A Microsoft outage disrupted Outlook, Teams and other Microsoft 365 services on Thursday, November 12, affecting users across multiple regions. The disruption also affected Microsoft 365 connectors used by OpenAI and Anthropic services.

The outage lasted more than five hours for some users, while Microsoft worked to restore affected servers. Microsoft’s Copilot was also among the services affected.

Downdetector recorded a peak of 138,058 reports of Outlook problems in the United States at around 1 p.m. ET. Reports fell to about 34,490 later in the evening. These figures reflect user reports to Downdetector, not a verified count of affected customers.

The disruption spread quickly. Downdetector said reports came from 74 countries and territories within about an hour. Microsoft initially said it was investigating reports in the Americas, then confirmed that Microsoft 365 connectivity could be affected in any region.

Microsoft pointed to infrastructure strain as the cause. The company said part of its infrastructure was experiencing elevated load across multiple regions, likely contributing to connectivity errors. It also said some servers had unexpectedly become unavailable, placing additional strain on the remaining infrastructure. Microsoft said it had not identified a cyberattack as the cause.

In updates on X, Microsoft said it had paused deployments, moved traffic away from affected infrastructure and restarted servers. “We’re aggressively pursuing all pathways to reduce impact,” the company wrote.

A Microsoft spokesperson said the company was investigating problems with Microsoft 365 services, primarily Exchange Online. At 5:47 p.m. ET, Downdetector showed nearly 13,000 reports of Microsoft 365 problems and 763 reports involving Teams. Microsoft said some restored servers remained stable and that it was moving traffic to alternate routes where possible.

Recovery was gradual. After 6 p.m. ET, Microsoft said it was seeing limited recovery as users reconnected. The company cautioned that service could return gradually as capacity came back online. It also said some recovery attempts were not working as expected, after earlier reporting positive results.

The effects reached AI tools that connect to Microsoft 365. OpenAI said it was investigating an issue in which new Microsoft connector connections could fail and some tools might not work as expected. Anthropic also reported an issue with an upstream provider affecting its Microsoft 365 connector.

The incident follows an Azure outage in October that affected Microsoft services and other online platforms. Thursday’s disruption showed how problems with shared infrastructure can affect email, collaboration tools and connected AI services at the same time.

Related post

View all