Nvidia CEO Sees Massive AI Needs for Self-Driving Cars
This article first appeared in Data Sheet, Fortune’s daily newsletter on the top tech news. Sign up here.
Greetings from rainy Las Vegas, where last night I interviewed Nvidia CEO Jensen Huang, CEO of the 25-year overnight success chipmaker worth $134 billion.
Huang’s and Nvidia’s story is an epic tale of a very good market turning into a market for the ages. Nvidia for years made something called graphical processor units, chips that solved “massive computational problems” that made video games sing. The central processing units Intel (INTC) sold powered a generation of PCs, while Nvidia (NVDA) occupied a lesser-followed but still lucrative niche.
A niche no more, Nvidia’s market has morphed into supplying the brains for artificial intelligence. The same smarts that create a “virtual reality” for gamers now envision what the world looks like for all sorts of industrial applications. Chief among them is self-driving cars, and Nvidia announced partnerships this week with Uber and Volkswagen.
Huang, the CEO for Nvidia’s entire existence, is enjoying his moment. I asked why Nvidia hadn’t experienced the same kind of security concerns Intel had, and he scoffed. Asked if today’s automobiles could be retrofitted with AI-worthy chips, Huang essentially said: Not so much.
Huang painted a nuanced view of what autonomous vehicles will look like. Cars driving fixed routes in “geo-fenced” area will come first, followed by over-the-road truckers. True self-driving cars will be further off, and even once they come people will still drive for “entertainment,” he said.
The Nvidia CEO’s most fascinating observation is that his chips will power software that will write software. In other words, AI will beget AI to solve problems humans didn’t know they had.