Wayve’s AI will reach Mercedes cars within two years, Kendall says
Wayve's CEO, Alex Kendall, has predicted that its AI technology will be integrated into Mercedes consumer cars within the next two years, with Tokyo robotaxis following by the end of the year. This announcement came during his speech at HumanX in Amsterdam on Wednesday.
"It’s going to take them a while. Maybe they don’t know that yet," Kendall remarked about humanoid robot makers.
Kendall emphasized that self-driving cars, not robots, will be the first physical AI to reach widespread adoption, similar to the impact of ChatGPT.
Mercedes Partnership and Future Plans
The recent global partnership between Wayve and Mercedes-Benz solidifies this timeline. Nissan and Stellantis have already integrated Wayve's software into their production plans, followed by Mercedes.
Wayve demonstrated its prototype cars equipped with the AI on Mercedes production hardware in London, Stuttgart, and San Francisco. The system operates on compact, low-compute computers, utilizing the cameras and radar commonly found in mass-produced vehicles.
The next step involves integrating the technology into production cars, focusing on latency, safety, and software compatibility with various car components.
Licensing Over Building Cars
Unlike Tesla, which designs and builds its own cars, Wayve licenses its foundation model to carmakers and fleet operators. This model has accumulated driving data in over 500 cities. Car manufacturers can customize the AI's behavior by adjusting parameters, while Wayve benefits from partner contributions, enabling it to train at a massive scale.
Kendall acknowledged the challenges ahead, including product design, manufacturing, and deployment cycles, as well as handling unforeseen scenarios like northern Sweden's roads or unfamiliar Japanese signage.
The Future of Physical AI
When asked about Nvidia CEO Jensen Huang's claim that physical AI has reached its "ChatGPT moment," Kendall disagreed. He pointed out that only a small fraction of the population interacts with physical AI today, suggesting that widespread adoption is still years away. He foresees business applications preceding consumer ones and mobility services before robot arms.
Currently, fewer than 10,000 robotaxis operate globally, while approximately 100 million cars are on the road.