Nvidia CEO Jensen Huang’s keynote Tuesday on the firm’s GTC 2025 convention caught with custom and was chock-full of bulletins. However the firm additionally snuck in somewhat historical past lesson.
Throughout the automotive portion of his speech, Huang referred to AlexNet, a neural community structure that gained widespread consideration in 2012 when it received a pc image-recognition contest. Designed by pc scientist Alex Krizhevsky in collaboration with Ilya Sutskever (who’d go on to discovered OpenAI) and AI researcher Geoffrey Hinton, AlexNet achieved 84.7% accuracy in an instructional competitors known as ImageNET.
The breakthrough end result led to a resurgence of curiosity in deep studying, a subset of machine studying that leverages neural networks.
Seems, AlexNet spurred Nvidia to go “all in” on autonomous automobiles, the way in which Huang tells it.
“The second I noticed AlexNet — and we’ve been engaged on pc imaginative and prescient for a very long time — the second I noticed AlexNet was such an inspiring second, such an thrilling second,” he stated onstage. “It brought on us to determine to go all in on constructing self-driving vehicles. So we’ve been engaged on self-driving vehicles now for over a decade. We construct expertise that nearly each single self-driving automobile firm makes use of.”
Nvidia has notched partnerships with quite a few automakers, automotive suppliers, and tech firms creating autonomous automobiles. Its newest, an expanded collaboration with GM, was introduced this afternoon.
Automakers like Tesla and autonomous car builders Wayve and Waymo use Nvidia GPUs for information facilities. Different firms faucet Nvidia’s Omniverse product to construct “digital twins” of factories to just about take a look at manufacturing processes and design automobiles. In the meantime, Mercedes, Volvo, Toyota, and Zoox have used Nvidia’s Drive Orin pc system-on-chip, which relies on the chipmaker’s Nvidia Ampere supercomputing structure. Toyota and others are additionally using Nvidia’s safety-focused working system, DriveOS.
The upshot: Nvidia DNA is embedded within the automotive — and extra particularly, the automated driving — trade.
