In a swipe at Tesla, Waymo says ‘cameras… aren’t enough’

A top software executive at Waymo recently published a blog post detailing lessons learned from driving over 200 million autonomous miles, which serves as a direct critique of Tesla's self-driving technology strategy.

The publication argues that relying exclusively on cameras is insufficient for achieving safe, scaled driverless operations. While Tesla has historically dismissed lidar as unnecessary and costly, preferring a vision-only approach, Waymo asserts that a redundant sensor suite consisting of cameras, radar, and lidar is essential to build a reliable worldview. This multi-sensor model ensures that a vehicle maintains situational awareness even if a single sensor fails, a safety precaution that some state lawmakers are considering making mandatory for robotaxis.

The debate also extends to the use of high-definition maps and the developmental pathway to full autonomy. While Tesla rejects pre-mapping in favor of immediate visual analysis to handle changing road conditions, Waymo utilizes continuously updated, AI-driven high-definition maps as a foundational guiding memory. Crucially, the Waymo executive argues that driver-assist technologies cannot simply be iterated until they achieve full autonomy, calling this approach a false path. Instead, true autonomous systems must be built from the ground up and rigorously validated without human supervision. Currently, Waymo maintains a massive commercial lead in active robotaxi deployments, while Tesla continues to rely on driver-supervised software as it prepares its own autonomous vehicle platforms.