Waymo Unveils Custom 5nm Chip for Autonomous Driving Platform

Waymo Develops Custom Silicon for Autonomous Driving

Waymo has revealed its custom 5nm application-specific integrated circuit as part of the computing technology supporting its Driver autonomous platform. The Waymo autonomous driving chip is designed to work alongside other computing components and Waymo-developed sensors to improve the performance of its Ojai robotaxis in complex urban environments. The development also gives the company greater control over the hardware supporting autonomous driving functions while reducing reliance on externally supplied computing technology. Waymo said the approach brings silicon, sensors and algorithms together during development, allowing the system to improve sensor fidelity, bandwidth efficiency and the execution of different machine-learning models.

Custom 5nm ASIC Targets Robotaxi Computing Performance

The new 5nm ASIC will operate alongside other chips and Waymo's internally designed sensing hardware to support faster responses and improved navigation in demanding urban conditions. The company said developing these technologies together enables it to handle a heterogeneous range of models, extending from sparse convolutions to dense transformers. Historically, Waymo has relied on semiconductor products from companies such as Nvidia Corp and Advanced Micro Devices Inc. Its move toward proprietary silicon therefore represents an expansion of its hardware strategy rather than an immediate replacement of all third-party computing components used by the autonomous driving platform.

Waymo Expands Hardware Control as AI Computing Costs Rise

The custom silicon effort also aligns with a broader push by Google and its parent company Alphabet to develop specialized computing hardware and gain greater control over the infrastructure required for artificial intelligence workloads. For Waymo, designing silicon alongside sensors and autonomous-driving software can provide a more integrated development path and potentially improve computing efficiency. The strategy is particularly relevant as increasingly sophisticated autonomous systems require substantial processing capacity for perception, prediction, planning and navigation. By developing part of this computing stack internally, Waymo is also diversifying its technology supply base while maintaining the ability to use products from external semiconductor suppliers.

Waymo Continues Integrating Autonomous Technology Into Robotaxis

Waymo currently imports Zeekr vans from China before installing its autonomous driving hardware and software at a facility in Mesa, Arizona. This operating model gives the company a defined vehicle platform on which it can integrate its sensing, computing and software systems before deployment. The introduction of proprietary silicon adds another internally controlled element to that technology stack. Combined with Waymo's own sensors and autonomous-driving algorithms, the custom ASIC is intended to strengthen the computing foundation supporting robotaxi operations as the company continues developing systems capable of handling complex real-world driving environments.

Frequently Asked Questions

What is Waymo's new autonomous driving chip?
Waymo's new autonomous driving chip is a custom 5nm application-specific integrated circuit developed to support its Driver autonomous platform and robotaxi operations. The chip will operate alongside other computing components and Waymo-designed sensors rather than functioning as an isolated replacement for every third-party processor. Its development is intended to improve computing performance, sensor processing and navigation capabilities in complex urban environments. Waymo has also emphasized the benefit of developing silicon, sensors and algorithms together, allowing its engineers to optimize hardware and software for different autonomous-driving models and workloads.

Why is Waymo developing its own semiconductor technology?
Waymo is developing its own semiconductor technology to gain greater control over autonomous-driving computing while diversifying its supply beyond external chip providers. The company has historically used products from suppliers including Nvidia and Advanced Micro Devices, but proprietary silicon gives it another source of computing capability tailored to its autonomous-driving requirements. The initiative also coincides with Alphabet's wider effort to develop custom chips and manage the rapidly increasing costs associated with artificial intelligence computing infrastructure. Waymo's approach therefore combines performance optimization with greater control over an important part of its technology stack.