The U.S. air traffic system runs on infrastructure from the 1960s. That sentence is not a rhetorical flourish. The FAA’s own announcement frames the $12.5 billion modernization effort as a replacement program covering more than 4,600 sites nationwide, equipment that has seen air traffic grow twelvefold since it was first installed. Against that backdrop, the September 2026 debut of the FAA’s Strategic Management of Airspace Routing Trajectories system, known as SMART, is the most compressed and consequential modernization bet the agency has made in a generation.
What SMART actually does
The name is bureaucratic, but the concept is straightforward. SMART uses artificial intelligence and cloud computing to predict and deconflict aircraft trajectories, not in real time as controllers do today, but strategically, potentially months in advance. Aviation Week’s reporting from the FAA/EASA annual safety conference on 15 June quotes FAA Senior Certification Advisor Steve Fulton explaining the system’s purpose plainly: “We’re going to introduce this strategic coordination of air traffic at high altitude. When the airplanes are en route, we’ll be deconflicting the projected trajectories from each aircraft, and we’ll be providing updated arrival times.” The system builds on trajectory-based operations, a concept the FAA has been developing for years, and sharpens it with AI and cloud computing to improve predictions of where aircraft will be in space and time. The September demonstration will be bounded: en route airspace at 24,000 feet and above only. That is a deliberate choice. High-altitude en route is the least congested and most forgiving environment in which to validate a new concept before expanding it into terminal airspace, where the margins are tighter and the consequences of a prediction error are more immediate.
The controller question
One thing Fulton was careful to establish is that SMART is not a controller replacement. The framing he used is worth quoting directly: the flight crew on a modern aircraft has moved from being a “real-time integrator of information to a manager of the system, intervening where necessary.” The FAA wants to make the same shift possible for air traffic controllers, who have historically been kept in a real-time integration role by the limitations of the tools available to them. SMART is designed to absorb the workload of trajectory deconfliction so controllers can manage exceptions rather than process every conflict from scratch. Route changes generated by the system will be pushed to aircraft through existing connectivity options, including airborne reroute. The FAA has been explicit that controller authority is preserved throughout.
Who builds it
The FAA has not formally announced a contract award, but Flying Magazine reported, citing multiple people familiar with the selection, that Air Space Intelligence (ASI) is the likely recipient. ASI’s Flyways AI platform already manages more than 40% of U.S. air traffic by its own account. Palantir and Thales developed competing demonstration versions. The choice carries industrial-policy weight well beyond the immediate contract: ASI is an AI-native company built specifically for airspace management, while Palantir and Thales bring different institutional relationships with government and defence customers. Whichever way the award goes, the other two vendors will not simply disappear from the ATM technology market.
The pace problem
Here is the detail that stands out most, given what is at stake in ATM modernization. Fulton acknowledged at the FAA/EASA conference that some airlines and operators are telling the FAA it is moving too fast. His exact words: “It’s not often you hear industry say to the FAA, ‘You’re going too fast.’” That is a remarkable admission. The FAA has spent decades being criticized for moving too slowly on modernization. NextGen, the agency’s previous flagship modernization program, stretched across more than twenty years and delivered results that fell well short of early projections. The fact that airlines and operators are now flagging concern about the pace of SMART’s deployment suggests the September timeline is genuinely aggressive. Ambitious undersells it. Compressing what would normally be a multi-year validation cycle into months means the FAA needs to “validate very quickly that the core concept of being able to strategically deconflict the entire national airspace system using cloud computing” actually works at scale, in Fulton’s own framing. That is not a small ask.
The broader modernization context
SMART sits within a much larger program. The FAA’s Modern Skies website, launched in May 2026, tracks more than 10,000 individual modernization projects. The $12.5 billion envelope, funded through the One Big Beautiful Bill, covers 5,000 new network connections, 27,000 new radios, 450 digital voice switches, and 612 radars by end-2028. That is the physical layer: cables, radios, and radar heads. SMART is the intelligence layer sitting on top. Both need to work, and they are being built in parallel rather than sequentially. The staffing picture adds another variable. The FAA has been contending with a shortage of trained controllers for years, and while recruitment has been active, progress has been uneven. Introducing a system designed to reduce controller workload into a workforce that is already stretched is either very good timing or a complication, depending on how the transition is managed.
Why this matters beyond the U.S.
If SMART works, it becomes the reference architecture for AI-enhanced air traffic management globally. Eurocontrol has been moving in a similar direction with trajectory-based operations under its SESAR programme, and ANSPs in Europe, Asia, and the Gulf will be watching the September demonstration closely. A validated, deployed system gives the U.S. a concrete proof point that no one else currently has. If it stumbles, the consequences run in the other direction. Trust in AI-based ATC is not yet established with the flying public or with the controller workforce. A high-profile failure at this scale, in this environment, would set back the broader adoption of AI in safety-critical airspace management by years. The FAA is betting its credibility on a tight timeline, a not-yet-announced vendor, and a concept that has never been operated at national-airspace scale. Enough time around ATM performance data teaches you that the gap between a validated concept and a working operational system is where most modernization programs lose their schedule. September is close. The next few months will tell us whether SMART is the beginning of a genuine inflection in how airspace is managed, or another reminder that aviation infrastructure moves at its own pace regardless of what the announcement said.