Big chipmakers go all in on robot cars
Santa Clara — Three major computer-chip manufacturers are moving into new territory by investing heavily to become the indispensable drivers of the driverless-car game.
Intel is best known for cranking up the speed of laptop and desktop computers. Nvidia supercharged graphics for video games. Qualcomm helped make smartphones truly smart. Today, all are vying to become major players as the auto and tech industries work to bring autonomous vehicles to the masses.
Santa Clara-based Nvidia took center stage last week in announcing a partnership with Toyota that puts its artificial intelligence hardware and software behind the automaker’s autonomous-car efforts. A week earlier, Intel unveiled its Innovation Center in San Jose and showcased its partnership with Delphi to create a turnkey driverless system that can be integrated into vehicle designs by any manufacturer.
And later this year, San Diego-based Qualcomm could win approval of its $47 billion bid to acquire semiconductor maker NXP — one of the top suppliers to the auto industry.
Intel, Nvidia and Qualcomm are no longer companies willing to simply see their chips used by others. They are neck-deep in the game and calling many of the shots.
“Right now, there really is a rush among silicon manufacturers ... to try and have the chip set that is, in some way, the core of the future automated vehicle,” said Chris Gerdes, who runs Stanford University’s Dynamic Design Lab.
The quest is all the more important in that many of the chipmakers face declines in their traditional revenue sources. At Qualcomm, which has been hurt by losses in cellphone sales, the hope is to be the chip provider to “cover all domains” of the driverless car.
“Everybody in the chip business has an automotive strategy now,” said Kevin Krewell, a principal analyst at tech-consulting firm Tirias Research. “They all see the increased amount of chips and silicon going into the automotive sector as one of the big growth-markets right now.”
But Krewell said it’s unlikely any one chip can dominate, despite the massive investments companies are making.
“The auto industry is just too big and too diverse,” he said.
Nvidia is best known in the videogaming community. After nearly a quarter-century in business, its graphics-processing units have evolved to crank out visuals that are more and more lifelike.
The demands of the videogame market may have set the company on an advantageous path as it moves full-force into driverless vehicles. Nvidia specializes in graphics processors, not the central processing units backed by Intel and Qualcomm.
Tim Wong, a member of the company’s technical marketing, said the fundamental difference that gives Nvidia its edge is that traditional central processors have only four to eight cores, or processing units. “Whereas, you have thousands of cores in an Nvidia GPU,” he said.
Much of the excitement around driverless cars centers on artificial intelligence, or teaching machines to learn. For autonomous vehicles, that means coaching a system to recognize objects and situations, and then respond accordingly. Nvidia officials contend this is similar to what its graphic processors have been doing all along in the gaming industry.
That work has already paid off in the form of high-visibility partnerships with Tesla Inc. and Volvo Cars Ltd., and last week, its announcement of the tie-up with Toyota.
Nvidia officials said they see plenty more action on the horizon.
“Artificial intelligence is keyed to be the next big thing in terms of the same kind of impact that the Industrial Revolution had,” said Nvidia’s Sean Wix, another member of Nvidia’s technical marketing team. “It will leave no market undisturbed, really.”
When wireless communications exploded in the 1990s and early 2000s, Qualcomm mastered technology that allowed multiple signals to share band space. It put the company and its chip sets into all kinds of telecommunication technologies.
That included the smartphone arena, where the company’s Snapdragon mobile processing chips boosted the company’s profile and portfolio. But waning demand for smartphones has Qualcomm looking for new areas of expansion.
Qualcomm is not new to the auto industry, having worked with General Motors on OnStar applications as early as 2002. The company’s wireless technologies have led to partnerships with “pretty much every automaker on the planet,” says Nakul Duggal, Qualcomm’s vice president of product management. Snapdragon processors made their way into the automotive field in 2014.
The company has made a series of acquisitions touching on wireless capabilities in the auto industry, including the $3 billion purchase of wireless networking company Atheros in 2011.
Qualcomm’s ability to push even further into the driverless-vehicle race will be boosted exponentially if the company’s bid for NXP is finalized by the end of the year, as company officials expect. Roughly 40 percent of NXP’s semiconductor sales are linked to the auto industry, where they are used for everything from vehicle connectivity, to audio systems and regulating engine functions.
Like Qualcomm, Intel’s foray into driverless cars stems in part from the search to supplement its declining traditional markets — in this case, laptop and desktop computers. Intel may have the most clearly defined game plan — at least publicly — of the three U.S.-based chipmakers.
Intel and partners like Delphi Automotive PLC, BMW AG and Ericsson AB are putting together what they describe as a turnkey driverless car system that can also help capitalize on Big Data — the information that connected vehicles can collect, sort, package and send to the “cloud.” That information can eventually be sold to a variety of customers in a business that could be worth $750 billion by 2030.
Intel’s decades of experience and its comprehensive approach to driverless and connected cars are what separate it from competitors, say company officials.
“Obviously, Qualcomm is focused on wireless connectivity — the wireless space — and they have other general computing (capabilities),” said Doug Davis, senior vice president and general manager of Intel’s Automated Driving Group. “Nvidia provides their GPU — a specialized computing element that can do things like sensor fusion and workload acceleration.”
He says his company tries looking at solutions more broadly: “We engineer our technologies to make sure they deliver all of the capabilities that are necessary for the automotive industry today.”