Waymo Challenges Tesla’s Camera-Only Autonomy Ahead of Cybercab Launch

Waymo’s critique, delivered through a blog post and interviews, frames its multi-sensor approach—combining lidar, radar, and cameras—as the only responsible path to full autonomy. Srikanth Thirumalai, a vice president of driving software at Waymo, warned that neural networks processing raw pixels are prone to black-box failures. He emphasized that in physical AI, there is no option to reboot or refresh when a system hallucinates, a direct jab at Elon Musk’s long-standing dismissal of lidar as a technological crutch.

This friction highlights a massive divergence in strategy. While Waymo operates roughly 4,000 robotaxis across 14 cities with a proven, albeit expensive, hardware-heavy model, Tesla is betting its future on an AI-first approach. By eliminating steering wheels and pedals in the upcoming Cybercab, Tesla aims for a cost-efficient, scalable design that could undercut competitors. However, the company faces significant pressure to prove its software can reliably navigate complex environments like school zones and harsh weather—hurdles Waymo has spent years documenting in over 200 million real-world miles.

Industry analysts remain divided, with some viewing Waymo’s stance as the defensive maneuvering of an incumbent. Yet, the stakes are undeniable. With Tesla recently registering Cybercabs in Texas and aiming to produce over 125,000 units annually, the competition is shifting from academic debate to a high-stakes race for a market potentially worth hundreds of billions of dollars. If Tesla demonstrates that its vision-based system can achieve true autonomy at scale, it could force a radical shift in how the entire industry approaches vehicle architecture and cost.