Microsoft has apologised for a widespread service outage that disrupted Outlook and other Microsoft 365 products, as the company worked to restore services nearly 24 hours after connectivity problems began.
“We are deeply sorry for the disruption this incident has caused our customers and partners. We know the trust they place in our services and take that responsibility extremely seriously,” a Microsoft spokesperson said on Friday.
The company said it was seeing broad recovery and stability across its services.
In a statement, Microsoft attributed the disruption to “an issue during an earlier network capacity upgrade”.
The software giant said the underlying cause had been addressed, but efforts were still underway to restore services for some customers.
Engineers began working to resolve the connectivity problems shortly before 1pm ET on Thursday.
Outlook’s email and calendar services, along with several Microsoft 365 products including SharePoint and Copilot, were largely unavailable during the morning in the United States on Friday and throughout business hours in Asia and Europe.
Several external services connected to Microsoft products were also affected. Users began regaining access to their emails and other services on Friday morning.
Recovery efforts continue
In an update at about 10.30am ET on Friday, Microsoft reported “encouraging recovery progress”, saying service availability and email delivery performance were improving across an increasing portion of the affected infrastructure.
In an earlier update, the company said email queues were approaching normal levels, although intermittent issues were expected as recovery efforts continued.
Microsoft had provided updates according to geographical regions, with services in Asia-Pacific among the first to begin returning to normal.
Engineering teams have been taking affected systems offline for repairs before gradually bringing them back online.
The company said it was prioritising the preservation of emails delayed during the recovery process, while working to reduce email queues and backlogs.
It had also extended the delivery timeout window to 48 hours, allowing delayed emails to continue attempting delivery and helping to protect messages from being lost.
Microsoft said it had not detected any malicious activity behind the outage.
The company initially attributed the disruption partly to a surge in traffic in sections of its infrastructure and problems with a component responsible for processing server health.
In its initial updates on Thursday, Microsoft said elevated loads across multiple regions had left a number of servers unexpectedly unavailable, placing additional strain on the remaining infrastructure.
Some servers were also failing automated health checks.
Microsoft said it would publish a full report once its assessment of the incident was completed.
Its shares traded largely flat on Friday.
Governments and businesses assess impact
Governments and businesses across the United States, Asia-Pacific and Europe said they were monitoring the situation as the outage caused widespread disruption.
The UK’s National Health Service said some of its services had been affected, but advised patients to attend their appointments as usual.
A hospital group serving South West London and Surrey warned on X that calls could face long waiting times because of the disruption.
Swiss luxury goods group Richemont experienced a slight delay in reporting its earnings on Friday due to the outage.
Electric truck maker Workhorse Group also rescheduled the release of its third-quarter results, originally expected on Nov 12, to Nov 16, citing business disruptions caused by the global Microsoft outage.
US Representative Josh Gottheimer, a Democrat from New Jersey and former Microsoft employee, told CNBC’s ‘The Exchange’ that he had requested a briefing on the incident.
He questioned whether the outage involved malicious activity, noting its scale, duration and timing.
Air India said it had temporarily rerouted emails to ensure it could continue receiving customer messages, adding that its flight operations were unaffected.
Australian Home Affairs and Cyber Security Minister Tony Burke said Microsoft believed the incident was similar in some ways to the 2024 CrowdStrike outage, in which a faulty software update triggered one of the world’s biggest IT disruptions.
The incident affected airlines, banks and other organisations worldwide.
Outage highlights reliance on technology infrastructure
The incident has highlighted the growing dependence of businesses and organisations on a relatively small number of technology and infrastructure providers.
James Owen, managing partner of Digital Risks at consultancy Control Risks, described it as a business resilience and dependency issue that had been developing over decades.
He said organisations had become increasingly reliant on a smaller segment of technology and infrastructure for their daily operations.
Owen said businesses should strengthen their recovery planning, including through manual backups, procedures that do not depend on internet connectivity, diversified suppliers and crisis management exercises.
He added that organisations were becoming more conscious of the geopolitical implications of their technology choices, including whether services were provided by companies in the United States, China or elsewhere.
Microsoft reported in 2025 that Copilot, its collection of artificial intelligence assistants and tools, had more than 100 million monthly active users.
Meanwhile, Amazon Web Services and Google users also reported outages on Thursday.
Amazon said on X that its services were operating normally, adding that an unrelated disruption elsewhere on the internet had prompted inaccurate reports about its services.
The company directed users to its health dashboard for accurate information.
At 4pm ET on Thursday, Microsoft said it had validated an alternative service health check method that allowed some affected servers to return online.
The company said it was applying the change more broadly and preparing automated recovery systems to detect and restart unresponsive servers to maintain service stability.
