When the cloud went quiet
A two-day Microsoft 365 failure left millions unable to send an email. For a region that has quietly moved its banks, ministries and classrooms onto one company's servers, the lesson is not technical. It is about who holds the keys.

REDMOND. It began the way these things always do, with a small thing that would not open. On Thursday, November 12, 2026, users across the Americas found they could not reach their Exchange Online mailboxes, according to SecurityWeek, and within hours the trouble had spread.
Outlook would not send. Teams calls dropped. By the time it was over, the failure had pulled in SharePoint Online, Microsoft 365 Copilot and even Defender XDR, the security tool many companies rely on to watch for intruders.
The disruption ran into Friday. As the Financial Times put it, Microsoft spent those two days racing to restore services after a global outage, and only by the weekend, Windows Central reported, was the company able to say that everything was up and running again.
For most people it was an irritation. For a great deal of the Caribbean's working day, it was a stoppage. Walk through the region's economy and notice how much of it now speaks through Microsoft's pipes.
Commercial banks run their internal mail and compliance on Outlook and Exchange. Government ministries schedule, file and sign on the same platform. Law firms, accounting practices, hotels taking group bookings, the universities running term on Teams, the small exporter confirming a shipment by email, all of them, for two days, were shouting into a line that was not connected.
When the mailbox will not open, the invoice does not go out, the letter of credit does not clear, and the customer in Miami or London assumes you are simply not answering.
What is striking about this outage is what caused it, which is to say, what Microsoft could not immediately say caused it. There was no hacker, no act of sabotage. The company disclosed no evidence of a cyberattack.
By its own account, reported by SecurityWeek, a large number of its servers simply became unavailable, failing the automated health checks that are supposed to keep the system honest, while overall traffic had not even risen. The machines built to guard the system turned on it. Microsoft logged the mess under two incident numbers, EX1550543 for the email failure and MO1550821 for the wider collapse, and set about testing a workaround.
It is worth noting the timing. The outage landed two days after the company's monthly Patch Tuesday, which this month fixed more than a thousand separate vulnerabilities, a reminder of how enormous and how fragile this single piece of global plumbing has become.
Here is the part the Caribbean should sit with. None of that plumbing is ours. We did not build it, we cannot inspect it, and when it breaks we cannot fix it.
We can only wait, like everyone else, for an engineer in Redmond to find the faulty valve. There is a genuine efficiency in that bargain, and it would be dishonest to pretend otherwise. No Caribbean government could build or secure an email system to the standard of Microsoft's, and the move to the cloud has given small institutions tools that once belonged only to giants.
But efficiency and resilience are not the same virtue, and a region that has lived through hurricanes knows the difference better than most. Resilience is what you have when the single thing you depend on fails.
Right now, too many of our institutions have no answer to a simple question: what is your plan for the day the cloud goes quiet? A bank that cannot email is a bank that cannot serve its customers. A ministry that cannot file is a government that pauses.
The outage was global, so no one could accuse the Caribbean of being singled out, but that is precisely the trap. When everyone depends on the same provider, everyone falls down the same hole on the same afternoon, and a small economy with thin margins feels the lost day more sharply than a large one.
The fix is not to abandon the cloud, which would be both impossible and foolish. It is to stop treating a single foreign vendor as if it were a public utility that cannot fail. That means boring, unglamorous things: a second communications channel that does not run on the same platform, contracts that put real money behind uptime promises, data that an institution can still reach when the main door is locked, and leaders who ask their technology officers the uncomfortable questions before the outage rather than during it.
Concentration is a risk even when the thing you are concentrated in is excellent.
For two days this week, the region discovered how much of its ordinary business now rests on infrastructure it neither owns nor controls. The servers came back, the mail flowed, and by Monday most will have forgotten. The ones who remember, and who build a little independence while the memory is fresh, are the ones who will keep working on the next afternoon the cloud decides to go quiet.
Read more at The Caribbean Ledger ›