Microsoft struggles to restore Outlook, Teams and Copilot as global outage drags on

• A global Microsoft outage disrupted Outlook, Teams and Copilot, with Downdetector recording problems across 74 countries

• Microsoft linked the disruption to servers unexpectedly going offline, placing additional pressure on its remaining infrastructure

• Recovery efforts suffered setbacks, with the company warning that restarted systems could become unresponsive again as users reconnected

Microsoft was struggling to restore access to its workplace services on Friday after a global outage disrupted email, collaboration tools and artificial intelligence applications for hours, affecting businesses, government agencies and other organisations.

The disruption began on Thursday and continued into Friday morning, with users reporting difficulties accessing Outlook, Teams and Microsoft’s Copilot AI service. Efforts to stabilise the systems produced some improvement, but Microsoft subsequently acknowledged that changes intended to restore reliable access had failed to deliver the expected results.

The company said it was exploring alternative measures to increase available capacity and address persistent imbalances in the way traffic was being handled across its infrastructure.

The prolonged disruption affected much of the US working day on Thursday, while problems were also reported in the UAE, the UK, Japan, Singapore and Australia. Microsoft warned that customers in any region could potentially experience difficulties.

According to outage-monitoring website Downdetector, reports spread to 74 countries within about an hour. Outlook-related complaints peaked at nearly 140,000 globally around 18:00 GMT on Thursday, more than 2,000 times the usual level.

Although reports subsequently declined, thousands continued to flag problems. Downdetector figures reflect reports submitted to the monitoring platform rather than a confirmed count of everyone affected.

Outlook users experienced problems opening mailboxes and sending or receiving messages. Microsoft initially investigated connectivity failures affecting Exchange Online, the service that supports email and calendar functions for many organisations, before confirming a broader impact across Microsoft 365.

Some customers also encountered difficulties accessing files or viewing calendar information. The list of potentially affected services expanded during Thursday to include file-sharing platforms, cybersecurity services and administrative tools.

The impact varied between applications and customers. Although Teams users reported problems, Microsoft said messaging and online meetings remained largely functional, indicating that the disruption did not affect every feature equally.

Microsoft traced the pressure on its systems to a number of servers unexpectedly becoming unavailable. It said the resulting strain was not caused by an increase in customer traffic: the remaining infrastructure was having to carry work that would normally have been distributed across more servers.

The company was still investigating why those servers had become unavailable and what was preventing a sustained recovery.

To ease the pressure, engineers began diverting connections through other parts of Microsoft’s global network. Recovery work also involved restarting affected systems and adjusting how traffic reached the troubled infrastructure.

Those measures delivered uneven results. Microsoft initially reported encouraging signs as some systems returned to service and customers began reconnecting. It later warned, however, that servers brought back online could stop responding once traffic resumed.

More than seven hours after the disruption emerged, the company said it was examining diagnostic information to establish what was triggering that behaviour. It also temporarily restricted certain types of traffic reaching affected infrastructure in some locations.

By Friday morning, Microsoft acknowledged that its latest changes were “not taking as expected” and said it was assessing other ways to restore capacity.

The outage also disrupted connections between Microsoft’s services and external AI platforms. OpenAI said it was investigating problems affecting its Microsoft integrations, while Anthropic identified difficulties involving its links to Microsoft 365. These reports concerned connected tools and integrations rather than confirmation of a general outage across either company’s AI services.

Government operations were among those affected. Michigan’s Department of Technology, Management and Budget warned that the state was experiencing technical difficulties with email and other Microsoft applications because of the wider disruption. It advised that responses to emails could be delayed.

The incident underscored the operational reach of Microsoft’s workplace software, which is used extensively across businesses, public bodies and educational institutions. Problems with a shared layer of infrastructure affected several services at once, interrupting routine communication and access to information.

Despite periods of partial recovery, Microsoft had not confirmed full restoration in the updates available on Friday morning. Engineers continued working to bring unavailable servers back into operation while preventing renewed failures as customers reconnected.

With agency inputs