Waymo has designed a custom 5nm processor for its robotaxis and published the seven suppliers it builds the computer with. Nvidia is still on that list.Key Highlights Waymo revealed a custom 5nm ASIC on 20 August 2026, already running in its latest robotaxi generation. The chip handles sensor data before the driving brain sees it, and Waymo credits it with over 1,000 TOPS of front-end machine- learning performance. Nvidia, AMD, Micron, Samsung, SanDisk, Socionext and TSMC all appear on the supplier list Waymo published alongside the chip.Waymo put out the design of its robotaxi computer on 20 August 2026, introducing a custom 5nm chip of its own and naming the seven companies it builds the system with, an unusual level of disclosure from a business that guards its hardware details. Waymo says the chip is in production in its latest robotaxi generation, which is the Ojai, opened to all riders in Los Angeles, Phoenix and San Francisco the day before. What it does is narrower than the headlines suggest, and more interesting: it decides what the car sees, while other silicon decides what the car does. Nvidia and AMD both sit on that supplier list.About The AuthorAt heart, I am a storyteller drawn to the watershed moments that bend the technology landscape. I braid narrative with data, humanise statistics, and trace the arc from first spark to world-changing impact. My reportage, features and reviews are witty, sardonic, visual and vivid, using anecdote to illuminate rather than eviscerate.
Waymo Built A 5nm Chip That Decides What Its Robotaxis See
Waymo has built a custom 5nmASIC that processes raw camera, lidar and radar data inside itsrobotaxis. The chip runs ahead of the main driving brain rather thanreplacing it, and Nvidia and AMD remain suppliers. Waymo says theASICs alone deliver over 1,000 TOPS of machine-learning performance.
Waymo's Ojai robotaxis use a 5nm ASIC that preprocesses sensor data at 1,000+ TOPS, offloading work before the driving stack. This signals that vertically-integrated sensor silicon plus commercial GPUs for inference is the battle-tested production model for autonomous systems.










