Microsoft has said it is "deeply sorry" for a widespread outage that caused disruption to millions of users, organisations and businesses across the globe – and that it has "addressed the underlying issue".

The company said the outage was a "technical issue" not related to cyber-attacks or AI agents, and that it continued to gradually restore services.

The outage of Microsoft's services, which began Thursday evening, lasted roughly 18 hours before services started to recover just before midday GMT on Friday. Microsoft said it was bringing back services gradually to ensure a stable restoration.

Earlier on Friday, it said email queues were "significantly reduced and approaching normal levels" as it brought back services. More than 430 million people use Microsoft 365 apps, and more than 90% of Fortune 500 companies use Copilot, Microsoft's AI chatbot, according to the company.

On Friday, government agencies worldwide, from municipalities in France to the US House of Representatives, reported problems with Microsoft services. The issues were patchy, with some but not all users reporting problems with services including Microsoft Teams and Outlook email.

The outage website Down Detector showed an increase in Microsoft 365 problems at about 6pm GMT on Thursday, with nearly 130,000 reports of problems in 74 countries affecting access to Word, Excel, PowerPoint, Outlook and Teams as well as enterprise services such as SharePoint and OneDrive. Copilot was also affected.

Microsoft has suggested the outage may have been caused by problems with the settings applied to servers responsible for keeping Microsoft 365 up and running.

Microsoft's executive vice-president of Copilot, Charles Lamanna, said: "We regret that today's outages caused disruptions for our customers, and we are working to address the issue."

The Microsoft outage also hit connectors used by ⁠OpenAI's and Anthropic's AI ⁠services.

OpenAI said it was investigating a problem with Microsoft connectors in which new connections might be broken and tools were not working as ⁠expected. Anthropic said it had identified an issue with an upstream provider ​affecting its Microsoft 365 ​connector.

The outage has affected businesses and organisations across the world that use Microsoft products.

In the UK, healthcare systems were disrupted and several British police forces urged people not to try to report crimes online and to phone them in instead.

Some schools in the UK notified parents that reporting attendance or absence issues had to be done via phone. National Rail, which provides information across the UK network, warned people might not be able to access tickets that had been emailed to them.

US government offices across the country have been affected by the outage, including congressional offices that heavily rely on Outlook email. The US House's chief administrative officer that handles IT for the chamber sent out a "campus-wide alert" to staffers on Thursday about the outage, Politico reported.

Multiple cities, including the City of Tampa, Florida, posted notices on their municipal websites to say that they were affected by the outage. Law enforcement offices across the US were also affected, with one sheriff's county in Colorado saying it was unable to process bond payments because Outlook email was down, and requesting that anyone who needed to post bond do so in person.

When stock market trading opened on Wall Street, shares in Microsoft dipped slightly, down 1.1% at $516.86.

In Australia, the financial impact of the outage was expected to cost billions of dollars, according to the risk advisory firm, RSM Australia. Several government agencies, including national security sevices, were unable to access or send emails.

Hospitals and health clinics were affected, although the Australian Medical Association on Friday said had been no impacts for patient care. Courts and schools also struggled to operate as normal.

The Indian airlines IndiGo and Akasa Air said their email support services were affected, and Taiwanese security firms, life insurers and investment trust companies said fund management services had been disrupted.

In financial markets some portfolio managers have been unable to check emails, making it hard to access on some trading platforms.

In Japan, the shipping company Sagawa Express said it was unable to accept online reservations for some of its services at airports and train stations while Taisho, a leading pharmaceutical company, said it was having to handle orders by phone.

Microsoft last month reported an ​outage of its Azure cloud platform that affected a ​suite of productivity software ‌and a range ​of industries worldwide.

Mark Gregory, an associate professor in the School of Engineering at RMIT University in Melbourne, said the outage appeared to be affecting the network layer in Microsoft's systems, and was potentially similar to the configuration issue that brought down Optus's network in 2023.

He said that when some of the servers had configuration issues, it meant customers were forced to share a smaller pool of servers to access the cloud services.

"If you lose part of that pool, then all of the users are trying to access the remaining servers, and that overloads them. And then they go down. And so you end up with a cascading problem. Then it takes time to find the underlying problem, in this case, a misconfiguration in the network layer."

Gregory said businesses reliant on cloud services such as Microsoft 365 should have backup options.