Microsoft restores all cloud services after Thursday evening outage

By Amara Osei

Microsoft restored every affected cloud service by Saturday morning (CET), ending an outage that began Thursday evening and left Outlook email, Microsoft 365 applications, the Copilot assistant and interfaces to third-party AI providers unusable for hours. The US software group confirmed on its website and on the short-message platform X that all systems were running again, making this the second Microsoft 365 disruption this year after similar problems in January.

Which services failed, and on what schedule?

The failure reached well beyond a single product. Users reported intermittent problems sending email, managing calendars and holding video conferences, and Microsoft 365 services generally degraded. Third-party connections broke intermittently as well, pulling down AI services from OpenAI and Anthropic alongside Microsoft's own Copilot assistant.

Microsoft wrote on X: "We have completed our actions to remediate the cause of the incident and restored service for the remaining affected customers and regions."

The timeline:

• Thursday evening: the disruption starts across Microsoft's central cloud offerings, hitting Outlook and the wider Microsoft 365 package.
• Friday afternoon: Microsoft reports stepwise progress toward recovery.
• Saturday morning (CET): the company declares all systems restored and remediation of the root cause finished.

Because Outlook and Teams rank among the most used office applications anywhere, the outage spread quickly from enterprises and consumers into government offices worldwide.

Which governments and agencies reported outages?

The Australian intelligence service and the Federal Police reported outages. Canberra activated an agreement with Microsoft to secure priority restoration of the affected systems. The US House of Representatives also reported restrictions for its users.

Germany's federal administration escaped largely intact. A spokesperson for the Federal Office for Information Security (BSI) said on Friday: "The federal administration is operational and essentially unaffected by the outage."

Was it a cyberattack?

No. Microsoft attributed the incident to a self-inflicted problem, ruling out a cyberattack and any AI activity, and stated there were no indications of malicious activity. The root cause sat in a previously executed upgrade to network capacity. A defect introduced during that upgrade, the company said, only surfaced after more recent network changes.

Microsoft also dismissed any link to a separate technical glitch that, since Friday evening (CET), has led to thousands of flight cancellations in the United States. The two events are unrelated, the company said.

Does this fit a pattern of cloud instability?

It does. Similar Microsoft 365 problems in January preceded the Thursday event, and an outage of the Azure cloud platform disrupted services worldwide last month. Three cloud failures inside roughly three months now sit on the record of a single provider whose services underpin email, office software, video conferencing and government back offices in several countries.

Microsoft ended a repair that ran from Thursday evening to Saturday morning with a plain admission: its own network capacity upgrade broke the service, and no attacker was involved.