Nvidia AI to Boost Development of Autonomous Vehicle Tech
Autonomous vehicles (AVs) are emerging as a use case for artificial intelligence, with Nvidia announcing partnerships with carmakers to develop self-driving technology using its hardware and software platform.
“The AV revolution has arrived after so many years, with Waymo’s success and Tesla’s success,” said Nvidia CEO Jensen Huang during a keynote speech at CES 2025 in Las Vegas Monday (Jan. 6).
Nvidia and Uber announced in a Monday press release that they will jointly develop AI-powered self-driving technology. Uber will use data from millions of trips taken in its vehicles and pair it with Nvidia’s new generative world foundation model, Cosmos. This enables Uber’s AI systems to train in virtual-world settings.
“Generative AI will power the future of mobility, requiring both rich data and very powerful compute,” said Uber CEO Dara Khosrowshahi in the announcement. “By working with Nvidia, we are confident that we can help supercharge the timeline for safe and scalable autonomous driving solutions for the industry.”
Toyota will build its next-generation vehicles on Nvidia’s Drive AGX platform to add advanced driving assistance capabilities, while Aurora and Continental plan to deploy driverless trucks at scale, according to a separate Monday press release. Aurora is developing self-driving systems for commercial vehicles like trucks, while Continental provides advanced trucking technologies and components. Other clients include Mercedes-Benz, Rivian, Volvo, Amazon’s Zoox, BYD and more.
The autonomous vehicle market “is going to be very large,” Huang said during the speech, predicting it will become the “first multitrillion-dollar robotics industry.”
Separately, Amazon Web Services (AWS) said in a Monday press release that it is collaborating with Here Technologies to use its mapping solutions to develop advanced driver assistance systems, automated driving and new digital car experiences.
The Holy Grail
Fully autonomous vehicles have long been the holy grail of companies like Tesla and Waymo. Despite the investment of billions of dollars, true self-driving technology has remained elusive and most recently claimed General Motors’ Cruise, which was its robotaxi division.
However, the technology is once again attracting venture capital dollars, which could signal a shift this year.
The reason why fully self-driving systems are difficult to develop is the nearly limitless scenarios an autonomous car will encounter on the road that it can be trained to anticipate. For instance, partially obscured stop signs or signs placed in unexpected areas can confuse a car. These are things that even an inexperienced human driver can easily handle.
There are five levels of autonomous driving, with level 5 being a vehicle that can drive itself in all conditions without human input, according to the U.S. Environmental Protection Agency. Levels 1 to 4 have varying degrees of human intervention. Tesla, despite having a full self-driving (FSD) feature, has not reached level 5.
Nvidia’s Drive AGX platform helps carmakers reach different levels of autonomy, from advanced driver assistance systems to fully driverless. Drive AGX is part of Hyperion, Nvidia’s broader reference architecture specifically built for autonomous vehicles. Huang said during the keynote that each car manufacturer will decide on the extent of their collaboration with the chipmaker.
Huang called Nvidia’s approach the “three computers,” referring to the training, simulation and in-car computer systems. The company’s most advanced chip architecture, Blackwell, is coming to autonomous cars. The Nvidia Drive Thor chip, the successor to Drive Orin, is built to handle intense processing workloads — like the billions of data coming from cars with sensors and cameras.
Regulatory Tailwind?
On the regulatory front, a group that represents autonomous car companies is urging the U.S. government to do more to accelerate adoption — or else cede the lead to China.
“The federal government is the one that needs to lead when it comes to vehicle design, construction and performance, and we just have not seen enough action out of the federal government in recent years,” Autonomous Vehicle Industry Association CEO Jeff Farrah said, according to a Tuesday (Jan. 7) Reuters report.
Members of the association include Waymo, Uber, Zoox and others.
Autonomous vehicle deployment has been hampered by high-profile crashes, including an October 2023 accident when a Cruise vehicle dragged a pedestrian 20 feet after she was hit by another car, the report said. The U.S. Department of Transportation has opened investigations into Cruise, Waymo and Zoox.
However, proponents have long argued that autonomous vehicles are much safer than human drivers, who can be distracted or impaired.
The trade group wants federal legislation to clearly state that human controls are not necessary for autonomous vehicles to meet performance standards and to allow car companies to disable an AV’s manual controls, per the report. It also called on the U.S. government to create a national AV safety data repository accessible by state transportation authorities.
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.
The post Nvidia AI to Boost Development of Autonomous Vehicle Tech appeared first on PYMNTS.com.