was a technical issue triggered by a network change, not a cyberattack and not related to AI.

・The impact spread because many features rely on data held in the email backbone, while Teams calls and chat kept running.

▼ Watch the Japanese-language video explainer

This video has Japanese audio and slides. To read English captions, open Settings → Subtitles/CC → Auto-translate → English, where available. Automatic translations may contain errors. If the option is unavailable in the embedded player, watch on YouTube. The accompanying English article provides the written explanation.

Microsoft 365 Email Outage Spreads to Calendars and AI Tools

Microsoft first notified customers at 16:20 UTC on November 12, 2026 (01:20 on November 13, Japan time), that users might be unable to reach their mailboxes in Exchange Online, the email service behind its business suite Microsoft 365. In Outlook, mail would not open or arrived late, and the problems spread to calendars, meeting scheduling and the AI assistant Copilot.

The company restored service region by region, in the order of Asia-Pacific, Europe and North America, but it has not disclosed how many users were affected.

In Japan, the Ministry of Foreign Affairs and the Ministry of Education, Culture, Sports, Science and Technology asked people on X to contact them by phone about urgent matters.

Osaka City also asked residents to call instead, and MBS News reported on November 13 that the work of about 8,000 Sakai City employees was affected. In the United States, electric vehicle maker Workhorse Group postponed its earnings call to November 16, citing business disruptions caused by the outage.

In a final update at 04:15 UTC on November 14 (13:15 Japan time), Microsoft declared service restored and said the incident ran from 10:05 UTC on November 12 to 16:10 UTC on November 13. Microsoft said an issue introduced during an earlier network capacity upgrade surfaced after a recent network change, and described it as a technical issue not related to cyber or to AI and agents.

Microsoft said it would publish a preliminary post-incident report within two business days and a final report within five business days.

What Stopped, What Kept Working, and How Service Came Back

Exchange Online is Microsoft’s cloud service that stores and delivers email, calendars and contacts for Microsoft 365 business customers, and Outlook is the app people use to read and write that data. Because calendars and some AI features draw on the same service, an email failure did not stay an email failure.

Exchange Online, the system behind Outlook

When someone books a meeting in Teams, the appointment is written into Exchange Online as well. Company documents and files are stored elsewhere, in services such as SharePoint.

According to Microsoft’s service updates, opening files directly in SharePoint kept working. Copilot, which answers questions by pulling together email, calendars and files, failed in several ways: it could not reach work data, its chat screen would not load, or it did not respond.

In other words, files could still be opened, but finding or summarizing them through the AI assistant was disrupted. Microsoft has not said which parts of Copilot failed or how.

Microsoft’s list of affected features also included meeting scheduling and security tools that watch for confidential information leaving the company. Microsoft said core Teams features, including calls, meetings and chat, were not affected.

AI tools from outside Microsoft were hit too. Anthropic, the developer of Claude, said on its status page that from 16:03 UTC on November 12 to 16:40 UTC on November 13, many requests through its connector for reading Microsoft 365 mail and calendars failed, and it attributed the errors to the Microsoft 365 incident.

OpenAI reported errors in a similar Microsoft 365 connector for ChatGPT on its status page, starting at 18:29 UTC on November 12 and resolved at 00:18 UTC on November 14, but it did not state a cause. There is no indication that the AI models themselves were at fault.

On the file side, Microsoft listed problems with specific operations, such as exporting reports that carry sensitivity labels like “confidential” and sharing files outside the organization. There is no indication that any file format was to blame.

A slow recovery, and why reconnecting was not the same as returning to normal

The repair did not work in one pass and moved forward in stages, region by region.

At 19:30 UTC on November 12 (04:30 on November 13, Japan time), Microsoft said the problem was not a rise in overall traffic but a large number of servers becoming unexpectedly unavailable, which pushed the load onto the rest. Later updates said earlier steps had not brought the expected relief, and that restarted servers stopped responding again once they took traffic.

The explanation also shifted along the way. At 19:58 UTC on November 12, Microsoft said it had not identified a recent change that lined up with the start of the incident, but at 01:12 UTC on November 13 it began rolling back a recent network configuration change that did. After it patched an upstream component that handles connection authentication and routing, at 07:24 UTC, service began to stabilize, and recovery moved through Asia-Pacific, Europe and North America in that order. In its 12:18 UTC update on November 13, Microsoft said Japan remained comparatively more stable.

