Microsoft 365 Back After Multi-Day Outage Hit Firms in 70+ Countries

Microsoft restored Outlook, Teams and Copilot on Saturday after an outage that spanned more than 70 countries and territories, knocking out email for governments, banks, airlines and courts. The fault, the company said, was a technical issue tied to an earlier network-capacity upgrade — not a cyberattack and not an AI agent gone rogue.

The disruption began around midday on Thursday in the United States and ran into Friday. According to The Associated Press, Microsoft initially hit problems with server availability and Exchange Online, its cloud email and scheduling service. Engineers then worked through a gradual restoration process, and Microsoft confirmed on Saturday it had fixed the underlying issue and brought back the remaining affected customers and regions. The company apologised and acknowledged how much customers rely on its services.

What broke, exactly?

Microsoft attributed the incident to a network-capacity upgrade that turned problematic after a subsequent network change. The company said it found no indication of malicious activity or of any AI agent causing the failure.

That distinction matters for risk assessment. Firms can spend heavily on cybersecurity and still fall to an operational failure inside the systems they depend on. Resilience means redundancy, tested recovery procedures and alternative channels — not just perimeter defence.

Where did the damage land?

The outage reached well beyond inboxes. AP reported disruptions to governments, courts, healthcare providers, businesses, airlines and banks, with authorities in Japan, Singapore and Australia reporting interrupted government operations.

In Britain the effects were concrete:

• A court trial in Manchester was halted and the judge sent jurors home.
• Police services struggled to receive online crime reports.
• The National Pharmacy Association warned of possible delays to prescription services.

How did businesses cope?

Some reached for technology of an earlier vintage. Reuters reported on Friday that businesses used telephones, fax machines and Bluetooth transfers to keep essential activities running. Euronext advised listed companies to use telephone communication. Insurer Aviva activated contingency arrangements for its earnings calls, and Maersk told customers to phone local representatives about urgent matters and to verify unexpected communications linked to the outage. In Taiwan, some financial institutions reportedly confirmed transactions by fax.

The lesson is uncomfortable. Moving processes online does not make them resilient; retiring older channels without working alternatives leaves fewer ways to operate when the primary platform fails. For firms spanning multiple jurisdictions, one outage can hit suppliers, customers, regulators and counterparties at the same time.

What does this change for cloud strategy?

Microsoft says more than 430 million people use its Microsoft 365 applications. When email, documents, meetings and collaboration sit on interconnected services from one vendor, a single failure can take down several functions at once. AP highlighted precisely this concentration risk, and reported that some lawmakers outside the United States have raised concerns about overreliance on American technology infrastructure.

The answer is not abandonment of major cloud providers. The commercial case — collaboration across locations, lower infrastructure burden, scaling without building everything internally — remains intact. The answer is supplier-failure planning: identifying which functions critically depend on a single provider, establishing alternative communication channels, keeping essential records accessible during disruption, and actually testing the backups. A contingency plan that exists only on paper protects no one.

Smaller organisations need not buy duplicate infrastructure. They need to identify essential processes, set up practical alternatives and ensure staff know what to do when normal systems go dark.

The restoration ends the immediate disruption. It does not answer the question the incident raised: how prepared is your organisation for the failure of the platforms your daily operations run on? For anyone who has centralised communications around a handful of providers, that question now belongs to operational resilience, not just IT.