TL;DR: The FAA has stopped installing a new Juniper Networks router model and disabled more than 60 of the units already running on its air traffic control network, five days after identifying that hardware as the trigger for the nationwide ground stop that canceled more than 9,000 flights on November 13. A second router in Amarillo, Texas malfunctioned the same way. Investigators still don’t know what caused the failure, and the FBI has been called in even though officials say there’s no sign of a cyberattack.

What actually happened on November 13
Just before 4pm Eastern that Friday, the FAA ordered every departing flight in the country to stay on the ground. It was only the third time in the agency’s history that happened nationwide, after the September 11, 2001 attacks and a software failure in January 2023 that took down the system used to warn pilots about hazards. About 3,500 aircraft already in the air landed without incident, but on the ground the damage piled up fast: airlines canceled more than 9,000 flights that day alone, with thousands more scrapped over the following two days as the schedule struggled to recover before the ground stop was lifted that evening, roughly five hours after it began.
The cause traced back to a contractor, L3Harris, which was in the middle of swapping out old copper telecom lines for fiber at an FAA facility in El Dorado, Arkansas. That modernization work included installing a new router, the same kind of box most people only think about when their home connection drops and they go looking up the difference between a modem and a router. On the scale of a national air traffic control network, that one box turned out to matter enormously.
One router, over 1,000 connection attempts every few seconds
FAA Administrator Bryan Bedford has said the newly installed router began flooding the network with more than 1,000 connection attempts every four to five seconds, a surge that lasted roughly 45 minutes before the agency, unable to pin down the cause in real time, ordered the nationwide stop out of caution. Routers in this part of the FAA’s network carry radar feeds, flight plans and weather data between air route traffic control centers, so a malfunctioning one doesn’t fail quietly. It drags everything connected to it down with it.
The router model comes from Juniper Networks, the networking brand owned by Hewlett Packard Enterprise, and its rollout is part of a much larger project to retire roughly 2,000 aging Cisco routers across the FAA’s network. Bedford said the same model had already gone into 68 other sites without incident before the unit in Arkansas started broadcasting bad traffic into the system, and that a second router in Amarillo, Texas failed in a similar way around the same time, prompting the agency to disable both immediately. HPE has pushed back on the idea that its hardware was defective, with a company spokesperson telling reporters the routers performed as designed and did not fail, leaving the FAA and its own vendor publicly disagreeing about what actually went wrong.
Suspended installs, dozens of routers pulled, still no root cause
Five days on, the FAA has acted on what it does know even though it can’t yet explain why. The agency has suspended new installations of that router model entirely and disabled more than 60 of the units already live across its network, a sweeping response for hardware that passed roughly a year of pre-deployment testing. Bedford has been candid about the gap: investigators know which router caused the problem, not why it happened.
That uncertainty is why the response reaches beyond networking teams. Transportation Secretary Sean Duffy has said national security agencies told him the incident wasn’t a cyberattack, but he asked the FBI to help examine the Arkansas hardware regardless, a sign of how seriously the agency is treating a single piece of equipment that was able to ground the entire US aviation system. Democratic members of the Senate Commerce, Science and Transportation Committee are pressing the FAA for a public accounting of the failure and a concrete plan to prevent a repeat, a familiar ask for anyone who watched Microsoft spend weeks explaining a Microsoft 365 outage that knocked out Outlook, Teams and Copilot earlier this year without a quick answer of its own.
The modernization project isn’t slowing down
If anything, the FAA and Duffy are using the incident to argue for moving faster, not slower. Duffy has renewed his call for Congress to approve an additional $50 billion to finish modernizing the country’s air traffic control infrastructure, framing the copper-to-fiber transition that caused this outage as the cure for a much bigger risk: staying on decades-old equipment. He has argued that an old, rickety network only gets less safe the longer it stays in place, and that the real danger is delaying the upgrade rather than pushing through it.
That puts the agency in an odd position. It’s simultaneously pausing new router installs for safety and asking Congress for billions more to install thousands of additional routers of the same general type down the line. Both things can be true at once. The specific unit in El Dorado clearly had a defect worth understanding, and the broader network it’s replacing is old enough that a prior FAA-commissioned review flagged parts of it as still dependent on decades-old time-division multiplexing technology. Neither fact makes the other less urgent, and neither answers the question Congress actually wants resolved before it writes another check.
The bottom line
A router that passed about a year of testing and worked fine at 68 other sites was enough to ground every flight in the country for an evening and cancel thousands of trips, and five days later the agency still can’t say why. That gap between identifying the hardware and understanding the failure is the part worth watching here, not the router model number. Pulling more than 60 units and freezing new installs buys time and looks decisive, but it doesn’t close the gap, and Congress is unlikely to hand over $50 billion to finish a modernization project before someone can explain with confidence what went wrong with the last phase of it.



