Waymo vs Tesla vs Zoox vs Motional: Four Robotaxi Stacks Compared

Four companies are now running driverless cars on American public roads, and they disagree about almost everything. Not branding. The actual engineering: what the car should see with, what shape the car should be, and who the rider should be booking through. This is the first time a Waymo vs Tesla vs Zoox vs Motional robotaxi comparison has been a comparison of genuinely different machines rather than four versions of the same one.

The sensor question splits the field first

Waymo's position has not moved in a decade: lidar, radar and cameras, fused, with the redundancy treated as the point rather than the cost. Its newest vehicle, the Ojai, is a purpose-built minivan carrying the current generation of that stack, now operating across San Francisco, Los Angeles and Phoenix.

Tesla runs cameras only. No lidar, no radar, on the argument that a system trained on enough video should be able to do what a human driver does with two eyes, and that the extra sensors are an expensive crutch. Zoox and Motional sit closer to Waymo, with multi-sensor suites.

Nobody has been proven right yet, which is worth saying plainly. Camera-only is cheaper per vehicle and scales better if it works. Multi-sensor is more expensive and degrades more gracefully when one input is compromised. Those are different bets about how often the hard cases show up.

Vehicle design is the second fork

Zoox has taken the most committed position of the four. Its vehicle has no steering wheel and no pedals, and it has now won landmark US approval to operate that way, which is the regulatory milestone the whole sector had been watching. There is no fallback to a human driver because there is no driver's seat to fall back to.

Waymo retrofits and now purpose-builds passenger vehicles that still look like cars. Tesla uses vehicles it already manufactures at volume, which is a real structural advantage if the software gets there. Motional runs modified production vehicles too.

Who owns the rider matters more than it looks

Waymo and Tesla run their own apps and own the customer relationship end to end. Motional operates through Uber, which means Uber owns demand and Motional supplies the vehicles. Zoox is building its own service.

That split decides who absorbs the cost of empty miles, who sets pricing, and who gets blamed when a ride goes wrong. It is the same structural question showing up in the drone delivery race, where some operators are building their own aircraft and others are partnering.

Permits are the scoreboard nobody watches

Fleet approvals are a better read on ambition than press releases. Nevada has cleared Tesla for up to 5,000 vehicles, Waymo for 1,000, and Uber for 1,000 through Zoox and Motional, with Zoox holding a separate 100-vehicle permit. Waymo's commercial territory is expanding across the Bay Area and Los Angeles, with Sacramento and San Diego in view.

A permit ceiling is not a deployed fleet. Tesla being cleared for 5,000 vehicles tells you what it asked for, not what is driving. But permit filings are the earliest public signal of where a company thinks it will be in eighteen months, and right now they are the only comparable number across all four.

What to actually watch

Not the announcements. Watch three things: whether Zoox's no-controls approval gets copied by regulators in other states, whether Tesla's camera-only stack clears a supervised-to-unsupervised transition in a city with weather, and whether Waymo's cost per mile falls fast enough to matter. Everything else is territory expansion, which is easy to announce and slow to prove. Autonomy in other domains follows the same pattern, from delivery aircraft to autonomous fighter jet testing: the demonstration comes years before the deployment.