By Phil Royle – Sept. 7, 2026
Colorado is about to become an autonomous taxi battleground, and the reason is snow. This year, Waymo is expanding its fully autonomous service into Denver, and that means its cars are about to face something that has always plagued camera-only autonomous driving: What happens when roads stop looking like roads?
Waymo is going to start robotaxi operations in Denver as winter approaches, which means we’re about to get a real-world test of what happens when snow piles up, lane markings disappear, the visual cues fully autonomous vehicles often rely on become significantly harder to interpret, and the vehicle’s cameras themselves get snowed in. The question isn’t whether Waymo will be perfect. Of course it won’t be. The question is whether its combination of sensors gives it enough information to keep going when the cameras aren’t getting a clean picture.
This is where Tesla is running blind. The fundamental difference between Waymo, Zoox, and Tesla robotaxis is the technology stack. Waymo and Zoox use LiDAR, radar, and cameras, giving the system multiple ways to detect what’s around. All sensors work together to draw a complete picture, and if one sensor struggles, another can fill in. That’s redundancy that fails not only “safe,” but “functional,” as a Waymo in Denver with a snow-obstructed visual camera should still be able to get its passengers to their destination.
Tesla, meanwhile, has been (some might say) stubbornly committed to vision-only autonomy. Visual-only cues may work remarkably well under perfect conditions, but “remarkably well” isn’t what’s needed for true, fully autonomous robotaxis. If the goal is Level 5 autonomy (which is autonomous driving that operates in all conditions, on all roads, at all times…no ifs, ands, or buts) then I don’t see how cameras alone get you there. At some point, the car has to be able to deal with the environment when the environment itself makes vision impossible.
Denver is about to provide a pretty useful demonstration of how much closer to Level 5 Waymo is compared to Tesla.
The timing is especially brutal for Tesla. While Tesla is talking up its Cybercab and trying to convince investors that its self-driving technology is the future, Waymo is heading to Denver, probably to prove a point.
Tesla’s biggest advantage has always been scale. If it gets its technology working at the level Elon Musk claims it can, Tesla could potentially put enormous numbers of autonomous vehicles on the road within hours. The problem is, Tesla is going to be stuck at Level 4 autonomy, where the vehicle can handle all of the driving tasks and safely navigate without human intervention…within specific limits.
To be clear, Waymo and Zoox robotaxis are categorized as Level 4, too. But in the race to achieve true Level 5 autonomy, Tesla is simply missing the underlying tech. Tesla’s robotaxis may be autonomous within defined conditions and locations, but the company’s camera solution can’t offer the anywhere, anytime requirement to be true Level 5 autonomy. Tesla can call its tech “Full Self Driving” all it wants, but the fact is, FSD comes with an asterisk.
Some (like Mr. Musk) will insist I’m wrong about Tesla’s tech, but eventually, the road itself is going to show us who’s right.
(Image courtesy Waymo)
