A developer kit for Nvidia’s self-driving car platform, the Drive PX, will go on sale in May for $10,000, Nvidia CEO Jen-Hsun Huang said today at the company’s GPU Technology Conference.
The platform is powered by two of Nvidia’s Tegra X1 chips and was first teased at CES earlier this year. Huang claimed it’s 3,000 times faster than DAVE, the autonomous vehicle technology developed by DARPA.
The notion is that with powerful enough hardware, self-driving vehicles will be better able to recognize what they’re seeing, learn from the environment and make the right decisions. Nvidia hopes to expand its existing partnerships with automakers like Tesla, Audi and BMW.
“It’s a system that can be trained, and retrained, with more data,” Senior Automotive Director Danny Shapiro wrote in a blog post. “Every time your self-driving car gets an over-the-air update, it can get smarter.”
“Just like Mars,” Huang remarked, a nod to Musk’s desired place of death.
After Musk took the stage, Huang asked him about a 2014 tweet that likened the dangers of artificial intelligence to nuclear weapons. Nearly all of Nvidia’s keynote touched in various forms on AI and deep learning.
“I said ‘potentially,'” Musk qualified. “We don’t really need to worry about autonomous cars. It’s a narrow form of AI.”
Musk said having the right hardware should come first in designing autonomous vehicles, but that change may come slowly. There are currently two billion cars and trucks on the road worldwide, and global manufacturing capacity is at 100 million vehicles per year. “Legacy” vehicles, he noted, will be around for a while.
And, Huang asked, what about security? Musk said the absence of steering wheels and brake pedals, “many years from now,” creates new challenges. At Tesla, his team is currently focused on “making it very difficult to do a multi-car hack.”
In the keynote, Nvidia also announced a new GPU, the Titan X, and a “supercomputing box” for researchers called the Digits Devbox, which will also debut in May for $15,000.
This article originally appeared on Recode.net.