What actually limits robotaxi scale: onboard compute or driverless deployment?
Waymo revealed its in-vehicle compute stack for the first time, including a custom 5-nanometer ASIC delivering 1,000 TOPS of front-end sensor processing and a redundant architecture designed for low-latency operation.[4] Separately, Robotaxi Tracker recorded all 170 monitored Tesla rides in Austin over two weeks as unsupervised, across 54 vehicles, although the apparent increase partly reflected revised data sources.[5] Waymo operates about 4,000 vehicles and roughly 500,000 paid trips weekly, while Tesla is preparing to introduce its steering-wheel-free, pedal-free Cybercab in Austin.[4][5]
The disclosures illustrate two separate scaling tests: Waymo must reduce the cost and complexity of data-center-class hardware operating inside a vehicle, while Tesla must demonstrate that unsupervised service can expand reliably ahead of a purpose-built vehicle rollout.[4][5]
Key insights
- Waymo’s architecture processes camera, lidar, and radar streams onboard so driving decisions can be made in milliseconds without human backup.[4]
- Waymo says its compute capacity has increased twentyfold in eight years and is designed around responsiveness, ruggedization, and redundancy.[4]
- The company builds its own front-end sensor-fusion silicon but relies on suppliers including AMD, Micron, Nvidia, Samsung, SanDisk, Socionext, and TSMC for the broader system.[4]
- Tesla’s observed Austin fleet remains smaller than Waymo’s local operation, which comprises more than 300 driverless vehicles.[5]