The Federal Aviation Administration has embedded Palantir’s Foundry AI platform into safety operations at major U.S. airports, using it to identify recurring risk patterns from surface radar feeds, NTSB filings, weather data, drone sighting logs, and live aircraft tracking — and the system has already triggered at least one concrete operational change.
A Politico investigation published June 19 revealed the scope of the FAA-Palantir partnership. Among the findings: in April 2026, the agency banned parallel landings at San Francisco International Airport after Foundry flagged a cluster of Traffic Collision Avoidance System alerts linked to landing procedures that had been applied incorrectly. It is the first publicly confirmed instance of AI-generated pattern analysis directly triggering an operational change at a major U.S. hub.
How FAA’s Palantir AI Maps Risk
Foundry sits on top of existing FAA data systems rather than replacing them. The platform pulls from surface radar, ADS-B position data, historical incident databases, and real-time traffic feeds — data the agency already generated — and runs pattern analysis across all of them simultaneously.
Human analysts review flagged locations, validate findings against operational knowledge, and decide whether to adjust procedures, mandate equipment, or modify arrival rates. The AI surfaces the signal; people make the call.
FAA Deputy Administrator Chris Rocheleau described the partnership plainly: “Palantir has been a great partner for us.” The agency has allocated nearly $4 million in fiscal 2027 funds to expand the initiative, an acceleration that began following authorization in a 2025 spending law.
LaGuardia and the Transponder Gap
The Foundry deployment has taken on added urgency in the wake of the March 22, 2026 collision at LaGuardia Airport, where a tragic chain of five simultaneous system failures led to an Air Canada Express CRJ-900 striking a Port Authority fire truck on Runway 4, killing both pilots and injuring 41 people.
NTSB investigators determined that the fire trucks lacked vehicle movement area transponders. Without them, LaGuardia’s ASDE-X surface detection system could not reliably track the vehicles — and generated no collision alert before impact.
The FAA is now using Foundry to identify which airports have similar transponder gaps and prioritize equipment upgrades accordingly. LaGuardia, Chicago O’Hare, and Reagan National have all been flagged as priority targets.
FODNews has previously reported on the scope of the underlying problem — the FAA’s own March 2026 update identified more than 150 U.S. airports with elevated surface collision risk, spanning 453 individual hot spots. Foundry adds a layer that static designations never had: the ability to surface emerging patterns before they become formal entries on an airport diagram.
Pattern Detection, Not Real-Time Intervention
Officials and safety researchers have been consistent about what Foundry can and cannot do. Transportation Secretary Sean Duffy has publicly emphasized that AI supports — rather than replaces — human decision-making in air traffic management.
Hassan Shahidi, president and CEO of the Flight Safety Foundation, framed the distinction clearly: Foundry is an early-warning system, not a real-time intervention tool. Its value is in catching conditions that accumulate over weeks and months at specific airports — not the cascading simultaneous failures that caused the LaGuardia collision.
NTSB Chair Jennifer Homendy, after an FAA demonstration, called the system impressive for identifying live hotspots in national airspace but noted its effectiveness depends entirely on how analysts program its parameters. Former NTSB Chair Robert Sumwalt cautioned against over-reliance while acknowledging the platform’s genuine potential to surface risks that fragmented legacy databases have historically buried.
A Question for the Industry
For airfield safety professionals, the FAA’s Foundry deployment raises a pointed question: if AI can cross-reference thousands of incident reports to identify vehicle proximity risks and procedural failures, the same capability could theoretically map debris-prone surface areas proactively — before a tire strike or engine ingestion event forces reactive action.
No formal announcement has extended Foundry’s analysis to FOD detection or debris pattern mapping. But the infrastructure is now in place, and the precedent — an AI-generated finding triggering a concrete operational change — has been set. The shift from reactive to predictive has begun. The question is how far it goes.
Stay informed — subscribe to FODNews for daily coverage of FOD incidents and prevention worldwide.
Sources
- Politico — “FAA Uses AI to Track Aviation Close Calls” (June 19, 2026)
- Air Traveler Club — “FAA deploys Palantir AI, banning parallel landings at San Francisco International” (June 20, 2026)
- FODNews — “FAA Report: 150+ U.S. Airports Are Runway Incursion Hot Spots as Safety Risks Climb in 2026”
- NTSB — Investigation DCA26MA161 (LaGuardia Airport, March 22, 2026)
- FODNews — “Three Months After LaGuardia: Five Simultaneous Failures Killed Two Pilots — And One Fix Is Already Underway” (June 23, 2026)
