Infrastructure Matters · Aviation and network infrastructure · EH 98
🔊 Listen to this article
Listen to the full article. Playback starts when you press Play.
Category: Infrastructure
US ground stop: The cost of an air traffic control outage
EH 98 · Version 1.10 · Updated November 16, 2026: expanded technical guidance on updates and recovery testing.
Author: MAGNETiX
A nationwide halt to departures on November 13, 2026 severely disrupted US air travel. Eventhacker estimates the economic scale using flight statistics and transparent assumptions.

A network problem with nationwide consequences
On Friday, November 13, the US Federal Aviation Administration (FAA) suspended almost all departures nationwide for nearly five hours. According to Transportation Secretary Sean Duffy, work on the communications network caused network equipment to become overloaded. The fault affected 17 of 22 regional en route control centers. Authorities are investigating whether faulty hardware, operation or installation triggered the incident. The ground stop was a safety measure; aircraft already airborne were still allowed to land.
Even after the restriction was lifted, airlines had to reposition aircraft and crews. Citing FlightAware, Reuters reported about 9,200 cancellations and 7,500 delays on Friday, followed by more than 6,400 cancellations and 8,000 delays on Saturday as of 9:35 p.m. US Eastern Time. Flight schedules were expected to take until Sunday to recover.
Multiple backups, one shared point of failure?
A video by Two Bit da Vinci explains a flight as a chain of interdependent connections: a flight plan passes between control centers, radar data provides the traffic picture, and remote radio stations connect controllers with cockpits. Its central argument is that multiple backup connections cannot protect against every disruption if they share the same underlying network infrastructure. The video describes the FAA's FTI telecommunications network and asks how independent its backup layers really are (from 06:27 and from 13:03).
The architectural principle can be illustrated by two doors leading into the same hallway: if one door is blocked, the other remains available; if the shared hallway is impassable, both are affected. A physically separate cable route therefore provides a different kind of protection from independently operated routing. New fiber does not automatically eliminate a shared cause of failure either. The key question is which faults can affect both the main and backup paths at once.
The video is an additional explanation, not a final investigation report. Its details on router requests, reconnection times, backup layers and the successor network FENS have not been independently confirmed here against primary technical documentation. We do not use it to establish a final diagnosis or additional loss figures. The actual independence of the systems and the precise sequence of failures remain questions for investigation.
Why not use the internet with a VPN as a fallback?
The obvious question is whether an encrypted connection over the public internet could provide an additional, independent fallback path. A VPN can secure connections. On its own, however, it does not establish whether that path remains reachable when the main system fails or meets the requirements of the application. Such a solution would need to be assessed particularly for latency and its variation, packet loss, available capacity, switchover time and the required safety assurance.
Actual separation would be equally important: do both paths use the same access connection, network operator, routers or power supply? Does switchover still work if the shared control system itself is disrupted? For the FAA incident, the sources assessed here do not establish whether such an internet fallback existed, was tested, or why it might not have been used. Internet plus VPN is therefore a specific question for the operators, rather than an immediate solution that can be inferred from the video.
Based on these flight figures, our model suggests an economic burden of approximately US$0.6 billion to US$1.4 billion. The middle scenario is around US$1 billion. The estimate concerns the consequences of the US ground stop following the FAA communications infrastructure outage. They are not a confirmed loss total.
What our estimate is based on
Not a confirmed loss total: The US$0.6–1.4 billion range is a rough scenario calculation by Eventhacker. It combines assumed additional airline costs with the monetary value assigned to passengers' lost time. It is neither an official accounting nor a verified payout total or a loss fully attributable to the FAA outage.
| Input | Value | Basis |
|---|---|---|
| Cancellations / delays | 15,600 / 15,500 | Rounded flight totals for Friday and Saturday; Reuters citing FlightAware, as of Saturday, 9:35 p.m. ET |
| Passengers per flight | 120 | Our model assumption, not a measured load factor |
| Passenger time value | US$47 per hour | Model input based on the Airlines for America reference |
| Additional net airline costs per cancellation | US$10,000 / 20,000 / 30,000 | Our assumption: low / middle / high |
| Time lost per cancellation | 4 / 6 / 8 hours | Our assumption: low / middle / high |
| Time lost per delayed flight | 1 / 2 / 3 hours | Our assumption: low / middle / high |
We use rounded figures of 15,600 cancellations and 15,500 delays across Friday and Saturday, an average of 120 passengers per flight, and a time value of US$47 per passenger-hour. The time value is based on Airlines for America. Passenger numbers, time lost and airlines' additional net costs are explicitly model assumptions.
