Microsoft restores global services after prolonged email disruption

Microsoft has restored Outlook, Teams and other Microsoft 365 services worldwide after a prolonged outage disrupted businesses, governments and essential public services across more than 70 countries and territories.
The company confirmed on Saturday that the remaining affected customers and regions had regained access, ending a disruption that began on Thursday and continued through much of Friday. It said the underlying technical problem had been addressed and that delayed email messages had been processed.
Microsoft ruled out a cyberattack and rejected suggestions that artificial intelligence systems or agents had caused the failure. Instead, it traced the incident to a problem introduced during an earlier network capacity upgrade that surfaced after a subsequent network change.
The company has not yet published a detailed technical account explaining precisely how the changes triggered such extensive failures. It said an investigation would continue and that further findings would be made available after its review.
“We are deeply sorry for the disruption this incident has caused our customers and partners,” Microsoft said, acknowledging the dependence of organisations on its services. It explained that restoration had been deliberately gradual, with engineers checking stability at each stage rather than bringing all affected systems back simultaneously.
The disruption affected Microsoft 365 applications connected to Exchange Online, the cloud platform underpinning business email. Users encountered difficulties accessing mailboxes, sending and receiving messages, scheduling meetings and using collaboration tools. Some also reported interruptions involving Microsoft’s Copilot assistant.
Problems first emerged around midday on Thursday in the United States, corresponding to early evening in Britain. Complaints spread as users in other time zones attempted to access workplace accounts, with successive waves of disruption affecting organisations across Asia, Europe and North America.
Downdetector recorded reports from users in more than 70 countries and territories. Such reports provide an indication of the geographical reach of an outage but do not establish how many individual customers were affected. Microsoft has not disclosed a definitive figure for disrupted users.
The consequences extended beyond ordinary office correspondence. Government departments, courts, healthcare providers, banks and airlines reported difficulties with systems dependent on Microsoft’s cloud services, illustrating how interruptions to shared technology infrastructure can affect unrelated institutions simultaneously.
Britain’s public services were among those affected. Some hospital workers resorted to paper records, while court proceedings faced communication difficulties. Police services experienced problems receiving online reports, and disruptions to electronic systems complicated aspects of healthcare administration, including prescriptions.
Government operations in Japan, Singapore and Australia also experienced interruptions. The scale varied between organisations, depending on which Microsoft services they used and whether alternative communication arrangements were available during the outage.
Businesses faced delayed correspondence and interrupted coordination across teams and offices. Email messages accumulated while systems were unavailable, leaving users waiting for outstanding communications even after access began returning. Microsoft’s recovery updates specifically referred to restoring mailbox access and processing delayed messages.
Customers in some regions regained access before others. During Friday, Microsoft reported improvements in mailbox availability and email processing, although it cautioned that recovery was incomplete. Engineers continued monitoring affected environments while working through the remaining problems.
Users in the United Arab Emirates began receiving Outlook messages again on Friday afternoon, with access returning on desktop and mobile applications. Elsewhere, the timing of recovery differed according to the service and region involved.
The incident also affected integrations between Microsoft 365 and third-party artificial intelligence products. OpenAI and Anthropic subsequently indicated that connections involving Microsoft’s workplace services had returned to normal, after disruptions to their respective integrations.
Microsoft directed administrators seeking further operational information to incident reference MO1550821 in the Microsoft 365 admin centre. The company’s service updates distinguished between correcting the underlying fault and completing recovery for customers whose email systems remained affected.
The disruption highlighted the operational importance of Exchange Online, which supports email and calendar functions for organisations using Microsoft’s subscription services. A failure in that infrastructure can interrupt several applications even when users encounter the problem through different interfaces.
Microsoft said its engineers had verified that previously affected environments were functioning again and that the backlog of email messages had been cleared. Its service health communications subsequently marked the incident as resolved, while the separate investigation into its technical causes remained open.