Today, September 21, 2026, the FAA switches on SMART — Strategic Management of Airspace, Routes, and Trajectories — at Reagan National, Dulles, and BWI. It’s the first AI system the agency has deployed that’s built to predict a delay before it happens rather than just manage the mess after. That’s a real shift. Every strike and weather event we’ve covered on this site this year — Italy’s ENAV walkouts, Charleroi’s 24 nights of controller strikes, O’Hare’s flood-driven cascade — played out the same way: something breaks, and the system reacts. SMART is the FAA’s bet that some of that can be caught weeks out instead.

Built by Air Space Intelligence under an $875 million, 12-year FAA contract, SMART isn’t a replacement for controllers and it isn’t live nationwide. It’s a 90-day trial at three airports in one metro area. Here’s what it actually does, what it doesn’t, and why the Washington region is where the FAA chose to find out.

Quick Verdict

Factor	Detail
System	SMART — Strategic Management of Airspace, Routes, and Trajectories
Launches	September 21, 2026
Where	Reagan National (DCA), Dulles (IAD), Baltimore/Washington (BWI)
Built by	Air Space Intelligence
Contract	$875 million over 12 years, awarded June 2026
What it does	Flags traffic conflicts up to 45 days out using schedules, weather, and runway capacity data
Current phase	90-day trial comparing predictions against real outcomes
Replaces controllers?	No — a decision-support tool, humans still run the airspace
Nationwide rollout target	End of 2028

What Is SMART, Actually?

SMART is a cloud-based platform that gives the FAA and airlines a shared, data-driven view of the National Airspace before flights ever depart. It continuously ingests airline schedules, weather forecasts, runway and airport capacity, and other operational constraints, then flags where those inputs are on a collision course — not literally, but operationally: too many flights aimed at too little runway, at the same time, in weather that’s going to make it worse.

That’s the pitch, anyway. The FAA has been candid that this is a test, not a finished product. NPR’s coverage of the rollout quotes FAA Administrator Bryan Bedford putting the current state of air traffic management bluntly: “The way we manage traffic today is chaotic, and we can do better.” That’s not spin. It’s the actual justification for spending nearly a billion dollars on this.

How Far Out Can SMART Predict a Delay?

Up to 45 days, according to reporting on the system’s capabilities. In practice, that window works in three layers:

1. Weeks out — SMART compares published airline schedules against known runway and airspace capacity to flag structural overbooking before a single flight is delayed.
2. Days out — weather forecasts get folded in, sharpening which specific days and hours are likely to strain the system.
3. Hours out — real-time operational data (actual runway use, live conditions) refines the prediction into something close to what a controller sees on the day.

The 45-day figure is the outer edge, not the typical case. Think of it less as “SMART tells you your flight will be delayed six weeks from now” and more as “SMART tells the FAA and airlines that a Thursday afternoon in November is shaping up to be a problem, six weeks before it’s a problem” — which gives everyone involved time to adjust schedules or routes before the day arrives, instead of after.

The $875 Million Bet on Air Space Intelligence

Air Space Intelligence won the contract on June 22, 2026, alongside a companion system called Flow Management Data and Services (FMDS), which becomes the new data backbone for the FAA’s Air Traffic Control System Command Center. SMART is the forward-facing prediction layer; FMDS is the plumbing underneath it. Together, they’re a 12-year, $875 million commitment — multiple outlets, including Quartz, the Washington Times, and Breitbart, independently confirm both the figure and the contract length.

For context on scale: that’s roughly $73 million a year, spread across a system meant to eventually cover the entire National Airspace System — every major US airport, not just the three getting it today. TechCrunch’s coverage frames the contract as the FAA’s actual answer to a controller staffing shortage it can’t hire its way out of fast enough. SMART doesn’t add controllers. It’s supposed to make the ones already working the boards more effective at catching problems before they compound.

Why Washington’s Three Airports First?

DCA, IAD, and BWI aren’t a random pick. It’s the FAA’s home turf — DOT headquarters sits on New Jersey Avenue in DC, and the National Capital Region gives the agency tight oversight over a live test without the added complexity of, say, running it at Atlanta or O’Hare first. It’s also not a low-stakes choice: this is the same airspace where a deadly midair collision near Reagan National last year put air traffic control staffing and technology under a level of scrutiny it hadn’t seen in years. Rolling out a predictive AI tool here first isn’t subtle. It’s the FAA testing its highest-profile fix in the airspace where the pressure to get it right is highest.