Microsoft’s first customer notice came at 16:20 UTC on November 12, but its final update puts the start of the incident about six hours earlier, at 10:05 UTC.

Getting connected again did not mean work was back to normal. Mail sent during the outage piled up in delivery queues, and at 09:03 UTC on November 13 Microsoft extended to 48 hours the period during which undelivered mail would keep being retried automatically.

Dave Plummer, a retired Microsoft engineer who publishes explainer videos, urged viewers to separate mail that failed to send from mail that is only delayed. Resending a message that is merely delayed can leave the recipient with two copies.

The Convenience of a Shared Platform, and the Weakness This Outage Exposed

A shared dependency on the email backbone widened the impact

Calendars, Copilot and security tools were all affected because each uses mail and calendar data stored in Exchange Online and depends on connecting to it.

That link is also what makes the suite convenient day to day. Copilot can look across email and files at once and answer a question such as “Which emails do I need to prepare for next week’s meeting?” because one company’s services are connected to each other. Creating a meeting in the calendar and sending the invitation from the same place works for the same reason.

Organizations also cite the ability to keep administrative settings and staff training on a single system as an advantage of using the suite as a whole.

Some of Microsoft’s protection features are hard to reproduce in the same way on another provider’s services, such as Google’s.

When Sony Network Communications, a Japanese internet provider, described at a Google event how it added Google’s workplace tools, it said fine-grained control that encrypts labeled files and limits who can open them could not be replicated in the same way on Google’s side.

The company filled that gap with employee training and log monitoring.

On the other hand, Teams calls and chat, and opening files stored in SharePoint, kept working. How far Microsoft was able to contain the failure within its own suite should become clearer in its detailed post-incident report.

The 2025 outages at AWS and Cloudflare stopped infrastructure that sits behind many websites and apps. This time, what stopped was a set of tools people use at work every day.

Where the impact was confirmed, offices switched to tools that still worked

The ministries and local governments whose responses we could confirm switched to features that were still running and to the telephone.

The Ministry of Foreign Affairs and the Ministry of Education, Culture, Sports, Science and Technology said they would take urgent matters by phone, and Osaka City asked residents to call. Sakai City temporarily relied on phones and other means, and, according to a Kansai TV report, its employees used Teams chat to reach one another. In place of email, they used features of the same Microsoft suite that were still working.

At a press conference on November 13, Japan’s Digital Minister Furukawa said government work systems had also been disrupted and that each agency has several ways to communicate, including phone, chat and text messages, according to press conference video released by Nippon TV. Like Workhorse in the United States, some companies moved scheduled announcements.

Plummer pointed out that an uptime commitment is not a substitute for a plan for keeping the business running when a service goes down. How long a provider takes to restore service and how its customers keep in touch while it is down are being treated as two separate questions. Microsoft apologized and said it would publish a detailed post-incident report.

Splitting or switching platforms serves different goals at different costs

Some organizations have moved part of their work tools to another provider, or moved off entirely, but each had its own goal, and outages like this one were not necessarily the reason.

Sony Network Communications added Google’s environment to get employees using AI in their daily work. It kept external email on Microsoft and took on the extra work of forwarding mail and syncing calendars so the AI tools could use them. Because documents sent outside the company are expected in Word or Excel formats, it has not moved over completely.

The German state of Schleswig-Holstein has spent years moving to open-source software, under a government policy of not being tied to any one vendor. Along the way, heise reported that judges were at one point unable to use email, which disrupted court proceedings.

The state government told the state parliament in an October 2025 document that the move carries training and labor costs, while staying with Microsoft would also bring costs for upgrades and a move to the cloud. The savings are the state’s own projection, starting several years out.

Japanese Reactions to the Microsoft 365 Outage

The following four posts on X, written in Japanese on November 13, 2026, are a small sample from individual users and do not represent public opinion in Japan. Descriptions of what stopped or what happened to mail reflect each poster’s own workplace and impressions, and times in the posts are in Japan time.

