Microsoft says its Microsoft 365 services are fully restored after a global outage that knocked Outlook, Teams and Copilot offline for more than a day, and it has named the culprit: a network capacity upgrade that went wrong. The company apologised to customers on Saturday, said the fault was a technical issue unrelated to cyber activity or artificial intelligence, and promised a full post-mortem. The outage, which began late on Thursday, disrupted email for businesses in more than 70 countries at its peak.

What went wrong
The problem started in Exchange Online, the email platform behind Outlook, and spread to every Microsoft 365 service that depends on it. Users could not open mailboxes, email delivery stalled, and Teams calendars, SharePoint, Copilot and Defender features that rely on email stopped working. Microsoft tracked the incident as EX1550543 for Exchange Online and MO1550821 for the Microsoft 365 impact.
At its peak, the outage generated more than 130,000 reports of Outlook problems on Downdetector, across more than 70 countries. Indian airlines IndiGo, Air India and Akasa Air said customer support was disrupted, and financial services firms reported trouble accessing research and reaching clients. Flight operations were not affected, and Air India rerouted its email services to keep messages flowing.
The cause Microsoft named
Microsoft said the outage was caused by a technical issue and was not related to cyber activity, artificial intelligence or AI agents. The problem originated during an earlier network capacity upgrade and surfaced after a network change, the company said. It addressed the underlying issue, then restored services gradually, checking stability at each stage because of the scale of demand.
Charles Lamanna, Microsoft’s executive vice president for Copilot, Agents and Platform, said the investigation had found technical challenges that complicated the recovery, including cascading failures and “thundering herds” at the network perimeter. He said he was working closely with the engineering team to complete the full post-mortem, which Microsoft will publish shortly with a step-by-step timeline and repair items.
Restoration and what is still open
Microsoft’s service health page displayed the message “All products are operational” on Saturday morning, covering Microsoft 365 business and enterprise services as well as consumer products. Its service health record for the incident was marked service restored in a final update at 04:17 UTC on Saturday.
The incident’s end was given as 16:10 UTC on Friday, and Microsoft is still monitoring the restored services and tracking the email backlog to ensure every message is processed. A full post-incident report has not been published yet, so the detailed root cause remains to be confirmed.
Our Read
The speed of the apology matters as much as the restoration. Microsoft named a cause within a day, ruled out the two explanations customers fear most — an attack and an AI failure — and put a senior executive’s name on the post-mortem promise. That is the playbook for an outage that hit the world’s email.
Own it fast, give people a mechanism, and commit to the detail in writing — that is what Microsoft did here. The test now is whether the post-mortem explains the mechanism, not just the trigger, and customers will want to know what changes before the next outage. Until that report lands, the cause stays preliminary.
Why It Matters
For the businesses that lost email for a day, the restoration ends the immediate disruption, though the backlog still has to clear. For Microsoft, the incident is the second large Microsoft 365 outage of 2026, and it landed on Outlook, Teams and Copilot — the services that carry its AI pitch to workplaces.
Reliability is now part of that pitch, and another long outage would cost Microsoft more than uptime. Customers who watched their email vanish for a day will judge the promised post-mortem on whether it names concrete fixes, not just a cause.
Background
This was the second large Microsoft 365 outage of 2026. It began late on Thursday 12 November, with Exchange Online failing health checks and servers unexpectedly unavailable. OpenAI and Anthropic both reported trouble with their Microsoft 365 connectors during the outage, and Claude suffered separate login problems on the same day.
Source: Microsoft restores services and the outage timeline


0 Comments