A bug lying dormant since July, a router in the middle of an upgrade, and one bank that dragged others down: that’s what explains a week of outages.
Between November 12 and 17, 2026, Microsoft 365, U.S. air traffic control, Chase online banking, and the NSA’s classified network went down. On social media, suspicions of a coordinated attack took hold before any official explanation emerged. The causes reported by those affected tell a different story: none point to an attacker, and almost all began with a change the operators themselves made to their systems.
That coincidence is no small thing. The Microsoft defect arrived with a July patch, the FAA failure occurred while copper wiring was being replaced with fiber optics, and the NSA outage followed routine maintenance. In all three cases, the problem stemmed from an upgrade, not an intrusion.
What connects the week’s events is not a common enemy, but a shared condition: systems that cannot be shut down for repairs and that millions of people depend on at once. Any change is made while they are running, and if it goes wrong, the error affects the entire system.
That doesn’t make what happened any less serious: failed maintenance caused damage comparable to an attack, with thousands of flights canceled and users in dozens of countries left without email for nearly a day. Three outages, one common cause
According to Microsoft’s preliminary report, the outage resulted from a combination of three problems: a software defect that no one had detected, increased traffic, and reduced server capacity. The defect had been latent since July, when a patch was deployed for the servers that act as the service’s gateway; at the time, it passed load tests without showing any failures.
At the FAA, the cause was more physical. In its official explanation, the agency confirmed that a telecommunications failure forced it to halt departures nationwide, and that about 12,462 flights were canceled that Friday. According to Aviation Week, the problem centered on a router that failed during the migration to fiber optics, and the investigation is ongoing to determine whether the cause was a contractor error or defective equipment.
Chase, for its part, attributed its November 16 outage to a software error in its own hardware systems. In the case of NSA Net, the agency has offered no public explanation, although a government official attributed the failure to routine maintenance.
It’s best to read all of this with caution. Each explanation comes from the affected organization itself, none has undergone independent review, and nearly all are preliminary: Microsoft’s final report is expected on November 20, and the investigation into the FAA contractor’s work remains open.