You are here

Nvidia introduces AI for self-driving cars, partners with Volvo

Volvo Car Corp. will use the new products in its public trials of autonomous vehicles in 2017, Nvidia said.

[SAN FRANCISCO] Nvidia Corp wants to get into self-driving cars.

The chipmaker, looking to expand its reach in the automotive market, introduced a new computer for vehicles that includes artificial-intelligence features to make them more autonomous.

Chief executive officer Jen-Hsun Huang introduced the Nvidia Drive PX 2: a set of processors featuring deep-learning capabilities and software that enables cars to better recognize objects around them, know where they are and plan what actions to take. He unveiled the new technology in a presentation late Monday at the International Consumer Electronics Show in Las Vegas. Volvo Car Corp. will use the new products in its public trials of autonomous vehicles in 2017, Nvidia said.

Under its founder, Nvidia is trying to expand the use of graphics chips into markets outside of personal computers, a business that's projected to decline for a fifth straight year in 2016. Still, the Santa Clara, California-based chipmaker has managed to expand sales by persuading carmakers, server companies and gamers to use its chips.

The new product is aimed at making Nvidia the center of efforts to make cars capable of piloting themselves regardless of whether their occupants can drive or not, Huang said. If those efforts are successful, particularly in complex urban environments, the impact could be far-reaching, he said.

"Self-driving cars are hard. It turns out driving is hard," Mr Huang said. "Humans are the least reliable part of the car." The new product will carry the same computing power as 150 MacBook Pros and require water cooling, the CEO said. That's the kind of horsepower required to make a car perceive its surroundings, he said.

Nvidia's focus on cars is part of a broader trend in semiconductors. As automakers increasingly try to differentiate their products based on new electronic functions and move toward autonomous vehicles, they need more chips and technical expertise fromĀ Nvidia and other technology providers.

Mr Huang has argued that graphics chips are a natural fit for the intense image-recognition computing needed for self-driving cars to navigate traffic without human help. The chipmaker, whose shares climbed 64 per cent in 2015, has already won orders for chips used to power infotainment systems in high-end vehicles such as Lamborghini and Tesla Motors Inc.