What Happens During the 90-Day Trial

This is the part worth understanding before anyone assumes SMART is fully operational starting today. According to reporting on the rollout, the initial phase runs for roughly 90 days, and the goal isn’t to let SMART start rerouting traffic on its own. It’s to check its work.

The FAA and participating airlines will pull specific flights, compare SMART’s predicted routes and conflict flags against what actually happened, and build a track record before trusting the system with anything closer to real authority. Transportation Secretary Sean Duffy has described the approach as deliberately incremental — a slow, methodical introduction into live airspace rather than a switch-flip rollout. Controllers stay fully in charge of every operational decision during this phase. SMART suggests; it doesn’t direct.

That distinction matters for a simple reason: an AI system that’s wrong about a 45-day-out prediction costs nothing. An AI system that’s wrong about a live conflict at 2,000 feet costs everything. The FAA’s caution here is the correct instinct, not corporate hedging.

Predictive vs. Reactive: Why This Matters After a Year Like This One

We’ve spent most of 2026 covering the same story with different geography. Italy’s ENAV controllers walking out and grounding flights that never touch Italian soil. Charleroi’s 24 consecutive nights of strikes. O’Hare’s record rainfall cascading into a summer-long capacity cap that reshaped tens of thousands of itineraries. In nearly every case, the system found out about the problem the same moment travelers did — and everyone spent the following days digging out.

SMART doesn’t fix strikes, and it doesn’t fix weather. It’s not designed to. What it’s built for is the structural overbooking and capacity mismatches sitting underneath a lot of those cascades — the version of the problem where an airport was always going to be scheduled past what it can safely handle on a given day, strike or no strike. Catching that six weeks out instead of discovering it in real time, the way the new DOT delay rule reshuffles what airlines owe you after the fact, is a genuinely different kind of fix. One works on the front end. The other works on the back end, after your day is already ruined.

It’s worth being honest about the limits, too. SMART is a forecasting layer, not a staffing fix, and it doesn’t touch anything airline-specific like crew scheduling or maintenance delays. TSA rolled out its own AI agent, Ace, earlier this month to handle traveler questions at scale — a completely different problem, but the same instinct running through federal aviation right now: throw predictive and conversational AI at the parts of the system that are currently reactive by default.

What SMART Doesn’t Do (Yet)

• It doesn’t cover your airport. Unless you’re flying through DCA, IAD, or BWI, none of this touches your itinerary today.
• It doesn’t replace controllers. Every account of the rollout, from the FAA’s own announcement to independent reporting, is explicit that humans make the calls.
• It doesn’t fix weather or strikes. Those causes aren’t going anywhere, and SMART’s forecasting doesn’t reduce their impact once they hit.
• It doesn’t have a track record yet. The entire point of the 90-day trial is that nobody, including the FAA, knows yet how accurate SMART’s 45-day predictions actually are in practice.
• It isn’t nationwide. The FAA’s stated target for broader National Airspace System rollout is the end of 2028 — over two years out.

Who Should Actually Care About This Right Now

Flyers through DCA, Dulles, or BWI this fall. You won’t notice SMART directly — there’s no passenger-facing feature — but if the system does what it’s supposed to, fewer of your delays should come as a same-day surprise.

Anyone tracking how the FAA is handling the controller shortage. SMART is a technology answer to a staffing problem the agency has been open about not solving through hiring alone. Whether that trade works is exactly what this trial is testing.

Airlines scheduling into the National Capital Region. SMART’s predictions are only as useful as the schedule data feeding them, and TechCrunch’s reporting notes friction points around how competitive slot scheduling interacts with a system that’s now flagging some of those schedules as structurally risky before the season even starts.

Who Can Skip This For Now

Anyone not flying through Washington’s three airports this fall has nothing to act on yet. There’s no app, no passenger setting, no way to interact with SMART directly. If the 90-day trial holds up and the FAA expands it, that’s the point to start paying attention to whether it’s touching your home airport.

The Bottom Line

SMART is the FAA’s most direct attempt yet to get ahead of delays instead of just absorbing them — a real, funded, $875 million shift from reactive to predictive air traffic management, built by Air Space Intelligence and live today at three Washington-area airports. It’s also, by design, a cautious 90-day experiment with no passenger-facing features and no nationwide reach for at least two more years. After a year of strikes and weather cascades that this site has tracked across three continents, a system built to flag the next one 45 days out is worth watching closely. Just don’t expect it to touch your flight unless you’re flying through the DC area this season.

---

Details current as of September 21, 2026, based on the FAA’s official announcement and reporting from NPR, Quartz, the Washington Times, and Breitbart. Trial scope and timeline can shift — verify current status at faa.gov before assuming wider rollout.