Automakers from Tesla to Volkswagen have used Nvidia’s graphics chips to power infotainment systems and displays in their cars. Now the tech firm sees an opportunity to expand into another field—driver-assist technologies and even self-driving cars.
Driver-assist features include things like sounding an alarm when you’re reversing into a wall or drifting too close to a car in the next lane. Key to Nvidia’s plan is its new Tegra K1, which combines four CPU cores and 192 GPU cores on a single chip that the company says can do the work of several.
“We’re going to see a consolidation of several driver assistance processors capable of being replaced by one Tegra K1,” said Danny Shapiro, senior director of automotive at Nvidia.
We saw the Tegra K1 in action at Nvidia’s headquarters in Santa Clara, California. The chip processed data in real time from a car’s front-facing camera to detect vehicles and other objects and to recognize traffic signs and lane markings.
The K1 was also used to process data from a Lidar sensor—the rotating sensor often seen on top of self-driving cars to map the vehicle’s surroundings.
The computerization of the auto industry has been a huge boon for chip companies. Features like anti-lock braking systems and passenger air bags are all controlled by microchips, and data from sensors and the Web is creating demand for more computing power.
Nvidia, best known for its graphics cards, says its mobile processors can bring “supercomputing horsepower” to cars. Automotive has become the fastest-growing segment in its Tegra division—Nvidia says there are 5.1 million cars on the road with its processors, more than double the number in 2012.
Car computers will soon be “more powerful than any computer you will ever own in your life,” according to Shapiro. They’ll run dozens of applications—“many of which might not even exist when you buy the car, but the ability to update the car with new software, much like you do with your phone and tablet, is the way it’s going to move in the future.”