The Dependency Nobody Talks About Enough

The emergence of autonomous vehicle technology is reshaping how travelers reach airports, creating potential disruption to airport revenue models — and its key concern is narrower still: the existing revenue model of airports faces a potential disruption that could impact the self-sufficiency of smaller airports.

Airports Council International’s 2025 Airport Economics Survey puts car parking at 43.2 percent of non-aeronautical revenue across North American airports — more than double the concentration seen anywhere else in the world. In dollar terms, FAA Compliance Activity Tracking System data shows US airports collected $6.0 billion of parking and ground transportation revenue against $13.2 billion in total non-aeronautical revenue in 2023. That revenue collapsed to $2.69 billion during the pandemic trough of 2021, a 44 percent decline from the $4.82 billion baseline of 2019 — and has since not just recovered but grown past it, with per-enplanement non-aeronautical yield climbing from $5.10 in 2019 to $6.60 in 2024.

At the airport level the parking share concentration gets more extreme. Charlotte Douglas booked $211.1 million of its $431.7 million in 2024 operating revenue — 48.9 percent — from parking, concessions, and rental cars combined. Dallas Fort Worth’s parking and ground transportation revenue hit $251.1 million in 2024, up 12 percent year-over-year. At large hubs generally, parking plus ground transportation represents 43.1 percent of non-aeronautical revenue, but the range runs from 20.4 percent to as high as 68.2 percent depending on the airport’s specific rate-setting structure — a spread that matters enormously for how exposed any individual airport actually is, a point we return to below.

What the Ride-Hailing Precedent Actually Showed

When Uber and Lyft scaled into airport access between roughly 2015 and 2020, 45 percent of large-hub airports reported parking revenue declines, and parking-plus-ground-transportation revenue per passenger fell an estimated 4 to 8 percent industry-wide. Los Angeles International’s parking revenue dropped 5.2 percent year-over-year from 2016 to 2017 — from $108.5 million to $102.8 million — even as enplanements grew, while Transportation Network Company (TNC)-generated revenue at the same airport nearly quadrupled, from $8.9 million to $33.7 million, over the same period. NREL’s research on the pattern found something specifically useful for forecasting the AV transition: parking revenue at newly-TNC-penetrated airports tended to peak 12 to 24 months after ride-hailing entry, then decline 3 to 7 percent annually thereafter — a slow bleed, not a cliff.

Crucially, airports did not simply absorb this as lost revenue. They built a fee structure around it. TNC per-trip access fees, surveyed across 54 large and medium hub US airports, now range from $1.00 (El Paso) to $7.00 (Orlando), with 43 of 54 airports charging pickup fees and 31 charging both pickup and dropoff. Fee revenue from this mechanism grew from roughly 1 percent of parking-and-ground-transportation revenue industry-wide in 2016 to 11 percent by 2024. At DFW specifically, TNC fees reached $39.0 million in 2024 — 19 percent of the airport’s traditional parking revenue — with Uber’s Rasier LLC alone contributing $19.1 million. This is the part of the ride-hailing story that matters most for the AV question: the industry did not out-innovate the disruption, it out-taxed it, converting a threat to landside revenue into a new, growing revenue line with a lower operating-cost footprint than a parking garage.

Why Autonomous Vehicles Are a Different Kind of Threat

Ride-hailing added a new mode alongside parking; a mature robotaxi network potentially substitutes for the trip-cost logic that makes parking attractive in the first place. Economics are the mechanism to watch. Waymo’s current per-mile pricing sits at $0.85–$1.20; Tesla’s stated cost target for its robotaxi operation is $0.30–$0.55 per mile. At that cost level, a round-trip airport robotaxi fare could run $15–$25 — undercutting the $30–$150 yield airports currently earn on a multi-day parking stay. That is the number that would, if realised at scale, force a genuine repricing of the parking asset class rather than a marginal share shift.

