%20--%3e%3cdefs%3e%3cstyle%3e%20.st0%20{%20fill:%20%23061b40;%20}%20.st1%20{%20fill:%20%23306af1;%20}%20.st2%20{%20fill:%20%235ce5cf;%20}%20%3c/style%3e%3c/defs%3e%3cg%3e%3cpath%20class='st0'%20d='M55,10.5h9v3h-9v9h12V7.5h-12v3ZM64,19.5h-6v-3h6v3Z'/%3e%3cpolygon%20class='st0'%20points='69%2016.5%2078%2016.5%2078%2019.5%2069%2019.5%2069%2022.5%2081%2022.5%2081%2013.5%2072%2013.5%2072%2010.5%2081%2010.5%2081%207.5%2069%207.5%2069%2016.5'/%3e%3cpolygon%20class='st0'%20points='95%2010.5%2095%207.5%2083%207.5%2083%2022.5%2095%2022.5%2095%2019.5%2086%2019.5%2086%2016.5%2095%2016.5%2095%2013.5%2086%2013.5%2086%2010.5%2095%2010.5'/%3e%3cpath%20class='st0'%20d='M40,1.5v21h11.6l1.4-1.4v-7.6h0c0,0-1.4-1.5-1.4-1.5l1.4-1.4V2.9l-1.4-1.4h-11.6ZM50,19.5h-7v-6h7v6ZM50,10.5h-7v-6h7v6Z'/%3e%3c/g%3e%3cpath%20class='st1'%20d='M23.1,24L14.7,4.8l-4.9,11.2h3.8l-1.8,4H3.3L12.1,0H2C.9,0,0,.9,0,2v20c0,1.1.9,2,2,2h21.1Z'/%3e%3cpath%20class='st2'%20d='M34,0h-16.8l10.6,24h6.2c1.1,0,2-.9,2-2V2C36,.9,35.1,0,34,0ZM32.5,20h-4V4h4v16Z'/%3e%3c/svg%3e)
In the vast landscape of physical AI, autonomous driving is seen as the first "hard nut" to crack. Recently, Wu Xinzhou, Vice President of NVIDIA, shared the company's grand blueprint in the field of intelligent driving at a communication session in Beijing. He not only detailed the "five-layer cake" system that supports assisted driving but also clearly outlined a specific timetable for the implementation of Level 4 autonomous driving.
The "Five-Layer Cake" Builds a Full-Stack Ecosystem
NVIDIA is no longer satisfied with just being a chip supplier. Instead, it aims to build a complete service system through "three computers" (vehicle-side inference, cloud training, and simulation verification). Wu Xinzhou metaphorically called it the "five-layer cake": from the underlying hardware platform Hyperion, to the operating system, open model Alpamayo, simulation toolchain, up to the top-level cloud infrastructure.
The core of this system lies in lowering the development threshold for automakers. Especially during the transition from modular to "end-to-end" architecture, NVIDIA leverages its strong simulation capabilities to perform 2 million scenario validations daily, greatly improving model training efficiency. Currently, NVIDIA is actively promoting major car manufacturers to join the Hyperion platform, aiming for a leap in standardization and scalability.
Stick to the Visionary Approach and Redundancy Strategy
In terms of technical routes, Wu Xinzhou is a firm "visionary". He believes that the pixel density and upper limit of visual sensors far exceed those of LiDAR, which are sufficient to support high-level assisted driving. However, for higher-level L3 and L4 systems, NVIDIA still considers LiDAR an indispensable safety redundancy. He revealed that NVIDIA is working with suppliers in Europe and the US to find stable hardware support for high-level intelligent driving solutions.
Countdown for Level 4: Covering 30 Cities by 2028
Regarding the industry debate about "jumping directly from L3 to L4", Wu Xinzhou takes a pragmatic approach. He believes that L3 has immediate value in freeing up human labor, while L4 requires substantial cloud operation capabilities. According to NVIDIA's roadmap: In 2025, it will deliver a mass-produced project in collaboration with Mercedes-Benz; in 2027, it will launch an L4 pilot with Google; by 2028, NVIDIA plans to provide driverless services during the Los Angeles Olympics in collaboration with Uber, targeting coverage of 20 to 30 cities globally.
Extension of Physical AI: From Cars to Robots