By Vincent E. Bianco III
ATC Correspondent, Leeham News & Analysis
May 31, 2026, © Leeham News:
The four-phase regulatory Strategic Management of Airspace Routing Trajectories (SMART) architecture LNA articulated in our April 30 analysis distinguishes among the deployment layers at which automation can enter Air Traffic Management (ATM).
The FAA’s Strategic Management of Airspace Routing Trajectories is an important element for future Air Traffic Management. Credit: FAA.
Phase 1 is data aggregation: the system ingests data, performs analysis on it, and produces output that informs human decision-making, with no real-time control authority.
Phase 2 is advisory at non-towered airports, where the system can make recommendations to pilots in the absence of a human controller, with the pilot retaining decision authority.
Phase 3 is certified advisory at staffed facilities, where the system can issue advisories that human controllers and pilots execute, with the regulatory architecture explicitly authorizing the advisory authority.
Phase 4 is the National Airspace System (NAS) infrastructure integration, in which the system becomes part of the air traffic management infrastructure and has the authority to coordinate flight paths in real time.
Airborne Collision Avoidance System X (ACAS X) has been operating in Phase 3 territory since the late 2010s. The cockpit functions as the staffed facility; the pilot is the staff. The Resolution Advisory is the certified advisory output. The pilot executes the maneuver. The authority structure is that during the conflict event, the system’s advisory overrides the ATC instruction, and the pilot is required to follow the system. The Phase 3 deployment has worked because the architectural discipline named in Part 1 of this analysis was applied correctly: bounded scope, bounded time, bounded authority, conditional on equipage, with verification commensurate to the safety stakes.
Procuring SMART
SMART, as currently scoped by the Federal Aviation Administration (FAA) and described by Transportation Secretary Sean Duffy at the Semafor Modern Skies Summit on April 17, is structurally a Phase 1 deployment. The system extends air traffic conflict prediction from the current 15-minute window to two hours by ingesting flight schedule, weather, and airspace state data, running 4D modeling on the aggregate data, and producing recommendations for controllers on scheduling adjustments that could prevent downstream conflicts. The recommendations are advisory. The controller retains decision authority. The pilots and dispatchers in the loop between the recommendation and the actual flight operation provide the human verification layer. The architectural placement is correct for a Phase 1 deployment.
The architectural placement is not the analytical concern with SMART. The procurement velocity is.
The FAA selected three vendors to compete for the SMART contract: Palantir Technologies, Thales SA, and Air Space Intelligence. Vendor selection is targeted for the end of this month. First operational deployment is targeted for September 2026. The total timeline from public announcement of the vendor competition to first operational deployment is approximately five months. Compare to ACAS X’s 15-year arc from research origin to Technical Standard Order (TSO) approval. Compared to NextGen, which has been 16% complete after $7.5bn invested over 15 years. The SMART velocity is anomalous for federal aviation procurement of safety-critical software, and the explanation for the velocity is political rather than technical.
Personal priority and political pressure
FAA Administrator Bryan Bedford, confirmed by the Senate in July 2025, has made SMART a personal priority and is driving the program from inside the agency. The Department of Government Efficiency (DOGE) applied external pressure, with personnel visiting air traffic control facilities and Elon Musk publicly committing to “rapid safety upgrades.”
Project Lift, a separate initiative, is directing FAA funds toward upgrading network communications. The political layer compresses normal procurement timelines because the political principals want the deployment to be visible before the political cycle moves on. The LaGuardia Airport collision in March created public urgency that the political principals want to address through visible action rather than through the structural reforms that the National Transportation Safety Board (NTSB) final report on the DCA collision identified as the actual institutional pathology.
The political pressure does not by itself indict the SMART program. It does raise the substantive policy question being suppressed by the procurement velocity itself: does the SMART procurement process include the verification discipline that made ACAS X successful, or is the political compression of the timeline the artifact that obscures whether the verification work is being done properly?
Company experience
The vendor positioning offers some signal. Air Space Intelligence has the only deployed AI flight-path optimization system at a major US carrier, Flyways AI, which has been advising Alaska Airlines dispatchers since 2021. This five-year deployment history at a Part 121 carrier is the closest analog to the verification work that would be appropriate for a federal deployment, even if the operational context is different.
Palantir has the deepest existing FAA integration through Foundry platform deployment and prior data-modernization sole-source contracts; the integration depth is real, but does not by itself constitute verification of the AI logic.
Thales has 85 years of air traffic management heritage and a dominant European ATM market position; the institutional credibility is real, but the U.S. deployment context is unfamiliar terrain.
None of the three vendors has the equivalent of MIT Lincoln Laboratory’s 15-year verification arc for ACAS X. Whichever vendor wins the SMART contract will be deploying a system on a procurement timeline that does not match the verification discipline that successful aviation-safety AI has historically required.
If SMART works as advertised, the failure mode is contained: flights run on slightly different schedules than they would have otherwise, occasional false predictions that controllers can override, and a marginal improvement in throughput.
The Phase 1 architectural placement is the worst case. If SMART does not work as advertised, the most likely outcome is that controllers learn to ignore the advisories, the system becomes shelfware, and the FAA spends the next decade deploying NextGen 2.0 while the underlying staffing and equipment problems remain unaddressed. The institutional risk is not catastrophic. The institutional risk is that the political theater around SMART consumes the bandwidth that should be addressing the harder questions.
Part 3, The Architectural Discipline Not Being Applied to the Harder Cases, will appear next Sunday.
Vincent Bianco.
Vincent E. Bianco III is a 37-year aviation safety professional and the principal of Marivin Consulting Services LLC. He spent 23 years inside the FAA Air Traffic Organization — including six years as an Operations Supervisor and Air Traffic Manager — followed by FAA Headquarters contractor tours in GPS Navigation (WAAS/LPV) and En Route procedures (RVSM), and most recently eight months as Senior Program Manager for the Leader Capability and Proficiency program at Boeing Commercial Airplanes. He serves as ATC Correspondent for Leeham News and Analysis. linkedin.com/in/vebianco3
Related
