Legacy AI Influences NVIDIA’s Self-Driving Vehicle Strategy
25 March 2025 · Uncategorized ·
Source: · https://technews.tw/2025/03/19/nvidia-gtc-2025-jensen-huang/
At the 2025 GPU Technology Conference (GTC), Nvidia CEO Jensen Huang upheld his tradition of delivering key highlights and shared a compelling historical anecdote.
During discussions about autonomous vehicle technology, Huang referenced 'AlexNet,' which garnered significant attention in 2012 for its victory in an image recognition competition. Developed by computer scientist Alex Krizhevsky with the assistance of Ilya Sutskever (later one of OpenAI's founders) and AI researcher Geoffrey Hinton, this neural network architecture achieved a remarkable accuracy rate of 84.7% in the ImageNet academic contest.
This achievement revitalized industry interest in deep learning technology—a machine-learning approach utilizing neural networks. AlexNet’s success spurred NVIDIA to fully commit resources toward autonomous vehicle research.
“When I first saw AlexNet – we had been researching computer vision techniques for some time already – it really inspired me and made me very excited,” Huang stated on stage. “This led us to decide to focus entirely on the self-driving car field, an area in which we have now invested for over ten years; today our technology is used by almost all autonomous vehicle companies.”
NVIDIA has partnered with numerous automakers, automotive manufacturers, and tech firms developing driverless cars, including expanding its collaboration with General Motors (GM). Huang declared that the era of physical AI had arrived, extending from vehicles to manufacturing plants. NVIDIA is working alongside GM to drive transformative change in transportation.
Currently, self-driving technology developers such as Tesla, Wayve, and Waymo leverage Nvidia GPUs for artificial intelligence training within their data centers. Other companies are utilizing Nvidia’s Omniverse platform to create factory “digital twins,” enabling them to test production processes virtually or design vehicles.
Furthermore, car manufacturers like Mercedes-Benz, Volvo, Toyota, and Zoox employ the Drive Orin in-vehicle computing chip based on Ampere supercomputing architecture. Additionally, some automakers—including Toyota—are adopting DriveOS, an operating system specifically designed to enhance autonomous vehicle safety.
(Lead image source: NVIDIA GTC)
During discussions about autonomous vehicle technology, Huang referenced 'AlexNet,' which garnered significant attention in 2012 for its victory in an image recognition competition. Developed by computer scientist Alex Krizhevsky with the assistance of Ilya Sutskever (later one of OpenAI's founders) and AI researcher Geoffrey Hinton, this neural network architecture achieved a remarkable accuracy rate of 84.7% in the ImageNet academic contest.
This achievement revitalized industry interest in deep learning technology—a machine-learning approach utilizing neural networks. AlexNet’s success spurred NVIDIA to fully commit resources toward autonomous vehicle research.
“When I first saw AlexNet – we had been researching computer vision techniques for some time already – it really inspired me and made me very excited,” Huang stated on stage. “This led us to decide to focus entirely on the self-driving car field, an area in which we have now invested for over ten years; today our technology is used by almost all autonomous vehicle companies.”
NVIDIA has partnered with numerous automakers, automotive manufacturers, and tech firms developing driverless cars, including expanding its collaboration with General Motors (GM). Huang declared that the era of physical AI had arrived, extending from vehicles to manufacturing plants. NVIDIA is working alongside GM to drive transformative change in transportation.
Currently, self-driving technology developers such as Tesla, Wayve, and Waymo leverage Nvidia GPUs for artificial intelligence training within their data centers. Other companies are utilizing Nvidia’s Omniverse platform to create factory “digital twins,” enabling them to test production processes virtually or design vehicles.
Furthermore, car manufacturers like Mercedes-Benz, Volvo, Toyota, and Zoox employ the Drive Orin in-vehicle computing chip based on Ampere supercomputing architecture. Additionally, some automakers—including Toyota—are adopting DriveOS, an operating system specifically designed to enhance autonomous vehicle safety.
(Lead image source: NVIDIA GTC)