Waymo Unveils Custom 5nm Chip
Alphabet-owned Waymo has disclosed new technical details regarding the computing system that powers its robotaxi fleet, highlighting a custom 5-nanometer application-specific integrated circuit (ASIC). The proprietary chip performs front-end processing to clean and combine data from the vehicle's cameras, lidar, and radar before passing it to the broader artificial intelligence system. Delivering 1,000 trillion operations per second—equivalent to one quadrillion operations per second—the component forms a central part of Waymo's latest onboard computing architecture.
Continued Reliance on Major Hardware Suppliers
Despite developing specialized silicon internally, Waymo's disclosure confirms that its system still relies on established semiconductor suppliers. The company named AMD, Micron, Nvidia, Samsung, SanDisk, Socionext, and TSMC as hardware partners supporting the platform. Waymo focuses its internal engineering resources on real-time sensor fusion and front-end machine learning, while utilizing third-party CPUs, GPUs, accelerators, memory, and storage for other workloads.
Waymo did not disclose specific contract values, purchasing volumes, or the exact components provided by each named supplier. Consequently, the disclosure represents an overview of the existing hardware supply chain rather than newly signed partnerships or major new semiconductor orders.
Scaling the Sixth-Generation Driver
Waymo has increased its onboard computing capacity twentyfold over the past eight years to handle the demands of responsiveness, durability, and redundancy required for autonomous driving. Immediate decisions must be processed within the vehicle to avoid the latency and connectivity risks associated with cloud computing.
The company is currently preparing to deploy its sixth-generation Driver across multiple platforms, beginning with the purpose-built Ojai robotaxi and integration into the Hyundai IONIQ 5. Waymo's factory in Mesa, Arizona, is scaling toward a potential manufacturing capacity of tens of thousands of autonomous vehicles annually, with the Ojai vehicle initially slated for selected riders in San Francisco, Phoenix, and Los Angeles.