The operative word is “if realised at scale,” because the deployment evidence as of this spring tells a much more contained story. Waymo operates driverless service at exactly four US commercial airports — Phoenix (since November 2022), San José Mineta (November 2025), San Francisco (curbside-only, January 2026), and San Antonio (March 2026) — running roughly 500,000 weekly paid trips from a fleet of about 3,000 vehicles across 11 metros nationally. Tesla’s robotaxi, by contrast, has zero driverless airport access anywhere as of this writing; its ~573-vehicle fleet operates only in Austin, Dallas, and Houston, and its California robotaxi permit remains pending with the CPUC.

Within airports’ direct control is the fact that airport-authority access approval happens independently, airport by airport, and does not automatically flow from a state operating permit. This precondition is the leverage point. Roughly 94 percent of large and medium hub airports currently retain the ability to deny AV-operator access outright — the same authority that let them negotiate the TNC fee structure from a position of strength rather than simply absorbing the volume. Between HD-mapping cadence (6–18 months per new metro for Waymo), grid queue depth, and airport-by-airport access gating, the realistic disruption window for AV-driven parking substitution sits in the 2030–2040 range, not next year — but it is a window, not a wall, and it narrows the longer airports wait to build a response into it.

The Finance Mechanics That Decide Who Is Exposed

Not every airport is exposed to this risk equally, and what separates them isn’t where they are or how big they are — it’s how their airline agreements and bond documents are written.

Some airports use a residual model, where the airlines agree to cover whatever the airport can’t pay for out of its other income. If parking and concession revenue comes in below plan, the gap is simply added to the airlines’ bill the following year. Those airports are effectively protected, because the airlines are the last line of defence. The risk doesn’t disappear, though — it just surfaces somewhere else. It shows up when the airline agreement comes up for renewal and carriers push back on the higher cost per passenger they’re being asked to absorb. That’s a genuine problem, but a slower one, and one you can negotiate your way through. Very few airports have adopted the residual setting methodology. Larger proportions are on a hybrid-compensatory basis.

Compensatory and hybrid-compensatory airports take this hit themselves. In a compensatory structure, parking, rental cars, concessions, and ground transportation are the airport’s own business risk — if that income falls, nothing automatically restores debt-service coverage, unless the airport specifically negotiated extraordinary coverage protection into its bond documents.

The pandemic gives us a test case, though a reassuring one only up to a point. Rental car (ConRAC) bonds absorbed a 75–90 percent collapse in rental days for 18–24 months without a single default, because their reserves are built in four tiers and can fund 36 to 60 months of drawdown. But that was a short, sharp shock. A scaled autonomous-vehicle scenario is a different shape of stress: a 40–60 percent decline that persists for years. Reserves sized for a two-year gap would run dry well before an airport could raise its customer facility charge enough to close it.

And the room to raise that charge is limited. Large-hub CFCs currently average about $6.90 per transaction-day, against a practical ceiling practitioners put near $25. Push much past that and the increase itself starts pushing customers away from rental cars — accelerating the decline it was meant to offset.

Avinia’s View

We don’t think the threat autonomous vehicles pose to parking revenue is overstated. We think the timing is.

The 2030–2040 disruption window described above is real. It rests on grid capacity and the sequence in which regulation gets written — constraints no single operator can engineer around, whatever the claims about cheaper vision-only systems.

TRB’s decision to convene an insight event on this topic, rather than waiting for the usual multi-year research cycle, is itself the signal worth noting: the industry’s own technical body sees genuine uncertainty here, not settled disruption and not a false alarm. The ride-hailing precedent shows airports can convert a landside disruption into a new fee-based revenue stream when they act with the access leverage they hold — TNC fees now generate 11 percent of parking-and-ground-transportation revenue industry-wide, up from 1 percent eight years earlier. The AV transition offers a longer runway to do the same thing again, provided airports treat the next five years as the building phase rather than the waiting room.

Source: TRB ACRP Insight Event 11-08; ACI World Airport Economics Survey 2025; FAA CATS FY2023; Dallas Fort Worth International Airport FY2024 financials; Los Angeles World Airports FY2016–17 financials; Charlotte Douglas International Airport FY2024 financials; NREL ride-hailing/parking demand analysis; Walker Consultants; DWU Consulting TNC fee and autonomous-vehicle finance research; Lawrence Berkeley National Laboratory grid interconnect queue data (2025).