In May 2025 the Transportation Secretary announced a new air traffic control system, to be built in about three years. Congress put $12.5 billion behind it that July. On September 14 GAO published its first full look at how the job is being run.
The one-sentence version, from GAO: "FAA has made progress implementing this initiative, but it could be prioritizing speed over costs and air traffic disruptions."
Why it matters
Controllers manage up to 45,000 flights a day on equipment GAO has been warning about for years. In 2024 it found 51 of the FAA's 138 air traffic systems unsustainable and 54 more potentially unsustainable. The January 2023 failure of the NOTAM system cancelled more than 1,300 flights and delayed almost 10,000. At Newark in April 2025, controllers lost radar and radio contact "for about 90 seconds."
Nobody disputes that the systems need replacing. The report is about whether the replacement is being planned well enough to finish on time and on budget.
The plan
The Brand New Air Traffic Control System comes in two phases.
- Phase 1, due by December 2028: 13 programs covering telecommunications, voice switches, radios, radars, surface-movement tracking, controller displays, weather sensors and more. Its projected cost is now $10.6 billion, $1.6 billion above the funds set aside for it.
- Phase 2: replacing the core automation that controllers work on, at about $10.2 billion more. The FAA "does not yet have an estimated date for when phase 2 will be started or completed."
The timelines have been pulled in sharply. Swapping copper lines for fibre was due in September 2030 and is now due December 2028. Replacing 1,581 radios was due in 2037.
What GAO found
Progress. "As of May 2026, FAA reported replacing 2,560 of 5,170 old and frail copper wires with high-speed fiber optic communications cables," about half, and 17 percent of radio sites. By March 2026 it had committed $1.8 billion, 20 percent of the money.
No full cost estimate. "FAA has not developed a comprehensive and well-documented lifecycle cost estimate." Its $10.6 billion figure leaves out government costs and most years of operations and maintenance, so it "underestimates the full costs." Asked for the analysis behind its numbers, officials said it "was not readily available and would require a tremendous amount of effort to locate and provide."
A reason GAO points to: in August 2025 the FAA Administrator issued a blanket waiver exempting every program in the effort "from nearly all of FAA's Acquisition Management System requirements."
No single schedule. "Instead, FAA has 11,389 individual project schedules that are not integrated." That matters on the ground. In GAO's example, one St. Louis site gets its voice-switch site work in April 2027 and its display site work about eight months later. Without one schedule, conflicts between installations at the same facility are hard to see. The FAA plans an AI tool to find them; it is "undergoing testing," with no time frame.
GAO's conclusion: "In essence, the FAA is buying delivery speed at the risk of cost uncertainty and potential disruptions." And the risk it names is familiar: "these risks increase the likelihood of FAA repeating its history of cost overruns, delays, and unmet performance targets." The FAA's last modernization program, NextGen, is expected to cost at least $35 billion through 2030.
What happens next
GAO made two recommendations to the FAA: build a full lifecycle cost estimate, and build one integrated master schedule for phase 1. The Department of Transportation "partially concurred with both recommendations" and did not say what it will do. GAO kept both as written.
What the report does not show
- It does not find that the 2028 deadline or the $10.6 billion figure will be missed. It finds the FAA cannot yet show they won't be.
- It ties no specific outage or near-miss to this program. The incidents above are the reason for the program, not its results.
- The progress figures are the FAA's own.