
Graham Hope
· 1 min read
Waymo Develops Its Own Chip for Self-Driving
Heather Diehl via Getty Images
Self-driving company Waymo has revealed it has developed its own custom-built chip for its latest generation of vehicles.
The chip was introduced in a blog post by vice president of engineering Satish Jeyachandran and compute lead Daniel Rosenband, which insight into the tech that has secured Waymo’s top position in the robotaxi rollout in the U.S.
The new chip plays a key role in Waymo’s ever-evolving autonomous driving system, which the company claimed, “would be considered impressive for a data center, with the added complexity of an in-vehicle operating domain and real-time requirements.”
The chip -- which is Waymo’s first-ever custom silicon -- has been created to convert raw data from a robotaxi’s sensors, including radar, lidar and cameras, before it reaches the vehicle’s core machine learning brain.
Waymo described the 5-nanometer application-specific integrated circuit as “a specialized machine learning powerhouse” that can deliver more than 1,000 trillion operations per second to ensure driving commands can be executed in milliseconds.
This responsiveness is one of three keys Waymo said are integral to the design of its compute, which has grown in size 20-fold over the past eight years and has now delivered more than 200 million miles of autonomous driving. The others are durability -- essential when robotaxis are expected to perform in widely varying road conditions and climates across the U.S. -- and built-in redundancy, for safety.
Waymo said it will continue to focus on infrastructure products. At the same time, Waymo plans to continue working with outside companies, citing non-machine-learning tasks in which it pairs its tech with CPUs, GPUs and accelerators, and naming AMD, Micron, Nvidia, Samsung, SanDisk, Socionext and TSMC as existing partners.
The new chip will be featured in the company’s latest Ojai robotaxi, which debuted earlier this year and is now in commercial service in Phoenix, Los Angeles and San Francisco.
Original source
This story was published by AI Business and written by Graham Hope. SyncAI.news shows a preview; the complete article is on the publisher's site.
Read the full story on aibusiness.com


