The FAA’s Plan for Air Traffic: What the $875M SMART AI Contract Buys

The Federal Aviation Administration is preparing to put SMART, an AI traffic planning tool, into live use around Washington, D.C. as soon as Monday, September 21, according to reporting from The Wall Street Journal. It is the first operational step in the FAA's plan for air traffic, a 12-year contract that Air Space Intelligence says is worth $875 million. The goal is to spot congestion before flights leave the gate, and the FAA itself has not yet announced the launch date.

The headline version of this story, that AI is coming to fix air traffic, is looser than the paperwork. The FAA's SMART fact sheet describes a cloud based planning layer that works alongside existing systems, and it says controllers remain responsible for keeping aircraft safely separated. None of the FAA material I reviewed describes SMART issuing clearances or operating without a controller in the loop.

I read the FAA's June 22 award announcement, the SMART fact sheet and the contractor's own release, then checked the launch reporting against Airways and other outlets. The money and the contract are real. The results are still promises, and as of September 18 the FAA has not published a single performance metric for SMART.

By Alex Carter | Published September 18, 2026 | The news broke September 17, 2026

What Did the FAA Announce, and When?

The contract is three months old. On June 22, 2026, Transportation Secretary Sean Duffy and FAA Administrator Bryan Bedford announced that the FAA had awarded Air Space Intelligence (ASI) a contract for two linked systems. Flow Management Data and Services (FMDS) becomes the backbone of the FAA's Air Traffic Control System Command Center. SMART is an enhancement inside FMDS that coordinates schedules and trajectories before aircraft depart.

The FAA's release did not state a price. The $875 million, 12-year figure comes from ASI's own announcement that same day, and from the Journal's reporting as relayed by TechCrunch. That makes it a company figure I could not check against a contract document. It also covers both systems, while TechCrunch's brief ties the full amount to SMART. Spread evenly, it works out to roughly $73 million a year, by my arithmetic.

What is new is timing. Airways reported on September 18, summarizing the Journal, that SMART could go live in the Washington area as soon as Monday, September 21. As of that date the FAA had not announced it, and the agency's own materials say only that initial operations begin in the fall of 2026. I could not open the Journal's article, so the launch details here come through Airways and TechCrunch.

How SMART Works Inside the FAA's Plan for Air Traffic

According to the FAA's SMART fact sheet (a one page PDF), the software analyzes airline schedules, weather, airport capacity, airspace conditions and operational constraints, then predicts traffic flows and flags conflicts before they happen. The idea is a shared view for the FAA, airlines and operators, so everyone agrees on routes and departure times before takeoff. Bedford described the aim in the June release as “improving how we manage airspace before flights depart.”

Per POLITICO's September 18 report, as summarized by Airways, SMART will offer advice for the three airports serving the capital. Airways identifies the region's main airports as Reagan National, Dulles and Baltimore/Washington. Recommendations, such as alternative routes around bad weather, flow from a SMART center at the Transportation Department to an established command center through existing FAA processes. From there they reach controllers through systems they already use.

That is advice, not control. It also fits the pace Duffy set on September 15 at a National Artificial Intelligence Association talk, where Politico quoted him saying, “We have to do a slow roll into the airspace.”

Is SMART a Drone Traffic Management System?

No. The fact sheet does not mention drones or unmanned traffic management (UTM). It does list advanced air mobility aircraft among those SMART should make room for, next to new airlines and cargo operators. Nothing I reviewed ties SMART to a drone traffic management system. It is a planning tool for the national airspace and one piece of the FAA's broader air traffic management program.

Where the Claims Outrun the Evidence

According to Airways, none of the fact sheet, the award announcement or Bedford's September 15 testimony reports any measurement of SMART in operational use, and the agency has not published the baseline it will use to judge a roughly 90 day first phase. Bedford told Politico that phase is mainly about testing the system and building airlines' confidence in its predictions.

Scope is the second gap. That was six days before the award, and none of the September coverage I reviewed says whether the limit still holds.

Airlines are the third. Politico reported that carriers spent weeks confused about the FAA's plans, including whether they could opt out, and that three airline officials said some questions still lacked granular answers. Their concerns eased, per the same report, once the FAA said use would be limited at first.

Vendor claims are the fourth. ASI's president of civil aviation, Bernard Asare, says the company spent nearly $100 million of its own money on a platform that is “operational today.” That is a company claim. One trade outlet credits Alaska Airlines' use of ASI's Flyways product with a 3 to 5 percent fuel burn cut on long haul flights, and Startup Fortune reports a conflict spotting window stretching from about 15 minutes to two hours. Neither figure appears in the FAA documents I read, and I could not verify either.

SMART is also not a staffing fix. The FAA is running a separate hiring plan that, as TechCrunch notes, the agency said would “erase the longstanding staffing shortage.”

