Technology & AI
Microsoft 365 fully restored after global outage
Hindu BusinessLine reports a hidden software defect; CBC News reports insured losses of $300m to $500m.
By the numbers
Figures from this story's sources.
- Downdetector reports worldwide during the outage Source for 7.4 million (opens in a new tab)
- 7.4 million
- of gateway servers unresponsive at peak, per Hindu BusinessLine Source for over 80% (opens in a new tab)
- over 80%
- Microsoft share close on Friday, flat on the day Source for $523.88 (opens in a new tab)
- $523.88
Microsoft said on Saturday it had fully restored Microsoft 365 after an outage that began on Thursday. BBC News reported that the disruption reached more than 70 countries.
The Guardian reported that problems reached US users around midday on Thursday and UK users around 6pm. BBC News said Microsoft first acknowledged at 17:54 GMT that users in the Americas could not reach Exchange Online mailboxes.
The two outlets measure the length differently. BBC News said services were beginning to recover by Friday afternoon UK time, nearly 24 hours after the first problems. CBC News put the outage at more than 24 hours, running to full restoration on Saturday.
Microsoft's first account, as the Guardian reported it, pointed to an upgrade-related issue after a recent network change. The Guardian also reported that Microsoft apologised on Friday and said the outage was "not related to cyber" or to artificial intelligence agents.
Hindu BusinessLine reports that Microsoft has since admitted a hidden software defect, a latent bug from a fix introduced in July, in its networking gateway servers. More than 80% of those servers were unresponsive at peak, the outlet reported.
“We are deeply sorry for the disruption this incident has caused our customers and partners.”
CBC News reported that Downdetector recorded more than 7.4 million worldwide reports, with the peak near 20,000 a minute. Morningstar DBRS, through Mario De Cicco, estimated insured losses at $300m to $500m, against a $400m to $1.5bn financial toll for the 2024 CrowdStrike outage.
The disruption stopped work across sectors. Some juries were sent home early, the Guardian reported, and IndiGo and Akasa Air said their email support was affected.
In the UK, cyber minister Liz Lloyd called the impact on critical frontline services "limited", BBC News reported. The Guardian reported that NHS England hospital staff used paper records.
A Microsoft spokesperson apologised to customers and partners for the disruption, BBC News reported. CBC News reported that some companies delayed earnings publication, and that Microsoft shares closed flat at $523.88 on Friday, up about 8% this year.
Why it matters
The sources give no total cost to any single business. The figures they do give are insurance estimates and reported counts. The clearer lesson is concentration: CBC News cited Ookla analyst Luke Kehoe on dependence on a few large technology providers, and reported that organisations hosting their own servers were somewhat spared. Firms that run approvals, payroll or customer contact through one suite carried the outage with them.
What’s next
BBC News reported that Microsoft has promised a full post-incident report once its review is complete.
Sources
Checked against 4 outlets
- BBC News — Reach of more than 70 countries, first acknowledgement at 17:54 GMT, recovery timing on Friday, Microsoft spokesperson apology, UK minister's assessment, and the promised post-incident report
- CBC News — Downdetector report counts, insured loss estimate and CrowdStrike financial toll comparison, share price, earnings delays, the over-24-hours duration account, and expert comment on provider dependence
- The Guardian Technology — Initial timing in the US and UK, the original network-change account, Microsoft's apology and cyber statement, and UK hospital and court disruption
- The Hindu BusinessLine — Revised cause (hidden software defect in gateway servers), over 80% of gateway servers unresponsive at peak, and the 15-hour and 14% gateway details