Waymo CEO Dmitri Dolgov says pure‑vision approach limits full autonomy

Waymo co‑CEO Dmitri Dolgov spoke at Y Combinator’s Startup School and argued that relying solely on cameras – the “weak sensing” approach championed by Tesla – cannot meet the safety standards required for fully autonomous driving. He said cameras can match human‑level performance and support driver‑assist systems, but they hit a safety ceiling well before the reliability needed for driverless operation.

Dolgov explained Waymo’s sensor‑fusion strategy, which combines high‑resolution cameras, LiDAR and millimetre‑wave radar. He cited real‑world cases where cameras failed – such as a dust storm in Phoenix and night‑time barrier breaches – while LiDAR and radar continued to detect pedestrians and obstacles. He also highlighted a simple failure mode: a single leaf covering a camera can halt a vehicle, whereas other sensors keep it safe.

The remarks were widely read as a direct challenge to Tesla’s “Tesla Vision” roadmap, which has eliminated radar and other sensors in favor of camera‑only perception. Dolgov referenced the exponential “nine” reliability curve, noting that each additional 9% in safety requires roughly ten times more effort, making pure‑vision a risky path for full autonomy. Waymo, with over 2.2 billion miles of autonomous‑vehicle testing and robotaxi services in 15 U.S. cities, continues to pursue the multi‑sensor model.