My read: this follows a pattern I have watched for years. A small advisory first phase gets described in national terms, and the success metric gets picked after the fact. The useful question is which numbers the FAA commits to at day 90, and against what baseline. The same problem of proving a system does what its builders say runs through BriefFlash's coverage of independent AI safety evaluators, and MIT's HardFlow method tackles a different version of it, keeping generative AI inside hard safety constraints.

Who This Affects

• Controllers and traffic managers. SMART is advisory and sits inside existing systems. An industry official told Politico it will not create new procedures for controllers. The FAA also says the goal includes easing controller workload, which is a claim, not a finding.
• Airlines. They get a shared schedule and routing view, plus open questions on opt outs. Airlines for America CEO Chris Sununu called the program an encouraging step in the FAA's June release and said carriers are coordinating closely with the agency.
• Travelers in the Washington corridor. Reagan National, Dulles and BWI are the likely first test sites. All three were among the 40 airports under the FAA's emergency flight reduction order in November 2025. That is context on how congested the region is, not a sign SMART will change anything there on day one.
• Vendors and builders. A 12-year award for software that is already running at airlines suggests the FAA will buy commercial AI rather than build it. Startup Fortune reports ASI beat Palantir and Thales for the job, which I could not confirm with the FAA or a second outlet.

What to Watch Next

First, whether the FAA confirms the Monday start or moves it. Second, the 90 day mark: if the launch holds on September 21, that lands around December 20, by my count.

Airways lists the numbers that would settle the question: delay minutes, the duration of ground delay programs and ground stops, cancellations, reroutes issued, fuel burned in holding, and how often airlines accept the recommendations. If the FAA publishes even a few of those against a baseline, the claim gets real. If it publishes only anecdotes, treat the results with caution.

Third, ASI says SMART and FMDS would be deployed over the next 12 to 24 months. Answers on the altitude limit and airline opt outs will show how fast that can realistically move.

Updated: none yet. If the FAA confirms or moves the September 21 launch, this note will read “Updated [date]: [what changed and why].”

Key Takeaways

• The FAA's plan for air traffic centers on SMART, an AI planning tool awarded to Air Space Intelligence on June 22, 2026 alongside a companion system, FMDS. The $875 million, 12-year figure comes from the contractor and the Journal, not the FAA's release.
• SMART is advisory and works before departure. The FAA's fact sheet says controllers remain responsible for keeping aircraft separated.
• First live use is expected around Washington, D.C. as soon as September 21, per the Journal via Airways. The FAA has not announced the date, and the first phase is a roughly 90 day test.
• Fewer delays, less fuel burn and better on-time performance are FAA goals, not measured results. No performance metrics or baseline have been published.

FAQ

Do air traffic controllers make 300k?

Not on the figures the government publishes. The Bureau of Labor Statistics put the median annual wage for air traffic controllers at $148,080 in May 2025. The FAA's hiring page, last updated April 21, 2026, says certified controllers average more than $155,000 a year within three years of graduating from the academy. Trade outlet Air Gazette reports that federal law caps combined base and locality pay at $228,000 for 2026. I could not find a published source for $300,000 as a typical or median figure, and whether individual controllers reach it with overtime is unverified.

Which 40 airports is FAA cutting flights?

The premise is out of date. The FAA ordered flight reductions at 40 high traffic airports during the 2025 government shutdown, starting November 7, 2025 at 4 percent with a planned ramp to 10 percent. The FAA ended that emergency order on November 17, 2025 at 6 a.m., and I found no current nationwide order of that kind. The list included Atlanta, Boston, Charlotte, Dallas/Fort Worth, Denver, Newark, JFK, Houston Intercontinental and the three Washington area airports, Reagan National, Dulles and BWI. The FAA's November 14, 2025 release has the full list, and the FAA's NAS status page shows live delay programs.

What does Charlie Charlie mean in aviation?

It is informal radio shorthand for yes, affirmative or correct, built from the phonetic name for the letter C. It is not standard phraseology. Pilot and controller threads on PPRuNe trace it to a marine signal code where C means yes, while a 2003 Airliners.net reply reads it as “copied correct.” Those are forum explanations, not official definitions, and one Australian controller there says he has never heard it used for yes and sticks with “affirmative.” It has no connection to SMART.

What is the 1/2/3 rule in aviation?

It is a pilot flight planning rule for instrument flights under 14 CFR 91.169. A pilot does not need to file an alternate airport if the destination has an instrument approach and, from one hour before to one hour after the estimated arrival time, the forecast ceiling is at least 2,000 feet above the airport and visibility is at least 3 statute miles. Fail either test and an alternate is required. It is separate from SMART, though both depend on weather forecasts.