Microsoft 365 outage knocks out Outlook, Teams and Copilot worldwide

Microsoft 365 went down for hours on Thursday, cutting off Outlook, Teams and Copilot for users around the world. BleepingComputer reports that Microsoft traced the disruption to servers that became unexpectedly unavailable, which strained the infrastructure still running. Recovery has been slow and uneven, and Microsoft hadn’t given a time for full restoration at the last update we could verify.

What broke and what Microsoft says

In a Microsoft 365 admin center advisory tracked as MO1550821, Microsoft confirmed that customers may be unable to connect to Exchange Online, Teams, SharePoint Online, Microsoft 365 Copilot and Defender XDR. The incident began at 1:26 PM ET on Thursday, although Microsoft had been investigating Exchange Online mailbox problems in the Americas earlier in the day. It later confirmed that users in any region could be affected. ABC’s live coverage lists further services that may be impacted, including Planner, To Do and Intune. ABC News Australia adds that Safe Links in Defender for Office 365 may be unable to verify URLs, which can affect links clicked in Outlook and Teams.

Microsoft first blamed elevated load across multiple regions, then revised the diagnosis. In an update on its @MSFT365Status account on X, it said the load wasn’t driven by more traffic but by a number of servers being unexpectedly unavailable, which strained the remaining infrastructure. The company says it paused all deployments, redistributed connection traffic across its worldwide infrastructure and began restarting the affected servers. Its later update admitted that those steps hadn’t provided the anticipated relief, and that a core process on the affected servers was becoming overwhelmed and causing capacity loss. Microsoft also said it was assessing whether reverting a small number of recent configuration updates could speed recovery. ABC News Australia reports that Microsoft later said the outage wasn’t believed to be due to “malicious activity”.

The disruption reached beyond Microsoft’s own apps. According to Reuters, as relayed by ANI, the outage also hit connectors that OpenAI’s and Anthropic’s AI services use. Reuters cites Downdetector figures showing Outlook problem reports peaking at 138,058 around 1 pm ET, before falling to about 34,490 by Thursday evening.

Recovery has been gradual

Microsoft’s 6:57 PM ET update said it had returned a portion of affected servers to service and that they remained stable. It also reported shifting traffic to alternate routes and observing limited recovery, and it warned that users may see a gradual return as their apps reconnect.

That stability didn’t hold everywhere. ABC News reports that Microsoft later said the restored servers can become unresponsive again once they receive traffic, and that it was working on a fix. Tom’s Guide quotes Microsoft as saying it was preparing an automated recovery process to continually detect and restart unresponsive servers. ABC News Australia’s Friday live blog says Microsoft observed “partial recovery” that wasn’t sustained, with no long-term fix yet.

There was also a separate Teams incident earlier on Thursday. Tracked as TM1550488, it affected users in North and South America with delayed or failed messages, errors creating or updating conversations, and trouble retrieving chats. Microsoft blamed resource contention in its infrastructure and says it’s fully resolved. It hasn’t said whether the two incidents are related.

What it means for the Gulf and Egypt

Microsoft hasn’t singled out the UAE, Saudi Arabia or Egypt, but its confirmation that any region could be affected means tenants in the Gulf and Egypt aren’t exempt. Because the main incident began around 10:26 PM UAE time on Thursday, much of the recovery has played out across the start of the next working day. We recommend IT teams check the admin center under MO1550821 and EX1550543 before assuming a local network or Outlook profile fault. Microsoft also points admins to status.cloud.microsoft if the admin center itself is unreachable.

Copilot is a growing part of that exposure. Microsoft is bringing UAE sovereign AI agents into Microsoft 365 Copilot with Inception42, so a suite-wide failure now takes AI features down alongside mail and chat. It’s not the first time Copilot has gone dark either; we covered the September Copilot outage and, earlier this year, the global Claude AI outage.