Waymo has revealed a custom 5-nanometer application-specific integrated circuit designed specifically for its autonomous-driving system, giving the Alphabet-owned company greater control over the computing hardware that processes enormous amounts of information from cameras, lidar, and radar. The specialized chip delivers more than 1,000 trillion operations per second of machine-learning performance and helps Waymo vehicles translate sensor information into driving decisions within milliseconds. The development represents an important step toward vertically integrating autonomous vehicles, potentially improving performance, energy efficiency, reliability, and ultimately cost as Waymo expands commercial robotaxi operations. The technology also illustrates a broader shift toward specialized silicon for physical artificial intelligence, where autonomous vehicles and eventually advanced robots will require powerful computers capable of operating reliably outside controlled data-center environments.
Key Takeaways
- Waymo designed a purpose-built 5nm ASIC capable of more than 1,000 TOPS of machine-learning performance, processing raw camera, lidar, and radar information before it reaches the central autonomous-driving system.
- The company’s onboard computing architecture emphasizes low latency, durability, energy efficiency, and redundancy, allowing autonomous vehicles to make decisions within milliseconds while operating without a human driver available as a backup.
- Waymo’s investment in proprietary silicon could provide greater control over performance and costs as robotaxi fleets scale, while demonstrating that competition in autonomous transportation increasingly depends on semiconductor engineering as well as artificial intelligence and automotive technology.
In-Depth
Waymo’s decision to design its own 5-nanometer application-specific integrated circuit marks another step toward vertical integration in autonomous transportation. The chip processes raw information from cameras, lidar, and radar before that data reaches the broader driving system, giving the vehicle the computational speed required to interpret its surroundings and respond within milliseconds. Waymo says the ASIC provides more than 1,000 trillion operations per second of machine-learning performance.
The significance extends beyond raw computing power. Robotaxis must carry data-center-class processing capability inside a vehicle, where electricity consumption, heat, vibration, physical space, and reliability impose severe constraints. Waymo has engineered its onboard system around responsiveness, ruggedness, and redundancy, including parallel computing capability designed to preserve operation if one side encounters a fault. Its sixth-generation system also processes feeds from 13 high-resolution cameras while supporting lidar and radar inputs.
This strategy could strengthen Waymo’s control over costs and technological development as autonomous fleets expand. Rather than relying exclusively on general-purpose processors, purpose-built silicon allows hardware, sensors, and artificial-intelligence models to be optimized together. Waymo nevertheless continues working with major semiconductor suppliers, including AMD, Nvidia, Samsung, Micron, Socionext, Sandisk, and TSMC.
The larger lesson is that autonomous transportation is increasingly becoming a semiconductor competition as much as an automotive one. Companies capable of controlling their computing architecture may gain advantages in performance, energy efficiency, reliability, and cost. Those advantages will matter as robotaxi operators attempt to move from impressive demonstrations toward profitable, large-scale commercial transportation networks operating across diverse American cities and driving conditions.
Sources
- https://waymo.com/blog/2026/08/look-under-our-trunk/
- https://techcrunch.com/2026/08/23/techcrunch-mobility-the-custom-chip-driving-waymos-robotaxi-ambitions/
- https://www.techrepublic.com/article/news-waymo-custom-ai-chips-robotaxis/
- https://siliconangle.com/2026/08/20/waymo-details-the-custom-chip-in-its-autonomous-driving-system/
- https://thenextweb.com/news/waymo-custom-chip-robotaxi-tsmc-ojai