IA · 14 November 2026 · 5 min read

FAA Expands AI to National Airspace: Inside the $875M SMART Rollout

In brief: The US Federal Aviation Administration is expanding SMART, an $875 million cloud-based AI advisory system, to manage airspace congestion across all 22 national air route traffic control centers. While commercial airlines praise documented reductions in flight delays and jet fuel burn, the national air traffic controllers union has pushed back, noting it was entirely excluded from development and testing amid growing concerns over system resilience.

by Team Mocchi's

FAA Expands AI to National Airspace: Inside the $875M SMART Rollout

Predictive Algorithms Take Over the American Skies

The US Federal Aviation Administration (FAA) has officially expanded its flagship artificial intelligence advisory tool, known as SMART, to all 22 air route traffic control centers across the United States. Following an initial operational trial launched in late September around the high-density airspace of the National Capital Region near Washington, DC, the federal agency decided to push the platform across the entire contiguous national airspace system.

SMART, backed by an $875 million federal investment, leverages cloud infrastructure and predictive machine learning models to ingest multifaceted operational inputs: real-time meteorological fronts, flight trajectory data, regional congestion metrics, and air traffic control staffing levels. The system's objective is to proactively resolve congestion points across the busiest flight corridors in the world, generating dynamic rerouting suggestions and arrival sequencing far faster than manual human planning could calculate.

Flight Times Drop, Airlines Cheer

During its initial test phase in the skies above Washington and subsequent expansions into southern hubs like Atlanta, Jacksonville, and Miami, performance data appeared encouraging to both federal officials and airline executives. As reported by Ars Technica, the FAA logged approximately 180 rerouted commercial flights, shaving an average of 10 minutes off flight times and conserving nearly 80 gallons of jet fuel per rerouted aircraft.

Major US carriers, including Delta Air Lines and American Airlines, promptly embraced the initiative as airlines face tightening operational margins and pressure to rein in emissions. Structurally, the FAA maintains that SMART operates solely in an advisory capacity: a dedicated human team of FAA traffic specialists in Washington reviews the system's recommendations before relaying them to field facilities, where local air traffic controllers retain full legal authority to accept or dismiss proposed flight paths.

Labor Tensions and System Fragility

Despite the agency's reassurances, the nationwide rollout has sparked significant labor friction and rekindled questions surrounding automated aviation infrastructure. The National Air Traffic Controllers Association (NATCA), representing over 15,000 controllers across the nation, stated that it was neither consulted nor involved in the design, testing, or implementation of SMART, declaring that it could not endorse the tool's impact on controller workload or situational awareness.

The deployment also arrived during an uncomfortably tense period for US aviation. As Wired highlighted, a recent nationwide ground stop triggered by an internal communications outage reminded passengers and regulators of the brittle systems underpinning modern flight logistics. While the administration affirmed that SMART had no bearing on the communications breakdown, the incident underscored the hazards of deploying sophisticated algorithmic layers over legacy foundations, particularly during ongoing controller staffing shortages.

Mocchi's Take

The FAA's nationwide expansion of SMART serves as a textbook study on deploying mission-critical artificial intelligence in zero-tolerance operational environments. For engineering teams and enterprise leaders building decision-support platforms, technical accuracy alone will not guarantee operational resilience if frontline stakeholders are bypassed during development. Maintaining a human in the loop as a legal safeguard is insufficient unless system designers actively invest in workforce trust, explainability, and rock-solid fallback architecture.

Further reading

All articles on the Mocchi's blog