Technology
FAA invests $875 million in AI system to tackle air traffic controller shortage
The agency is deploying an automated platform designed to optimize flight routing and reduce conflicts in U.S. airspace.
TechCrunch — AI
The Federal Aviation Administration faces a persistent shortfall of air traffic controllers and is turning to artificial intelligence as a potential remedy. An $875 million software initiative will deploy an AI-powered system to help controllers manage operations more efficiently and reduce safety risks across the country's busy skies.
What is SMART?#
The FAA's new platform, called SMART (Strategic Management of Airspace, Routes and Trajectories), is a cloud-based system built by Air Space Intelligence that integrates with existing FAA infrastructure. Rather than replacing human controllers, the software is meant to augment their decision-making by analyzing multiple data streams simultaneously.
The system processes airline schedules, real-time weather patterns, airport capacity levels, airspace conditions, and operational constraints. Using these inputs, SMART predicts how air traffic will flow and flags potential conflicts before they develop. This predictive capability could help controllers make faster, more informed routing decisions and prevent bottlenecks.
Rollout and timeline#
The FAA will spend $875 million over 12 years to deploy and maintain SMART. The system will first launch in the Washington, D.C., metropolitan area as a pilot, then expand to other regions based on performance and lessons learned from the initial deployment.
Broader FAA efforts#
This AI initiative complements other ongoing efforts to address controller shortages. Earlier in 2026, the FAA announced an aggressive hiring campaign aimed at eliminating the persistent staffing gap that has constrained operations for years. The agency is also advancing a larger modernization project to replace and upgrade the nation's aging air traffic management infrastructure, which relies on systems that are decades old in some cases.

