Elon Musk believes that anything other than cameras isn’t necessary for self-driving, but Waymo has thought deeply about integrating LiDARs into its self-driving mix.
Waymo CEO Dmitri Dolgov recently sat down for an interview to explain, from first principles, how the company’s three sensor types — LiDAR, radar, and cameras — work together rather than compete. His explanation cuts through the debate cleanly: it isn’t a question of which sensor wins, but of what each one uniquely contributes.
“They’re very complementary,” Dolgov said. “It’s all about blasting photons out there and then they bounce off something. They come back, you measure what comes back. The frequencies are very different.”
On LiDAR, he was precise: “Laser gives you very high resolution. Think of it as a laser beam that goes out, spins around. It shoots up millions of laser pulses per second, and then each one comes back — you’re kind of sampling the 3D structure of the world with very high resolution. LiDAR is great for very fine-grained mapping.”
Radar, by contrast, sacrifices resolution for resilience. “Radar has much lower resolution, but because of the physics of it, it degrades much better in adverse weather conditions — fog, snow, heavy rain. It can be occluded by verticals between it and the targets. Imagine driving in super dense fog. We’re close to San Francisco, so probably don’t have to think too hard about that. It can be really hard to see. Cameras degrade. Laser, depending on the size of the particulates, can degrade better or worse than cameras. Radar is not well affected.”
This has a concrete, counterintuitive implication: radar can effectively see through conditions where cameras go blind. “You can imagine driving on a freeway — radar will give you really good returns for cars that are absolutely invisible in the camera space.”
When asked whether that means Waymo switches to radar in bad weather, Dolgov pushed back on the framing entirely. The system doesn’t toggle between sensors — it fuses them continuously.
“It’s a combination of the sensors. Each one is noisy. How the noise characteristics show up in different environments is different. It’s not like we switch from one to another. It’s not like we estimate what’s happening in the world through cameras and then through radar and then through LiDAR and then compare. There’s an encoder for camera, there’s an encoder for LiDAR, and they all go into the system that gives you jointly the best view of what’s happening in the world.”
“So if it’s a nice, bright, sunny day, cameras are very valuable. If it’s pitch dark, or you have the sun in your face, or you’re blinded by the headlights from an oncoming car, the camera will degrade — there’s still some noisy signal, but it will degrade. And LiDAR is completely unaffected.”
Dolgov’s explanation reflects an engineering philosophy that has become Waymo’s central differentiator — and its sharpest point of contrast with Tesla. Musk has long argued that cameras alone suffice, that humans drive with just vision, and that adding LiDAR or radar only creates conflicting signals. Waymo’s position is the inverse: redundancy isn’t a liability, it’s the whole point. If a camera is blinded by glare, radar keeps tracking. If fog scatters laser pulses, radar returns remain clean. Each sensor’s failure mode is different, and fusing them — rather than arbitrating between them — is what makes the perception stack robust.
The practical stakes have grown considerably. Waymo recently began operating on freeways across the San Francisco Bay Area, Phoenix, and Los Angeles — a capability Musk had suggested its sensor complexity would prevent. The company has since expanded to 10 U.S. cities and claims its vehicles are involved in 11 times fewer serious injury collisions than human drivers. Its 6th-generation sensor suite — 13 cameras, 4 LiDARs, and 6 radar units — represents years of iterating on exactly the sensor fusion logic Dolgov describes. Andrej Karpathy, Tesla’s former AI director, has acknowledged that Waymo will have early gains from its sensor-rich approach, even as he bets on Tesla’s camera-only method winning long-term on scalability.
That debate remains open. But Dolgov’s framing is a reminder that Waymo’s bet isn’t on any single sensor — it’s on the idea that the world is too noisy and too varied for any one modality to handle alone.