One Cut Cable, 5,600 Delayed Flights: Give your business true redundancy

On Monday 21 September 2026, construction workers in New Jersey accidentally severed a fiber optic cable, and within hours, air travel across the entire US Northeast ground to a halt. Ground stops and delays hit Newark, JFK, LaGuardia, Philadelphia, and Boston. FlightAware tracked more than 5,600 delayed or cancelled flights nationwide, including roughly 1,200 at the three major New York-area airports alone, with over 100 flights diverted elsewhere. The timing could hardly have been worse: it landed the day before the UN General Assembly, with scores of world leaders due to fly into New York.

The mechanism was almost mundane. A construction crew cut a cable while digging, after site markers had reportedly been placed incorrectly. That cable fed into the Terminal Radar Approach Control (TRACON) system, the infrastructure controllers rely on to guide aircraft during takeoff and landing. Lose that link, and the safe answer is to stop flights, not risk them.

Two connections can still have one point of failure

Here's the part of the story that matters far beyond aviation: the cable the construction crew severed was the backup circuit. The FAA's primary circuit had already failed, independently, for unrelated reasons, and the fault had gone unnoticed until controllers tried to fail over to the backup, only to find it freshly cut as well.

In other words, this wasn't a single point of failure. It was two points of failure, both down, at the same time, and nobody knew until it was too late. On paper, the FAA's air traffic control network had redundancy. In practice, it had two cables in the ground, likely following similar routes, both vulnerable to the same category of risk: physical damage to buried fiber. Redundancy that shares a failure mode isn't redundancy, it's just a second copy of the same weakness.

This is not a criticism reserved for the FAA. It's a pattern that shows up constantly in enterprise and public-sector networks: a "resilient" design that turns out to be two ISPs reselling the same underlying fiber, two data circuits sharing a duct, or a primary and failover line that both terminate in the same street cabinet. It looks diverse on a network diagram. It isn't diverse where it counts, in the ground, in the duct, or in the last mile.

What real network diversity actually requires

A backup connection only earns the name if it fails independently of the primary, a different path in the ground, a different access medium, a different provider's infrastructure, and ideally automatic detection rather than discovering the fault only when someone reaches for the backup and finds it gone too. As we like to put it: air and fiber don't share a duct.

This is precisely the gap Blue Wireless's Global Managed LEO (GML) and Global Wireless Access (GWA) services are built to close. Rather than layering a second fixed-line circuit on top of a first, which is exactly the setup that failed in New Jersey, GML and GWA pair a site's primary connection with genuinely independent access media: LEO satellite and cellular, sitting on infrastructure that shares no physical path, no duct, and no digger risk with a buried fiber line.

That matters most for anything where "we'll get it fixed by end of day" isn't good enough: air traffic and transport infrastructure, financial and payment networks, emergency and safety systems, retail estates that can't afford a till system going dark. With automatic failover between cellular and LEO, a fault on one path is detected and routed around within seconds, well before it becomes a story.

The Northeast outage resolved within a day, and the world leaders arriving for the UN General Assembly mostly made it in regardless. But it's a useful, low-cost reminder, because this time the damage was inconvenience and delay, not something worse, that a network is only as resilient as its least diverse point of failure. Anyone auditing their own "redundant" circuits this week might find it worth asking a simple question: if a digger hit our primary link tomorrow, would our backup notice, and would it care?