The automotive industry is beginning one of the most significant transformations in its history.
"The evolution of the self-driving vehicle from automated to autonomous vehicles is already under way, and significant investments are being made by individual companies and the entire automotive industry to accelerate the pace of innovation and do actual prototyping on public roads," Thilo Koslowski vice president and distinguished analyst at Gartner, said in October.
Most of the glitz and glamour of self-driving machines goes to the vehicles themselves, such as Tesla's (NASDAQ:TSLA)recently announced Model S semi-autonomous system, and Audi's "piloted driving" system on the 560-horsepower RS7 research car.
But under the hood (or sometimes in the trunk) are advanced computer systems that string together algorithms and machine learning with cameras, LIDAR, GPS, and a host of sensors. This is where the real magic behind autonomous cars is made -- and potentially the big profits as well.
Mobile-chip makers NVIDIA (NASDAQ:NVDA) and Qualcomm (NASDAQ:QCOM) understand this opportunity. Each company is going after autos in its own unique way, even as they take on more established automotive chipmakers like Freescale Semiconductor and STMicroelectronics.
Slow and steady
Qualcomm is one of most dominant chipmakers for mobile devices, and last week the company showed off some its processor prowess at the international Consumer Electronics Show on a concept Cadillac XTS.
The company tapped its Snapdragon processor -- running the Android OS -- to run head-tracking software that monitors a driver's alertness level, with a lane detection warning system, in addition to a gesture-controlled infotainment panel.
To date, Qualcomm's automotive focus has been primarily on infotainment systems and technology that helps a car see what's on the road around it.
But the company is also creating ways for vehicles to communicate with other vehicles (V2V) and the surrounding infrastructure (V2I). That's an important part of automotive autonomy, as cars will need to talk with each other in order avoid accidents. The National Highway Traffic Safety Administration has already set draft standards for V2V and will formalize them by 2020. The company is in a unique position to tackle V2V and V2I because of its dominance in mobile wireless connections.
But even with all of Qualcomm's automotive ambitions, it doesn't come close to what NVIDIA is doing right now.
Life in the fast lane
While NVIDIA is well known for it graphics processors and gaming, the company has been building its automotive technology for some time, and its Tegra chip powers the infotainment system in Tesla's Model S.
Last week. the company highlighted three automotive areas it is focusing on: processors, infotainment platforms, and autopilot technology.
NVIDIA's newest processor is the Tegra X1, which boasts a 256-core GPU and an eight-core CPU. The company says Tegra X1 is as powerful as supercomputers from 15 years ago. NVIDIA is using the processor to power both its Drive CX infotainment platform and its Drive PX autonomous driving tech.
While Drive CX is impressive in its own right, the Drive PX autopilot platform could be the company's true automotive silver bullet. Powered by two Tegra X1 chips, Drive PX uses deep learning and situational awareness to autonomously drive vehicles. The system pairs with up to 12 high-resolution cameras to see and assess its surroundings for driving and parking itself, and can even pick out different types of vehicles on the road (like knowing the difference between an ambulance and a delivery truck).
NVIDIA's Drive PX is much more than a pipe dream. Audi, which has produced some of the most advanced autonomous vehicles to date, last week drove a piloted driving A7 550 miles from San Francisco to Las Vegas with minimal human input -- powered in part by NVIDIA.
As of fiscal third-quarter 2015, NVIDIA reported that more than 6 million vehicles on the road use infotainment systems powered by the company, and revenue from its Tegra chips jumped 51% year over year to $168 million. The company said last year the automotive segment is the fastest-growing part of its Tegra chip sales.
I think Qualcomm will eventually become a bigger automotive tech player -- as the company's mobile industry experience and technology can certainly push it to front of the pack quickly -- but for now NVIDIA is doing much more in the automotive space. The company is already creating advanced systems on the cutting edge of automotive autonomy, and it shows no signs of slowing down.