FAA Ground Stop: SMART AI Ruled Out, Legacy TDM Blamed
Airport departure board showing cancellations during the FAA ground stop

Three days after the FAA expanded an AI-supported traffic tool to all 22 of its air route
traffic control centres, it ordered the first nationwide FAA ground stop in years. The timing invited
the obvious question, and officials have been answering it all week: no, the AI did not do it.

That answer is more interesting than the conspiracy. The cause was a 1980s-era telecom
network, a router deployment, and a flood of connection attempts the legacy gear could not
absorb.

What this covers:

What the FAA Ground Stop Was Not

It was not a cyberattack. The FAA states that on all available evidence there is no indication
of an attack or an outside actor.

It was not the SMART system. The agency’s answer is blunt: the two systems are not
related, and SMART does not feed information into the air traffic control system.

It was not the modernisation programme either. The FAA points at legacy TDM equipment and says
none of the upgraded towers that received recent modifications went down.

It was also not a total loss of contact with aircraft. Some frequencies and radar feeds stayed
reliable throughout, which is how roughly 3,500 aircraft already in the air landed safely.

What Actually Caused the FAA Ground Stop

The FAA’s account is specific enough to be useful. Its legacy equipment became unstable
when data began flooding the FTI network at a rate of about 1,000 connection attempts every four
to five seconds.

That flood disrupted communications and radar feeds. The initial investigation points to two
locations as the source: a radar site in Amarillo, Texas, and a radio communications site in
El Dorado, Arkansas.

L3 Harris, the FAA’s network service provider, had been working at the El Dorado site,
including deploying a new router. When the data inconsistencies appeared, the provider blocked
those two locations and the network stabilised.

As a precaution, the broadcasting function was disabled on all the new routers.

The FAA describes the underlying mechanism as instability in the legacy TDM system, the copper
infrastructure being retired as it reaches obsolescence. Two sites and one router deployment were
enough to destabilise a national network.

Timeline of the FAA ground stop, from 3:30 pm to 8:15 pm ET
Four hours of stops at 3:30 pm, on a legacy network built from copper being retired.

How One Router Destabilised 17 Centres

The instability began at 17 of the 22 Air Route Traffic Control Centres and spread into
several Terminal Radar Approach Control facilities.

The ground stop went out at about 3:30 pm ET on Friday 13 November. Traffic resumed at 8:15 pm
the same day, after the agency held the system stable for a period rather than restarting
immediately.

The FAA has been asked why router replacement work happened during the day rather than
overnight. Whatever the answer, the sequence is a familiar one in any organisation.

A routine vendor change, a defensive reaction at the network layer, and an operational
consequence far larger than the change itself.

Dealing with this at your office? Talk to a local IT expert →

What SMART Actually Does

The Strategic Management of Airspace, Routes and Trajectories system, SMART, expanded to all
22 centres on 10 November after a trial in the National Capital Region that began in September.

It aggregates data on weather, flight paths and traffic to advise on congestion, with the
advertised benefits being fewer delays and less fuel burned.

The FAA’s key characterisation is that it is a data aggregator and does not feed
information into the air traffic control system. That is the difference between an advisory tool
and a control system, and it is why the outage did not originate there.

The FAA Ground Stop and the AI Trust Problem

The public reaction was not irrational. A new AI system goes live across the country on
Monday, the aviation system melts down on Friday, and people draw a line between the two.

What makes that instinct worth taking seriously is how organisations usually introduce
automation: quietly, alongside ageing infrastructure nobody has finished replacing, with the
reassurance that the new tool is only advisory.

When something breaks, the newest change gets suspected first. That is a communication problem
as much as an engineering one, and it is the reason a change log matters more than a press
release.

FAA Ground Stop: Cancellations and Recovery

The FAA reports approximately 12,462 cancellations on Friday 13 November, including 9,082 by
major airlines, while noting that not every cancellation was directly caused by the ground stop.

Officials describe recovery as slow and steady, with airlines repositioning aircraft and crews
after several hours without departures. The agency has said it and L3 Harris continue to
investigate root cause.

For travellers, that tail is the part that matters: the four hours of the stop, then the two
days of cancelled flights and displaced crews that follow.

FAA Ground Stop: What South Florida Businesses Should Take Away

South Florida runs on Miami International and Fort Lauderdale-Hollywood. When the national
system stops for four hours, the effect reaches staffing, shipments and client meetings well
beyond the airport.

Four things this incident demonstrates, in language that applies to any business:

  1. Your oldest component is your real risk. The FAA’s copper TDM network,
    not its newest AI tool, caused the outage. Inventory anything on the way out and know what
    breaks when it does.
  2. Vendor change windows need oversight. A router deployment at one site
    triggered a national reaction. Ask what your providers are changing, when, and how they roll
    back — the same discipline we describe in
    why vendor patch windows need a schedule.
  3. Alert storms look like attacks. One thousand connection attempts every few
    seconds was enough to destabilise the network. Tune thresholds so a torrent cannot masquerade
    as normal traffic.
  4. Test the human response, not just the failover. The technology held. What
    needed a plan was the four-hour decision to stop everything, and the two days afterward.

If travel disruption is a continuity risk for your business, the preparation looks a lot like
continuity planning for South Florida businesses: named owner, alternative
suppliers, and a way to reach staff when the usual channel is degraded.
Book a call with Nextek IT.

FAA Ground Stop Questions

Did the AI system cause the FAA ground stop? No. The FAA states the two systems are
unrelated and that SMART does not feed information into the air traffic control system. The
vendor and the Transportation Secretary have said the same.

Was it a cyberattack? The FAA says that on all available evidence there is no
indication of a cyberattack or an outside actor.

What caused it, then? A flood of connection attempts against the legacy FTI
network, traced initially to a radar site in Amarillo, Texas and a radio site in El Dorado,
Arkansas, where a router deployment was underway.

Did aircraft lose contact with controllers? Not entirely. Some frequencies and
radar feeds remained reliable, and every aircraft landed safely.

How many flights were cancelled? Approximately 12,462 on 13 November,
including 9,082 by major airlines, per the FAA, which notes not all were caused directly by the
ground stop.

Is it safe to fly now? The FAA says the system is stable, operations are
returning to normal, and it is safe to fly.

Need IT that actually answers?

Nextek IT supports South Florida practices and businesses — medical, dental, legal, accounting — with managed IT, cybersecurity and HIPAA-focused support.

Talk to a Local IT Expert → Or call 954-780-7159

Sources:
the FAA’s questions and answers,
Scientific American on the SMART
rollout
,
Flying Magazine on the 22-centre
expansion
,
AVweb’s report on SMART being ruled
out
, and
The Miami Times.
Timeline, cause and cancellation figures are as published by the FAA.