The Robotaxi That Threw Out the Steering Wheel

Zoox built a car with no steering wheel and no pedals. The NHTSA decides this month whether it can charge anyone to ride in it.

That ruling — on whether Zoox can operate up to 2,500 vehicles commercially — is the moment the 500,000-person Explorer waitlist either becomes a business or stays a very expensive science project. It's also the sharp end of a philosophical split the autonomous vehicle industry has been working through for years: retrofit a vehicle designed for a person, or build something purpose-made for a passenger who will never touch a control.

Zoox went purpose-built. A new Ars Technica deep-dive published Tuesday traces the engineering consequences: symmetrical front and rear ends so direction doesn't matter, sensors perched on little ledges at all four corners for unobstructed situational awareness, and a redundant electrical architecture that duplicates the battery pack, drive motor, and HVAC system at both ends. If one end fails, the other finishes the fare.

"A robotaxi is not a car; it's not a human-driven vehicle, and the requirements are wildly different," said Chris Stoffel, director of robot industrial design at Zoox. "We kind of got the kitchen sink of redundancy as we wanted to make sure this first product really could complete the mission."

The approach is expensive. Zoox spent roughly six years building what competitors shipped faster by bolting hardware onto existing designs. Waymo operates more than 800 vehicles in the Bay Area alone — retrofitted Lexus and Jaguar models with conventional controls — and charges fares today. Tesla's Robotaxi service operates in parts of Texas without a safety driver, having expanded from Austin to Dallas and Houston in April; the Bay Area remains unconfirmed. Tesla's cameras-only engineering bet — no lidar, no radar — is the industry's most contested approach.

The incident record is where the comparison gets awkward for both of them. In the second half of 2025, Zoox submitted 22 incident reports to the National Highway Traffic Safety Administration; Waymo submitted 449. Total for Zoox across all modes: 123 NHTSA-reported accidents as of mid-March 2026, across roughly 50 vehicles. The ratios are not directly comparable — Waymo's fleet has logged orders of magnitude more miles — but per-vehicle, the numbers favor Zoox.

Jesse Levinson, Zoox's co-founder and CTO, has spent years arguing the purpose-built case. "The premise that led us to start the company was that when you have autonomous technology, instead of retrofitting cars designed for human drivers, we should create a whole new type of vehicle designed to be a robot taxi," he told Car and Driver. "So it doesn't have to have any of the legacy constraints of cars."

The design has one structural advantage that is hard to replicate by retrofitting: sensor placement. Because there is no traditional hood eating up frontal real estate, Zoox's sensor ledges sit higher and further forward than anything bolted to a production car's roof. "Because we don't have a traditional hood, we've optimized our frontal coverage in a way that would be nearly impossible on a retrofitted vehicle," Stoffel told Ars Technica.

That architectural freedom is also why Zoox holds the only NHTSA exemption in the US allowing vehicles without steering wheels or pedals to operate on public roads. That exemption currently covers demonstration rides — free ones, through the Explorer program, in Las Vegas, San Francisco, and Austin, with Miami next. The company says it has logged nearly 2 million autonomous miles and carried more than 350,000 riders. All rides have been free.

Commercial operation — meaning Zoox can actually charge — requires a separate authorization. The NHTSA comment period on Zoox's application closed April 15. A ruling is expected this month.

Waymo, meanwhile, is expanding to ten new US cities this year, targeting more than 20 total, and has begun international pilots in London and Tokyo. The company reported $350 million in annual recurring revenue as of earlier this year and was valued at $126 billion in a recent funding round. It charges fares today and operates vehicles with conventional controls. It is, by any conventional metric, the commercial leader.

Zoox's bet is that the math changes when you design for the rider instead of the driver. Whether that produces a durable advantage or simply a more expensive vehicle with better sightlines depends on what the NHTSA decides in the next few weeks — and whether the per-mile incident ratio holds as the fleet scales.

The 500,000 people on that waitlist are about to find out if the ride they signed up for actually exists.