The new chip delivers over 1,000 TOPS of computing performance, placing it on par with Nvidia’s latest DRIVE AGX Thor processor. By handling raw data streams from the 13 high-fidelity cameras mounted on Waymo’s Ojai vehicle before the information reaches the core software brain, the system gains the reaction speeds necessary for dense urban environments. While Waymo maintains control over this custom architecture, it continues to rely on a collaborative supply chain, partnering with firms such as AMD, TSMC, and Samsung to manufacture and support the underlying compute infrastructure.
Waymo bets on custom silicon to scale its robotaxi fleet
Waymo is moving deeper into hardware vertical integration, unveiling a custom 5 nm ASIC chip designed to process the massive data loads required for its sixth-generation self-driving system. This proprietary silicon serves as a critical performance upgrade, aiming to bolster efficiency as the company expands operations across major U.S. cities.
This push for hardware efficiency coincides with broader industry shifts toward autonomy. While Waymo refines its own tech stack, other players are recalibrating their strategies. Serve Robotics, for instance, has diversified its partnerships with Grubhub and DoorDash following Uber’s decision to exit their agreement. Meanwhile, the autonomous sector faces growing regulatory scrutiny, from the Idaho National Laboratory’s investigation into potential security risks of Chinese-made lidar sensors to the National Highway Traffic Safety Administration’s ongoing inquiry into a Waymo robotaxi collision involving a child.
Comments (0)
No comments yet. Be the first!