| Item (US$ million) | Low | Middle | High |
|---|---|---|---|
| Additional airline costs for cancellations | 156 | 312 | 468 |
| Time lost through cancellations | 352 | 528 | 704 |
| Time lost through delays | 87 | 175 | 262 |
| Total (rounded) | 595 | 1,015 | 1,434 |
In the middle scenario: 15,600 cancellations × US$20,000 in assumed additional net costs = US$312 million. Added to this are 15,600 × 120 × 6 hours × US$47 = US$527.9 million in time lost through cancellations, and 15,500 × 120 × 2 hours × US$47 = US$174.8 million through delays.
The range assumes US$10,000–US$30,000 in additional costs per cancellation, four to eight hours lost for cancellations and one to three hours for delays. The flight statistics may include other causes of disruption; we cannot attribute every flight to the FAA outage. Rebooking may result in double counting, and actual load factors and avoided operating costs were not determined individually. Time lost does not automatically constitute a compensable monetary loss. Cargo impacts, missed business opportunities and technical restoration costs are not separately included. The model provides context, not a final accounting of losses.
Two disruptions on the same day — no proven connection
Notably, the ground stop coincided with the ongoing Microsoft 365 outage on November 13. Based on their findings so far, US authorities treat the FAA communications disruption as a separate event. A causal link to Microsoft has not been established; timing alone does not support that conclusion. Eventhacker covers the Microsoft outage and its effects at Phantasialand in a separate article. The estimated aviation losses above are not added to Microsoft's losses.
Updates, upgrades and a fallback that can be tested
CERT-EU recommends prioritising patches for exposed systems, controlled updates for critical assets, network segmentation, explicit recovery time and recovery point objectives, and documented restoration tests. These are general security recommendations, not findings about the FAA incident (Security Guidance 22-001).
Our practical test plan for venues: record versions and configurations before a change, define an abort criterion and a tested recovery procedure, and update only a limited part of the system first. Then deliberately disconnect the primary connection during an agreed test window. Check whether ticketing, point-of-sale systems and production communications still meet their agreed operating requirements on the fallback path. Record detection time, switchover time and lost transactions. A reachable router alone does not establish that the service works. Also test the return to normal operation.
Define the maximum acceptable interruption and the amount of data that may be lost separately. A backup file is useful only if it can be restored within those targets. The exact limits must follow the venue's operational and safety requirements. This test plan is an editorial recommendation; it does not demonstrate that the FAA had or lacked any particular fallback.
What the events industry can learn
The case shows how disruptions in central networks can affect dependent operations. Staged changes, independent fallback paths, monitoring for network overload, tested recovery plans and clear criteria for aborting a change also matter for festivals, clubs and venues. Particularly in safety-critical systems, faults must remain locally containable.
In practice, this means that a second connection alone is not enough for ticketing, point-of-sale systems and production communications. Operators need to examine which dependencies both paths share, which functions must continue during an outage, and whether switchover has been tested under realistic failure conditions. The video suggests limiting excessive network requests as a protective measure; without a technical investigation, this does not establish which specific protection was missing at the FAA. The benchmark remains a demonstrated fallback operation for the system in question.
Infrastructure Matters. Eventhacker — More show. Less chaos.
Sources and data
Two Bit da Vinci: “The US Grounding Flights – It's Worse Than We Thought…” — fully transcribed locally and assessed editorially on November 16, 2026. Relevant passages: 06:27–08:58 (network structure and modernization), 09:51–11:07 (suspected sequence of failures), 13:03–16:23 (shared failure causes and protective measures). Technical details have not been independently confirmed. The internet-with-VPN question is an additional editorial question for investigation.
Reuters, November 14–15, 2026 — FlightAware figures as of Saturday, 9:35 p.m. ET, and statements by authorities.
Reuters, November 14, 2026 — investigation into the technical cause.
GeekWire, November 13, 2026 — preliminary assessment by US authorities regarding the absence of an established Microsoft connection.
Airlines for America — reference for passenger time values.
Editorial note: Please send corrections, additions and further substantiated loss data to redaktion@eventhacker.de.