Waymo is publicly challenging one of Tesla's most fundamental autonomous-driving assumptions, arguing in a new technical explainer that cameras alone are insufficient for safe Level 4 autonomy just as Tesla prepares to widen Cybercab deployment.
200 Million Miles Behind The Multi-Sensor Case
Drawing on more than 200 million fully autonomous miles, Waymo says reliable driverless operation at scale requires redundant perception from cameras, lidar and radar. Tesla has taken the opposite line, arguing that increasingly capable neural networks can reach autonomy primarily through vision, on the grounds that humans navigate roads mostly through sight. The disagreement is no longer academic: Waymo is preparing a 2027 Munich robotaxi launch, and Tesla is scaling its purpose-built Cybercab in the United States.
Regulation May Pick The Winner
Regulators are increasingly weighing in. A proposed New Jersey robotaxi framework would require multiple sensors on any commercial autonomous vehicle, a rule that would effectively exclude a pure vision system. Waymo says its multi-sensor architecture offers redundancy when individual systems degrade, while Tesla argues that eliminating lidar dramatically lowers per-vehicle cost and makes autonomy easier to scale — a philosophical divide that echoes debates seen in trucking, where Gatik's driverless push takes yet another middle path.
Two Bets On Physical AI
The dispute represents two very different bets on physical AI: whether massive datasets and increasingly capable vision models can substitute for hardware redundancy, or whether safety-critical machines require both. As commercial deployments scale into more cities and highway corridors, the regulatory answer will do as much as engineering to define what a robotaxi can look like.
Reporting based on coverage from The Verge.