Modernizing the airspace, from the cockpit

The FlightMind team · · 7 min read

On the morning of January 11, 2023, the FAA ordered a nationwide ground stop because the system that carries Notices to Air Missions had failed. More than 1,300 flights were cancelled and more than 10,000 were delayed. The cause was not exotic: the agency traced it to a damaged database file, and its later review found that contract personnel had unintentionally deleted files while correcting synchronization between the live primary database and its backup.

For most people it was a one-day story about airline delays. For pilots it was a reminder of something we rarely say out loud: a great deal of what we depend on to fly safely sits on infrastructure with a long history, and some of that history goes back decades. The good news is that, for the first time in a long time, the national airspace system is being rebuilt in earnest. We think that is overdue and welcome. We also think the most interesting part of the story has barely been written.

Where things stand

A quick, dated tour, because the details matter and the news moves quickly. Figures are as of early November 2026, with sources linked.

The new air traffic control system

In July 2025 Congress provided $12.5 billion as what the Department of Transportation called a down payment on a new air traffic control system. In December 2025 the DOT and FAA named Peraton as the prime integrator. As of this writing, the stated goal is to finish the work by the end of 2028.

Progress has been real. By the April 2026 update AOPA reported, half of the copper lines had been replaced with fiber and 17 towers had moved from paper strips to electronic ones. A GAO report in September noted that the FAA had replaced about half of the system’s roughly 5,170 copper lines by May, and about 17 percent of its radio sites. It also asked for a reliable lifecycle cost estimate for the program and an integrated master schedule for its first phase, which spans more than 11,000 individual project schedules. The Department of Transportation partially agreed with those recommendations. The first phase will cost up to $16 billion, and in September the Secretary asked for $30 billion more. A bill to provide it, the Modern Skies Act, was introduced in the Senate on September 28, with a House companion on October 1. As of early November it has not passed.

AOPA is a founding member of the Modern Skies Coalition, now some 60 aviation organizations and companies backing the effort. We think the effort is warranted. Anyone who has flown through an area with a patchy frequency, or waited on a clearance that depended on a tired phone line, knows why.

NOTAMs

The January 2023 outage accelerated a replacement already in motion. The new NOTAM Management Service is a cloud-based system meant to replace platforms dating to the 1980s. Its first phase began in the fall of 2025, and on April 18, 2026 the legacy U.S. NOTAM System was retired. The FAA announced that milestone in May, more than a year ahead of its original schedule, with a second phase planned to retire the remaining Federal NOTAM Service infrastructure. AOPA expected the changeover to be invisible to pilots, and for most of us it was. The point of this kind of work is that you do not notice it until the day it would have failed.

Weather

Aviation weather has been moving from text toward graphics for years. The Graphical Forecast for Aviation replaced the text Area Forecast for the contiguous states in 2017, and for the Gulf, the Caribbean and Hawaii in January 2025 (Alaska still has a text one), and the graphical AIRMET is more precise in both time and space than the legacy text AIRMET it replaces. In July 2026 the National Weather Service announced it would retire the BUFR file format for G-AIRMETs in late August, keeping the XML version. That last item is small, but it shows the direction: products are being defined as data first and text second.

People

Equipment is only half of it. The FAA reported hiring2,235 controller trainees in fiscal 2026, the most in its history and above its goal of 2,200. Certification takes time, often more than two years for a new hire depending on the facility, so the benefit of that hiring arrives gradually. Controllers are the reason the system works at all, and a modern system that helps them is good for everyone in it.

Who it is being built for

Read the list above and a pattern appears. Fiber, radios, radar, electronic strips, a single authoritative NOTAM service, a modern scheduling backbone. These are investments in the core of the system, and they are mostly designed around the needs of the people who run it: controllers, airline operations, airports, the FAA itself. Even the language around the new system, digital twins of the airspace and predictive analytics, is the language of managing traffic at scale.

That is the right place to start. You cannot build a good house on a bad foundation. But it leaves a question open, and it is one the pilot flying a single-engine airplane for their own reasons is well placed to ask: what does all this make possible for me?

The gap on our side of the radio

A commercial crew launches with a dispatcher, a meteorologist and an operations center behind them. Most general aviation pilots brief alone, in the time they have, from a handful of sources. The data they need is mostly out there. What is missing is the person, or the thing, that reads all of it against their specific flight and says what matters.

Part of the reason is that so much of the system was never meant to be read by machines. A NOTAM is a terse line of abbreviations written for a human to decode. A forecast is prose. A pilot report is a sentence typed in a hurry. The modernization now underway moves in a different direction. NOTAMs are being moved onto a cloud service designed for faster distribution and easier machine use. Weather is increasingly published as structured, georeferenced products. The more of the system that is exchanged as clean, current, machine-readable data, the more a program can do with it on behalf of one pilot.

Modern infrastructure is not the same as modern situational awareness. The first is built for the system. The second has to be built for the person in the seat.

What we believe

We are an aviation AI company, and we think this is one of the most important opportunities in general aviation. Not because AI is fashionable, but because the shape of the problem fits. A pilot has a particular airplane, particular limits and a particular flight. The relevant data is large, scattered and perishable. The traditional answer was experience, a good briefer and a lot of time. Those still matter, but they do not scale to the volume of information that now exists.

As the national airspace becomes more digital, we believe artificial intelligence can turn that data into situational awareness and decision support for each pilot. Not a replacement for judgment. A virtual right seat that has read everything, knows your limits, and tells you plainly what it found and what it could not.

Three principles guide that belief, and we would hold them whatever the technology looked like.

  • The pilot is always in command. Intelligence informs the decision. It does not make it. Part 91 puts responsibility on the pilot in command, and any tool that blurs that line is working against safety.
  • Facts first, then judgment. What can be worked out precisely should be. Language and reasoning come on top, and stay tied to real reports.
  • Missing data is never clear. If a report is absent, the honest answer is to say so. Silence is not good news.

Caution is part of the job

We are careful about AI in aviation, and the FAA is too. Its first AI roadmap, published in July 2024, takes an incremental, safety-focused approach and distinguishes between systems that are trained and then fixed, which can be tested, and systems that keep learning, which are much harder to assure. That is a sensible way to think. Decision support that sits beside the pilot, shows its sources and defers on every call is a very different thing from automation that flies the airplane, and the bar for each should be different.

The same caution applies to the data. FAA guidance describes weather delivered over ADS-B as advisory, not something that can satisfy a regulatory requirement, and that is the right posture for any tool built on top of it. Better access to information should make pilots better prepared, not more comfortable than the facts justify.

What we hope to see

We do not expect the airspace to be finished in 2028, or any other year. A system this large is never finished. What we hope is that, as the foundations get replaced, the people building tools for general aviation pilots are in the room, too: asking for data that is open, timely and well structured, and building the intelligence layer on the pilot’s side of the radio with the same seriousness that the system side is getting.

If that happens, we think the pilots who fly alone, for themselves, stand to gain the most. The advantage a two-pilot crew has always had, a second set of eyes that reads everything, could belong to everyone.

Brief your next flight with us.

FlightMind is in private beta. Join the waitlist and we will save you a seat.

Join the waitlist