ひなきち (@mt_kainalu), November 13, 2026:

娘の体調良くないから病院行くために午前半休したいのに、Outlookが死んでて会社に連絡できない😭
仕方なく無断欠勤中😭

メールがOutlookで電話はteamsの弊社終わってる😇
出社して混乱ぶり見たかった😇
— ひなきち (@mt_kainalu) November 13, 2026

My daughter isn’t feeling well and I want to take the morning off to take her to the hospital, but Outlook is dead and I can’t contact my company 😭

So I’m absent without notice for now 😭

My company is hopeless, with email on Outlook and phones on Teams 😇

I kind of wish I’d gone in to see the chaos 😇 (translated)

こむぎこ。 (@komugiko_uta), November 13, 2026:

弊社、Outlookそもそも開けない組と、
開けるけど更新されない組がいるんだけど
おれぴは開ける組のため
「○○さん次○○会議室！」「○○さんは○○会議室！！！」と伝える仕事をさせられている(?)
— こむぎこ。 (@komugiko_uta) November 13, 2026

At my company there’s the group that can’t open Outlook at all,

and the group that can open it but it won’t update.

I’m in the can-open group,

so I’ve been given (?) the job of telling people, “Mr./Ms. So-and-so, your next meeting is in Room such-and-such!” “Mr./Ms. So-and-so, you’re in Room such-and-such!!!” (translated)

Pレコ (@reconguistainP), November 13, 2026:

Outlookとりあえず復旧したっぽいけど、止まってた間のメールはインターネットの海に飲まれたっぽいな…
— Pレコ (@reconguistainP) November 13, 2026

Looks like Outlook has more or less recovered, but the emails from while it was down seem to have been swallowed by the sea of the internet… (translated)

よんてんごP (@yontengoP), November 13, 2026:

一部では
「何で弊社はMSメールしか使えないんだ、
サーバとかルータも冗長化するんだろう
こうしてOutlookが止まったら仕事が止まるじゃないか

メールも、MS365Outloot、Gメール、Yahooメールとか複数準備しておくべきじゃないのか」

みたいな話も出てるらしいんすけど、そんな会社あるの…？？？
— よんてんごP (@yontengoP) November 13, 2026

Apparently in some places

people are saying things like, “Why can our company only use MS mail?

We make our servers and routers redundant, don’t we?

If Outlook stops like this, work stops.

Shouldn’t we have several email services ready too, like MS365 Outloot [sic], Gmail and Yahoo Mail?”

but does any company actually do that…??? (translated)

What One Platform’s Outage Revealed

Microsoft declared service restored on the afternoon of November 14, Japan time, but a detailed account of how the failure cascaded will have to wait for its post-incident report.

What the outage showed was how many features depend on data in the email backbone, and which tools people turned to on the day it stopped. In Japan’s ministries and local governments, the calls and chat that kept running, along with the telephone, held work together.

How much to consolidate on a single platform, and what to run work on when it stops, depends on an organization’s size, its business partners and the work it does. As more tasks are handed to AI, the factors behind that decision may change too.

Frequently Asked Questions

What caused the Microsoft 365 outage in November 2026?

Microsoft said in its November 13, 2026, update that an issue introduced during an earlier network capacity upgrade surfaced after a recent network change. It described this as a technical issue not related to cyber or to AI and agents, and said a detailed post-incident report would follow, with a preliminary version within two business days and a final version within five business days.

How long did the Microsoft 365 outage last?

According to Microsoft’s final update on November 14, 2026, the incident ran from 10:05 UTC on November 12 to 16:10 UTC on November 13, about 30 hours. Recovery came region by region, in the order of Asia-Pacific, Europe and North America, and Microsoft extended to 48 hours the period for automatically retrying delayed mail.

Were Microsoft Teams calls and chat affected?

No, not their core functions. Microsoft said Teams calls, meetings and chat were not affected, while Teams features that depend on Exchange Online, such as calendar scheduling, did fail; in Japan, Kansai TV reported that Sakai City staff used Teams chat to contact one another during the outage.

How did Japanese social media react to the Microsoft 365 outage?

Among the Japanese-language posts on X we reviewed from November 13, 2026, one user said they could not reach their employer to take time off for a sick daughter, and another described relaying meeting room changes by voice because the calendar would not update. Others worried that mail sent during the outage had been lost, or questioned whether any company really keeps several email services ready. These posts are a small sample, not a measure of Japanese public opinion.

Sekahan on YouTube

We publish video summaries of articles like this one, along with short clips built around Japanese reactions.

Subscribe: Sekahan Japan Reacts