FAA tests $12B SMART AI

Air traffic control is the domain here — and the real issue is not some sci-fi autopilot taking over the tower. The FAA is testing a planning tool called SMART, short for Strategic Management of Airspace Routing Trajectories, that tries to see congestion before it snowballs into delays. Three companies are in the running to build it out: Palantir, Thales, and Air Space Intelligence. The agency says the contract award is coming “soon,” and the point is to predict traffic flows and adjust departure times before the system clogs up. ### What is SMART actually supposed to do? Basically, SMART is not there to tell controllers how to separate planes in real time. It is there to organize demand earlier — before aircraft push back, before one weather snag ripples across a region, before a busy hub turns into a nationwide mess. Thales described it as a way to manage demand on airspace and airports proactively, and the FAA framed it as a system that predicts traffic flows and tweaks departures to resolve conflicts. (politico.com) ### Why does the two-hour window matter? Current traffic management tools are much better at the short game than the long game. The selling point for SMART is that it could move the FAA from reacting to problems minutes ahead to planning around them much earlier. That matters because once planes are already queued, crews are timed out, gates are occupied, and connections are missed, the delay machine feeds itself. The whole idea is to make fewer bad situations start in the first place. (politico.com) ### Is this replacing controllers? No — at least not in the way people hear “AI” and immediately worry about. The companies involved have been explicit that SMART is not meant for safety-critical separation decisions. Human controllers still handle the second-by-second job of keeping aircraft apart. SMART sits upstream from that — more like a traffic forecaster than a digital controller. Think of it as trying to meter cars onto a highway before the jam forms, not steering each car through the jam. (thenextweb.com) ### Why is the FAA reaching for this now? Because the broader system is under strain from both old tech and staffing pressure. FAA handles about 44,360 flights a day on average, and NATCA says the agency has nearly 11,000 fully certified controllers. At the same time, the FAA’s own modernization push now includes a brand-new air traffic control system targeted for the end of 2028, with replacements for radar, telecom, software, and other core infrastructure. SMART fits into that bigger rebuild. (politico.com) ### Wait — wasn’t NextGen supposed to modernize this already? Yes, and that is part of why this story lands. NextGen has been the FAA’s long-running modernization program for years. But GAO said in 2023 that progress was mixed, milestones slipped, and FAA had already reported spending just over $14 billion through fiscal 2022, with at least $35 billion in federal and industry costs projected through 2030. So SMART is not a fresh start from zero — it is another layer in a modernization effort that has already been expensive and uneven. (faa.gov) ### What’s the catch? The catch is integration. It is easy to describe an AI layer that predicts bottlenecks. It is much harder to plug that layer into the FAA’s dense patchwork of existing systems, airline schedules, weather inputs, and controller workflows without creating new confusion. Even Politico’s reporting noted that what SMART looks like in practice — and how it fits into the current mesh of systems — is still unclear. (faa.gov) ### So what should travelers watch for? Not instant magic. The near-term news is a test and a vendor competition, not a nationwide rollout. If SMART works, the payoff would be fewer cascading delays at the busiest hubs during bad weather and peak travel days. But the bigger signal is this: the FAA is moving from just replacing old hardware to trying predictive software on top of it. That is a much more ambitious bet. (politico.com) ### Bottom line? SMART is the FAA admitting that the hard part of modern air traffic control is no longer just tracking planes — it is managing complexity early enough to stop congestion from spreading. If the system works, passengers may never notice it. That would be the point. (politico.com)