Waymo has deployed its own 5nm chip in the Ojai robotaxi and is reducing its dependence on Nvidia and AMD
Waymo has deployed its own 5nm ASIC chip (over 1000 TOPS, manufactured by TSMC) in the Ojai robotaxi, which operates in Phoenix, Los Angeles and San Francisco. The goal is to reduce costs and dependence on chips from Nvidia and AMD; the vehicles have driven over 200 million miles.
Waymo (a subsidiary of Alphabet) confirmed that it has developed an in-house chip for the latest generation of its vehicles — its first-ever in-house processor. It is a 5nm ASIC chip manufactured by TSMC that, according to Waymo, delivers over 1000 trillion operations per second (TOPS) and processes raw sensor data (radar, lidar, cameras) before it reaches the main machine learning model that makes driving decisions. According to Waymo, this architecture is intended to ensure that driving commands are executed within milliseconds; the company cites response speed, resilience in various climate and road conditions, and built-in redundancy for safety as key features of the chip.
The chip is deployed in the Ojai robotaxi model, which debuted this year and is now in commercial service in Phoenix, Los Angeles and San Francisco. Waymo says its computing system has grown twentyfold over the past eight years and that its vehicles have driven over 200 million miles in autonomous mode. The Ojai vehicle is being developed in collaboration with the Chinese manufacturer Zeekr, but the chip itself is manufactured by TSMC.
Waymo began developing its own chip to reduce costs and limit its dependence on chips from Nvidia and AMD, which it had relied on until now. According to the company, it continues to partner with AMD, Micron, Nvidia, Samsung, SanDisk, Socionext and TSMC for tasks outside machine learning, namely CPUs, GPUs and other accelerators. Waymo has not disclosed how much the chip development cost; in February this year, it secured an investment of 16 billion dollars at a valuation of 126 billion dollars, and earlier in the year it announced the milestone of half a million paid rides per week. The company now operates in 11 US cities, lists another 19 as planned, and is preparing for international expansion to Tokyo and London.
Why it matters
An in-house chip gives Waymo greater control over costs and the supply chain for a key component of autonomous driving, replacing its previous practice of purchasing chips from Nvidia and AMD. For the autonomous mobility and semiconductor sectors, this signals that large robotaxi operators may increasingly replace external GPU/AI accelerators with their own specialized silicon, which could affect demand for hardware from these suppliers in this segment over time.
What was added since the original report
Verified updates
-
Specific vehicle model: Ojai; Specific deployment locations: Phoenix, Los Angeles, San Francisco; Waymo has completed 200 million miles of autonomous driving; The chip is described as an ASIC (application-specific integrated circuit); The chip directly processes sensor data (radar, lidar, camera) for real-time ML decision-making
- Specific vehicle model: Ojai
- Specific deployment locations: Phoenix, Los Angeles, San Francisco
- Waymo has completed 200 million miles of autonomous driving
- The chip is described as an ASIC (application-specific integrated circuit)
- The chip directly processes sensor data (radar, lidar, camera) for real-time ML decision-making
Relevant practical impact
What this means
For a business
Vertical integration at Waymo (an in-house chip instead of purchases from Nvidia and AMD) demonstrates a strategy of reducing costs and dependence on suppliers for autonomous vehicles, which is a relevant signal for companies in the semiconductor and autonomous mobility sectors monitoring competition with Nvidia and AMD.
Strategy More business impacts →Check the original
Event sources
confirmed by 2 independent sources · 2 publishers, 2 independent. We count feeds from the same owner only once.