By Phil Royle – Aug. 21, 2026
If you want a glimpse of where Tesla’s robotaxi ambitions stand, Las Vegas is a pretty good place to look. Recently, Tesla reportedly wanted to operate as many as 5,000 autonomous robotaxis in Las Vegas. Nevada’s answer was an unequivocal pull to the reins, handing them a permit for 10. That’s quite a difference, and a blatant tell as to what the city thinks about Tesla’s self-driving tech.
The decision is particularly interesting because Tesla is hardly a stranger to Sin City. The company has invested heavily in Vegas via its underground Vegas Loop, yet when Tesla came asking for a fleet of thousands of robotaxis, regulators approved a number that’s just 0.2% of what the company requested.
Shifting gears, Zoox (which is owned by megacorporation Amazon), is already operating autonomous vehicles in Las Vegas. The vehicles have no steering wheels or pedals, and Zoox has apparently been operating about 65 vehicles under a permit allowing as many as 100. The company has also received approval to transition from offering free rides to charging passengers for the thrill of living on the autonomous edge. /p>
At least on paper, Nevada is allowing Zoox to operate 10 times as many autonomous vehicles as Tesla. That should get Tesla’s attention.
The most interesting part isn't simply the numbers, it’s the underlying technology that drives the cars, and what this decision says about each company’s approach to self-driving tech.
Tesla has famously pursued a vision-based approach, relying primarily on cameras and image processing. Zoox, meanwhile, uses LiDAR as part of its camera and sensor suite, and that distinction matters because the fundamental question in autonomous driving is whether cameras alone can provide enough information to make a vehicle truly autonomous in every possible situation.
There’s certainly disagreement about this. Some people in the autonomous-driving industry believe LiDAR is necessary for unadulterated autonomy. Others argue that vision systems supplemented by radar can work. Tesla is sticking to a vision-only equation for its self-driving tech, while Zoox throws the kitchen sink at it, blending cameras, radar, and LiDAR.
Nevada’s decision provides an interesting real-world data point here, because Tesla isn't being given thousands of vehicles to prove its theory. It’s being given 10.
There are restrictions attached to those robotaxi Teslas, too. Tesla's autonomous cars are limited to 45 mph and cannot pick up passengers at the airport. That's a significant limitation for a Las Vegas robotaxi service, where the airport would seem like one of the most obvious destinations.
Taken together, the situation makes me wonder whether Nevada is effectively saying: We're interested, but we're far from convinced.
That seems especially notable given how much Tesla has invested in Las Vegas. This isn't a jurisdiction hostile to the company. Yet even here, regulators appear to be taking a cautious approach.
And maybe that's the right approach. Level 5 autonomy is an extraordinarily high bar. The goal is not merely to make a car drive itself most of the time. Level 5 means operating without human intervention in all conditions. A camera can be obstructed, and at highway speeds, the margin for error becomes very small when there’s no additional data to rely upon.
Zoox isn't necessarily proof positive that its sensor-heavy approach offers the magic that’s needed, but its larger approved fleet certainly make the comparison uncomfortable for Tesla.
The Vegas experiment may ultimately prove Tesla right. Maybe cameras really can do it (at least in the desert). But for now, Vegas appears to be saying: prove it first.
(Image courtesy)
