NVIDIA Corp. is chasing autonomous cars to drive revenue as the company tries to break into new markets and shift from its decadeslong reliance on computer graphics.
The chipmaker on March 18 said it struck a deal to widen its partnership with Toyota Motor Corp. to design and validate self-driving car technology. The companies will also focus on building artificial intelligence technology, an essential element needed for the navigation of self-driving cars. The partnership is focused on simulating autonomous car situations and developing the technology based on the results. The agreement could lead to Nvidia's technology being used across many Toyota vehicle models and types.
 Autonomous car with Nvidia tech |
"For Nvidia, this will broaden their appeal to other [original equipment manufacturers] that have broader portfolios, while for [Toyota], they can utilize the Nvidia platform, which is arguably ahead of their competitors in terms of software and research from years of work with other companies," said Nina Turner, a research manager on IDC's Enabling Technologies and Semiconductor group, which focuses on new semiconductor markets like automotive. IDC is a research and consulting services provider for the technology market.
The chipmaker is not building a self-driving car, but building the underlying circuitry, design capabilities and intelligence necessary for the vehicles. The core technology will revolve around its graphics chips, which are also built for use in gaming and supercomputing.
Nvidia currently is providing automotive technology to several carmakers. AB Volvo will deploy the first Level 2+ autonomous cars based on Nvidia technology in the early 2020s. Daimler AG's Mercedes-Benz brand will also use Nvidia technology for its next-generation autonomous vehicle. Other Nvidia customers include Volkswagen AG, Bayerische Motoren Werke AG and AUDI AG.
Enabling autonomous cars is a big computing challenge, and the partnership with Toyota will help make driving safer, said Nvidia CEO Jen-Hsun Huang at a keynote at the GPU Technology Conference on March 18.
"This is so exciting. This is how we can make a difference in the future of transportation," Huang said.
Nvidia said in its most recent earnings release that automotive revenue in the fourth quarter of 2018 was $163 million, up 23% from the same quarter a year ago. That was a small chunk of the $2.21 billion overall quarterly revenue, which declined by 24% year-over-year on graphics processor weakness. Annual automotive revenue was $641 million, growing 15%.
As cars move toward electrification and partial autonomy, the revenue opportunities will grow, analysts said. Yet questions remain about how quickly the industry will see the promise of autonomous cars turn into a real opportunity.
"It is more Toyota signing on to Nvidia's notion that simulations can effectively replicate or even replace the real world testing that other autonomous car pioneers are pursuing," said Charles King, principal analyst at technology research firm Pund-IT. "Only time will tell whether Toyota has made a good bet."
Nevertheless, Nvidia is in a strong position to cash in when autonomous cars become a standard, King said.
Nvidia's main competitor is Intel Corp., which is trying to break away from its reliance on personal computers by expanding data-center revenue and into new areas like automotive and AI. The company acquired Mobileye N.V., a provider of autonomous driving tech, for $15.42 billion in 2017, and in 2018 announced 28 design wins and launches of 78 vehicle models with various levels of autonomy. The company did not break down specific revenue numbers but has previously said the autonomous car industry is a $7 trillion opportunity. Like Nvidia, Intel provides AI chips and computing capabilities for autonomous cars.
Qualcomm Inc. also competes in the market, and like Intel, considers 5G, the next generation of mobile telecommunications, a big element in the success of autonomous cars. Both of the companies are making 5G baseband chips for cars, which enable 5G radio communications over spectrum, an area where Nvidia is lacking.