Microsoft Says Global Outage Fully Resolved After 70-Country Disruption

Microsoft declared early Saturday morning Brazil time that all its services were fully operational again, closing a global outage that disrupted public agencies, financial market operators and companies in more than 70 countries. "All products are operational," the company wrote in an update on its service status page.

Microsoft said it had processed the backlog of pending email messages and fixed the problem in Microsoft 365, the subscription bundle (software licensed monthly rather than bought outright) that serves more than 450 million corporate users.

What went wrong, and who was hit?

The outage hit Outlook, Teams and the Copilot AI assistant on Friday. Users reported missing meeting invitations and unresponsive web pages. Downdetector, a service that aggregates user complaints, recorded hundreds of thousands of reports across more than 70 countries.

Previous interruptions rarely lasted more than a few hours, which made the duration of this failure notable. Office workers and government employees from Tokyo to London and New York struggled to adapt, often falling back on 20th-century analog tools such as the telephone and, yes, paper.

The disruption carried real financial consequences. The Swiss luxury group Richemont and the American truck and bus manufacturer Workhorse were both forced to postpone the release of their financial results. Essential services also took the hit: the UK's National Health Service (NHS) had to reschedule some procedures, according to people familiar with the matter.

In the United States, some corporate employees celebrated the Outlook outage as a kind of "snow day for adults," a reference to days when schools and offices close for snowstorms. Employees on Google Workspace, meanwhile, watched their Outlook-dependent colleagues enjoy the break with some envy. The levity does not cancel the lost earnings reports or the rescheduled medical procedures.

How did Microsoft respond?

A spokesperson said the error was introduced during a network capacity update — and nothing more. The company posted hourly status updates that revealed several failed attempts as engineers hunted for the root cause. More than a thousand engineers and other staff joined videoconferences led by Scott Guthrie, head of the cloud division, and Charles Lamanna, responsible for M365 and other workplace applications. The meetings began Thursday night and ran through Friday morning.

Microsoft said it would provide more details once the outage ended. It stated there was no evidence of malicious activity and, per a spokesperson, no connection to artificial intelligence or AI agents.

"We know our customers rely on our services to do important work every day, and we regret the impact these disruptions had on those who count on us," Lamanna said in an emailed statement. "Our teams are reviewing what happened and how we can prevent similar issues in the future."

An apology is not a root-cause explanation, and the promised post-incident detail has not yet arrived.

Does this change anything for buyers?

The failure and the time needed to fix it mark a setback for Microsoft, whose Windows operating system and productivity applications remain essential infrastructure in the AI era. Critics again pointed to the risks of concentrating so much power in a single vendor, though many concede that abandoning Microsoft's ubiquitous products is unlikely in the short term.

Investors, for the most part, shrugged. Microsoft shares showed little movement.

"We do not expect the incident to meaningfully change Microsoft's competitive position in the near term, given the high cost and complexity of replacing productivity workflows that are deeply embedded into companies' operations," wrote William Blair & Co. analyst Jason Ader in a report. "Over the long term, however, the episode could make some customers more receptive to alternatives, particularly Google Workspace and the new productivity offerings from Anthropic, unveiled earlier this week."

The unanswered question — what exactly a routine network capacity update did to take down a 450-million-user ecosystem — is the one Microsoft's customers will want addressed when the fuller technical account arrives.