Alphabet’s (NASDAQ: GOOGL) Waymo co-chief executive Dmitri Dolgov has publicly questioned the viability of camera-only driverless technology, a strategy central to Tesla’s (NASDAQ: TSLA) autonomous vehicle ambitions.
Speaking at Y Combinator’s Startup School, Dolgov stopped short of naming Tesla directly but made clear his position on single-sensor approaches to self-driving systems.
Dolgov described camera-only systems as “weak sensing,” arguing that while such technology may develop rapidly at first, it ultimately hits a lower ceiling of capability and long-term performance.
The argument for camera-only systems has long rested on two pillars: lower costs and the fact that human drivers rely entirely on vision to navigate roads safely.
Dolgov acknowledged that a camera-only system could reach human-level driving performance, but argued that merely matching human ability falls short of the actual goal, which is building technology that is demonstrably safer than human drivers.
Waymo uses three sensor types simultaneously, combining cameras, LiDAR, and radar into a unified picture of the vehicle’s surroundings, a fundamentally different approach from Tesla’s cost-focused strategy.
“These different sensing modalities, they’re not backups to each other,” Dolgov said during the presentation, adding that the fused data produces a view of the world that is “vastly superior to what you get with any one sensor.”
Real-world scenarios illustrate the gap clearly, as a whiteout snowstorm could render a camera-only system effectively blind while LiDAR would still reliably detect obstacles and pedestrians on the roadside.
Even a mundane event such as mud covering a camera lens could fully disable a camera-only vehicle, whereas a system equipped with LiDAR and radar could navigate safely back to base without relying on visual input.
For Tesla investors, the criticism carries added weight given that Morgan Stanley automotive analyst Adam Jonas values autonomous driving and the robotaxi business at 41% of Tesla’s total valuation, compared to 34% from core automotive and energy operations and roughly 25% from its Optimus robotics potential.
The stakes are further raised by Tesla’s existing challenge of needing to replace or fairly compensate owners of roughly 4 million vehicles after the company admitted its Hardware 3 chip is not powerful enough to deliver unsupervised self-driving as originally advertised.
Waymo’s public critique lands at a moment when Tesla’s driverless strategy faces scrutiny on multiple fronts, making the camera-only approach a question that investors and regulators alike will be watching closely.
