Which robotaxi firm has the best business model?

Robotaxis keep popping up across the US.

You can hail one in 11 cities. With three more in testing phase.

And with riders spening more than 4 million hours inside robotaxis already, it's only a question before they spread everywhere.

For the longest time, investors have been asking whether or not robotaxis can work.

Now, that we've proven they do, the bigger question is which robotaxi firm has the best business model? Who owns the best machine for turning miles into $?

Waymo? Tesla? Zoox?

As I'll show you, it's neither of those.

Waymo has riders and scale.

Waymo says it now provides about 400,000 paid rides per week across six major U.S. metros and has surpassed 20 million lifetime rides. It delivered 15 million rides in 2025 alone.

Those aren’t vanity metrics. They show two things investors care about. First, the service works repeatedly in messy real-world cities. Second, customers trust it enough to keep using it. If you assume even a modest $15 average fare, 400,000 weekly rides implies roughly $6 million in weekly gross bookings. Gross bookings means the total value of rides before any fees or costs. The exact fare mix varies by market, but the takeaway is the same. This is now a volume business, not a science project.

But Waymo’s business is still heavy. It doesn’t just build software. It also runs a fleet and expands city by city. That means costs are tied to how much the cars are used and how fast they lose value, not just software margins.

In robotaxis, time is money in a literal way. If a vehicle does about 60,000 paid miles a year, or about 165 per day, then every extra 1,000 paid miles adds real revenue. Every hour it sits idle is lost revenue. Waymo can be the best operator and still find that the “hard” costs—vehicles, maintenance, charging, remote support, and incident response—take a long time to come down.

That’s why Waymo can have the best operating business today, but not yet the cleanest economic model.

Tesla has the biggest upside.

Tesla isn’t playing the same game. It’s trying to reset the cost curve entirely. If it works, everything changes fast.

Tesla’s robotaxi model is powerful on paper. It uses cameras instead of the heavier sensor setup Waymo uses, trains on massive real-world driving data, and leans on Tesla’s manufacturing scale. The goal is a large autonomous fleet with lower hardware costs per vehicle.

Tesla owners have logged billions of miles with FSD (Supervised), which is a driver-assist system that still requires attention. That’s far more data than Waymo’s fully autonomous miles, and it could become a major advantage if Tesla’s AI system reaches full reliability.

But that’s still an if. And the economic risk is whether it works cheaply and safely enough for regulators and cities to allow rapid scaling. Tesla itself says it began removing the safety monitor from Austin rides in January on a limited basis. That shows progress, but also that real-world rollout happens step by step.

So Tesla does not have the best business model today. But the most explosive model if the technical bet works. If Tesla can get per-mile costs meaningfully below the market price, it can attack the entire stack. Until then, it’s still climbing the same hill Waymo already climbed. Consistency, edge cases, and the unglamorous cost of running a fleet in public.

Zoox is taking a different approach.

Zoox, subsidiary of Amazon, is a new entrant on the market.

It’s operating in Las Vegas and San Francisco, with Austin and Miami planned next. That’s real progress. But unlike Waymo or Tesla, te footprint is still limited, and the economics are still being tested. Amazon’s backing gives it time and capital, but not a shortcut to a working model.

What makes Zoox different is its approach. It’s not retrofitting existing cars. It’s building a purpose-built, fully autonomous vehicle from the ground up. No steering wheel, no pedals. That gives it long-term design advantages. But it also slows things down. New hardware takes longer to validate, deploy, and scale.

Right now, it’s still in the early playbook. Limited or free rides. Controlled zones. Tight operational boundaries. The goal isn’t profit yet. It’s data, safety validation, and public trust. It may become a worthy contender one day, but it's nowhere near that point today.

Which robotaxi firm has the best business model? Uber.

Uber’s model is simple. Let others fund the expensive autonomy stack. Then have them use the app consumers already use. That’s why it keeps stacking partnerships. The Rivian deal is a good example. Uber said it will invest up to $1.25 billion and begin with 10,000 Rivian R2 robotaxis in 2028, with an option for 40,000 more later. Uber also hosts Waymo on its platform.

Uber provides riders, payments, dispatch, and local operating infrastructure. That means it can earn from multiple winners without carrying the full R&D burden of building the self-driving system.

The math is straightforward. Assume a mature robotaxi does 60,000 paid miles a year at $2.00 per mile in gross bookings. That’s $120,000 of bookings per vehicle. If Uber’s net take rate, which is its cut after paying fleet partners, settles at 15%, that’s $18,000 per vehicle per year flowing to the toll booth without owning the car. On 10,000 vehicles, that’s about $180 million a year. On 50,000, it’s about $900 million.

And as fleets grow, fleet owners compete to keep their cars busy. Uber can monetize that competition through things like preferred dispatch or guaranteed supply tiers, without taking on vehicle depreciation.

That model has one obvious weakness, though. If Waymo, Tesla, or Zoox can drive enough demand through their own apps, Uber’s take rate gets squeezed. But network businesses are hard to dislodge, and Uber already has a massive base of riders, pricing data, and city-level logistics.

Also, autonomy doesn’t remove the marketplace problem. It makes it more important. When supply scales, distribution becomes more valuable.

If I had to buy one robotaxi stock today, Uber would be it. It sits above everything, doesn’t need to spend as much capital, and can benefit no matter who wins underneath.

  Uber (UBER) has a Disruption Score of 2. Click here to learn how we calculate the Disruption Score.