FAA lifts nationwide ground stop after 4 hours, 827 flights diverted

Quick summary

The FAA lifted a nationwide departure ground stop at 8:15 p.m. EST on Friday, November 13, 2026, ending a restriction that began around 3:54 p.m. and lasted more than four hours. More than 3,500 aircraft were already airborne when the halt was issued, and at least 827 flights had been diverted by 7:05 p.m., according to Cirium data.

Investigators found no evidence of a cyberattack or AI involvement, though the underlying cause remains under investigation. Airlines warn cancellations will persist into the weekend as aircraft and crews reposition.

A four-hour halt that began with overloaded switches

The United States stopped departures across its entire airspace on Friday afternoon. Not weather. Not an attack. The switches and routers carrying air traffic control communications simply stopped keeping up.

The FAA issued the nationwide ground stop at roughly 3:54 p.m. EST and lifted it at 8:15 p.m. — four hours and twenty-one minutes during which no commercial departure left a US airport. Transportation Secretary Sean Duffy confirmed communications had been restored at air traffic control facilities, but cautioned that airlines needed more time before schedules returned to normal.

For anyone flying to, from, or within the United States this weekend, the practical takeaway is uncomfortable: a confirmed departure time is provisional until the operating carrier confirms the aircraft and crew are actually in position.

The failure traced back to telecommunications hardware inside the FAA’s own infrastructure. Officials pointed to overloaded switches and routers, and to a split between old and new equipment that decided what kept running. Newly installed digital systems rode out the incident. Older analog gear did not.

What the four-hour shutdown actually involved

More than 3,500 aircraft were already airborne when the order landed. That figure explains both the speed of the decision and the slowness of the recovery — every one of those flights still had to land, and every airport receiving them absorbed a schedule it had never planned for.

By 7:05 p.m. EST, Cirium had logged at least 827 diversions. Each diverted widebody can displace several hundred passengers and, more damagingly, a crew whose legal duty hours keep ticking while the aircraft sits on an apron it was never meant to visit.

The FAA’s official advisory recorded the stop as covering all destinations and facilities, with communications failure as the stated reason. Its preliminary public statement said the problem affected multiple Air Route Traffic Control Centers — so this was never a single-airport event.

Two boundaries deserve stating plainly. The FAA did not lose radar or communications entirely; exceptions were granted for medical flights, critical cargo shipments, and law enforcement operations. And this was one infrastructure fault, not a systemic collapse of the air traffic system — no airworthiness directive followed, and none was warranted.

As departures resumed, larger general aviation charter flights carrying 60 to 100 passengers got priority over smaller aircraft. That is a capacity calculation rather than a courtesy: bigger hulls clear more stranded passengers per departure slot.

The agency also moved quickly to separate its newest tool from the failure. SMART feeds recommendations to controllers but is not wired into STARS or ERAM, the systems that actually run terminal and en-route traffic. Three days before the outage, the FAA had announced that SMART was scaling up nationally.

For travelers connecting onward to Asia-Pacific, the knock-on effect is timing rather than routing — see current flights from North America to Asia-Pacific for the connections most exposed to a displaced US departure bank.

The weak point is older than the software

Friday’s outage drew attention to something the FAA has been quietly managing for years: a backbone that runs on a mix of legacy analog and newer digital equipment. When the two layers diverge under stress, the older one fails first — and that is precisely what officials described.

Agencies rarely volunteer that a new system is innocent unless they expect the question to be asked repeatedly. The FAA clearly expects it. AI adoption in air traffic control is under active scrutiny, and a nationwide ground stop is exactly the moment critics reach for the newest technology in the room.

The standard sequence from here is familiar to anyone who follows the agency: facility-by-facility resumption, a root-cause report, then a funding request aimed squarely at the hardware that broke. Watch whether the replacement timeline for those analog switches gets attached to a budget line — that, not the AI question, is where this story goes next.

How to protect a US itinerary this weekend

Facility-by-facility recovery means the ground stop is over but the disruption is not — aircraft and crews are still out of position, and your connection is the most vulnerable part of the trip.

• Check the operating airline, not the FAA. The agency’s ground-stop status clears faster than airline schedules rebuild. Confirm aircraft and crew are assigned before you leave for the airport.
• Protect the whole itinerary. If a US connection breaks, ask the carrier to rebook the full journey rather than accepting a partial fix that leaves you stranded mid-route.
• Confirm bag routing. Diversions scatter checked luggage. Verify the new tag destination matches your rebooked flight before you board.
• Build slack into tight connections. A 90-minute domestic-to-international transfer is not enough on a recovery day. Two hours is the working minimum.

Watch: the Department of Homeland Security’s next statement on the cyberattack investigation — it remains formally open, and a definitive finding would change how the FAA frames the incident.

Questions? Answers.