What a mile costs, and what it takes to make it free.

What a mile costs today

In a big city, a rideshare costs a rider several dollars a mile, and a car you own costs well under one.

Most of a fare pays for the driver. In New York, drivers’ pay came to about 79% of the base fare in October 2025, and UC Davis and the Institute for Transportation and Development Policy put drivers at “50% or more of ride-hailing costs” (Three Revolutions in Urban Transportation, 2017). The driver’s own car costs about 87 cents a mile to run, by the city’s latest study of driver expenses.

Transit already runs on support. The average U.S. bus trip costs $8.24 to run, and the fare covers about 90 cents of it. The rest is paid by someone other than the rider (Vugo arithmetic, National Transit Database, 2024).

How low it can go

Take out the driver and run the vehicle on electricity, and the research says a mile costs a fraction of today’s fare. The projections range widely, because their assumptions do.

Bösch’s peer-reviewed model has an urban taxi’s cost per passenger falling from 2.73 to 0.41 Swiss francs a kilometer once it drives itself: about $4.46 down to 67 cents a passenger-mile (Vugo conversion, 2017 prices). Prices from 2017 run about 37% higher in 2026 dollars.

What the projections assume:

• Bösch: insurance half as costly and the vehicle 20% dearer; cleaning counted, at almost a third of a driverless taxi’s cost; no separate line for remote supervision.
• RethinkX: full approval for driverless vehicles, cars that last 500,000 miles and insurance 90% cheaper; no remote operations, cleaning or depots.
• ARK: counts remote operators. ARK’s funds also hold Tesla.
• BCG: an operator needs a fleet of 15,000 to 20,000 vehicles to break even on operations.

Across a whole city, UC Davis and ITDP found that vehicles that are driverless, electric and shared together could cut the cost of vehicles, roads and transport services “by over 40%” by 2050, against business as usual, with savings “approaching $5 trillion per year” worldwide.

Sharing the vehicle changes the math

What matters to a rider is the cost of a passenger-mile: the cost of a vehicle-mile divided by everyone on board, averaged over every mile the vehicle drives, empty ones included.

For scale, cars carry about 1.52 people a mile (NHTS 2022), vanpools about 4.5 and city buses about 6.5 (National Transit Database, 2024). An MIT simulation found that 2,000 ten-passenger vehicles could meet 98% of the trip demand served by New York’s 14,000 taxis (Alonso-Mora and colleagues, 2017). UC Davis and ITDP expect small buses of 8 to 16 seats, once automated, to become “the cheapest per-passenger-kilometer on-demand travel option in the world.”

What makes a ride free

A ride is free when what it earns covers what it costs: revenue per passenger-mile at least the cost of a vehicle-mile divided by the riders on board.

Where the revenue comes from:

• Media in the vehicle. People who aren’t driving watch, and what they watch carries advertising. Vugo bills in-vehicle screens per validated play, not per person: screens don’t sense who’s in the car. So we publish no revenue-per-screen figure here; it depends on how much of each hour is sold.
• The passenger experience. Intel and Strategy Analytics, in the 2017 study that named the “passenger economy”, put new consumer services in driverless vehicles at “US$200 billion in revenues” by 2050: dining, entertainment, tourism, healthcare and services that come to the passenger. They expect the vehicle “to finally become a media and entertainment hub.” Intel gives no separate figure for advertising.
• Commerce, and the ads that come with it. Amazon is an e-commerce platform, not an ad company, yet the ads its sellers buy have made it one of the largest advertising businesses in the world. Services sold to passengers will carry the same kind of advertising.
• Others have made the case. McKinsey’s Monetizing car data (2016) described “free mobility”: rides offered in exchange for watching targeted ads or making purchases, or free rides to a store paid for by the store. RethinkX (2017) expected advertising, data, entertainment and product sales to “open a road to free transport” in shared fleets.
• Public money. Transit already pays about 89% of a bus trip’s cost from sources other than the fare (Vugo arithmetic, National Transit Database, 2024).

Free rides already happen

Vugo offers sponsored rides: a brand pays for a ride as part of its campaign, and the rider sees who paid. Rides like that have happened in limited numbers, and still do. Free rides for everyone haven’t arrived. Vugo co-founder James Bellefeuille set out the goal in The Case for Free Transportation in 2018, and Vugo CEO Robert Flessner made the case on CNBC that March. Fast Company had asked the question a year earlier: Will ride-hail be free by 2021?

What it means for households

Transportation is the second-largest bill an American household pays, behind only housing. In 2025 the average household had an income of $110,847 before taxes and spent $81,905: $27,790 on housing, then $13,492 on transportation, 16.5% of all spending. That’s more than food ($10,633) and about twice healthcare ($6,470) (Bureau of Labor Statistics, Consumer Expenditures 2025). Across 138 million households, it comes to about $1.86 trillion a year (Vugo arithmetic).

This doesn’t push drivers out. In our view, building fleets of driverless vehicles will take years, and people will keep driving and earning from it. Drivers who own their cars could also put them to work on the same networks.

The counterpoints

• Advertising alone may not be enough. A 2019 Fraunhofer study of driverless taxis in New York found ads came to about 2% to 5.6% of the fare, “not sufficient to guarantee a free-to-use taxicab fleet.” It modelled single-party robotaxis, not shared vehicles; it measured ads against a taxi fare, not against the cost of a driverless shared vehicle; and its summary doesn’t say what ad prices it assumed. If they were the taxi screens of the time, a fixed loop with no targeting that we estimate sold for $5 to $10 per thousand impressions, they were low: streaming-service ads average about $26 to $27 (EMARKETER, December 2025), and Vugo’s targeted screens typically run $20 to $25. Read it as a lower bound for ads alone in single-rider robotaxis.
• Robotaxis have to stay busy. Nunes and Hernandez (Transportation Research Part A, 2020) found a single-rider autonomous taxi would need its utilization to rise from 52% to 100%, and its profit to fall, to match the cost of owning a car; with shared rides, occupancy would need to rise from 1.67 to 2.2.
• Timelines have slipped. RethinkX expected its 16-cent mile by 2021, and a 2016 McKinsey scenario had up to 15% of new cars fully driverless by 2030. Every projection here is a projection, with its year.
• Company statements aren’t evidence. Tesla’s 20 cents a mile “over time” is a claim. Waymo and Zoox don’t publish a cost per mile.