Nvidia’s vision for powering the auto industry’s technological future runs through its data centers.
Sharing a stage in Taipei in mid-October, executives from Foxconn and Nvidia announced plans to build a new class of data centers powering artificial intelligence technology they’re calling “AI factories.”
These factories will power artificial intelligence technology, and Nvidia is keen on using them to continue increasing its automotive revenue. According to Nvidia CEO Jensen Huang, “A new type of manufacturing has emerged – the production of intelligence, and the data centers that produce it are AI factories.”
Displaying a vehicle diagram, Nvidia Chief Executive Officer Jensen Huang illustrated the concept and elaborated on their capability to continually receive and process data from autonomous electric vehicles, enhancing their intelligence. “This car would, of course, go through life experience and collect more data. The data would go to the AI factory. The AI factory would improve the software and update the entire AI fleet.”
Nvidia’s automotive segment is gaining traction, closing in on nearly $1 billion in annual revenue. Over the past decade, its auto revenue has grown by more than 812%. The company’s share price has more than tripled in the past year, driven by investors’ rush to get in on the AI boom.
The company’s fiscal year ending January 29, 2023, demonstrated robust financial performance, recording an annual revenue of $26.9 billion and gross profits totaling $15.4 billion. This marked an impressive 61.8% increase in annual revenue and a 47.7% rise in gross profits compared to 2021. Wall Street analysts expect Nvidia to finish the current fiscal year (ending in January 2024) with earnings of $10.43 per share, a potential increase of over 212% from the year prior.
This notable boost was primarily attributed to the successful sales of their highly sought-after H100 chip, an advanced chipset pivotal in the early AI surge, retailing for as much as $46,000 each. As of the last market close on Friday, the company’s shares were valued at $492.98.
Now, through its enhanced partnership with Foxconn, Nvidia is intensifying its artificial intelligence chip production for autonomous vehicles, making a calculated bet on the future of self-driving cars in a potential $300 billion EV market.
However, some critics and lawmakers remain skeptical of the technology’s readiness for the open road. Despite advancements, concerns linger regarding the reliability and safety of autonomous driving vehicles, particularly in complex and dynamic real-world scenarios.
Nvidia has maintained its stronghold in the artificial intelligence chip-making sector since the introduction of its renowned H100 chip early last year, catalyzing the AI boom as companies rushed to implement the technology across various industries.
The demand has bolstered Nvidia’s revenue significantly, turning it into a trillion-dollar corporation. However, the company’s expanded collaboration with Foxconn, a company best known for manufacturing Apple products, is a strategy to unlock fresh revenue streams in an AI market it believes is on the cusp of massive growth—autonomous vehicles.
While Nvidia first began developing autonomous driving vehicle partnerships in 2021, significant progress in the technology has been made since then. Late last year, the Nvidia Drive operating system received safety certification from TÜV SÜD, one of the most experienced and rigorous assessment bodies in the automotive industry.
More recently, Nvidia unveiled its ‘Drive Thor’ platform, scheduled for production in 2025. These all-in-one chips streamline car functions by combining digital entertainment, advanced driver assistance systems, virtual cockpit, driver monitoring, and more into a single chip. This integration simplifies car design, cuts costs by eliminating the need for multiple chips, and allows manufacturers to allocate the chip’s 2,000 teraflops of processing power for autonomous driving.
While Nvidia has chip auto partnerships with other vehicle manufacturers such as Ford, Toyota, and Mercedes, only Foxconn will be implementing this more advanced chip platform in its production, at least for the time being.
To be sure, Nvidia’s ability to find buyers for its AI chips relies on light U.S. export controls. The same day Nvidia announced its enhanced partnership with Foxconn, U.S. legislators tightened an existing Chinese chip export ban to include its more advanced chips. That said, it’s unlikely the new data centers will be built in China, where Foxconn does much of its manufacturing and employs hundreds of thousands of workers.
Despite this, one bright spot working to Nvidia’s advantage in developing this market is its limited competition. Tesla stands as Nvidia’s primary autonomous driving competitor. Unlike the new vehicles being developed in partnership with Nvidia, Tesla’s vehicles use mostly cameras to assist their vehicle’s AI. In contrast, Nvidia uses what many consider to be a more advanced LIDAR system.
LIDAR, or Light Detection and Ranging, is a remote sensing technique using pulsed lasers for precise imaging and is especially resilient in harsh lighting conditions compared with camera systems used by Tesla. Still, some critics argue its effectiveness is limited to well-trafficked areas, with performance dropping on remote or rural roads.
But more trouble is seen in most individual consumers’ current lukewarm interest in autonomous vehicles and their limited performance capabilities so far.
“I think personal usage is not going to work for a very, very long time,” says Akshay Singh, a partner at PwC Strategy specializing in the automotive industry. “Most of the companies are targeting the Robo-taxi market, which they think could potentially be a very big market.”
Cruise, a General Motors company, suspended driverless vehicle operations nationwide in October, two days after a California state regulator suspended its license to operate. It is one of the few robo-taxi companies currently in existence, and the suspension came after individuals and emergency services made complaints and reported several safety incidents.
Reached for comment, GM referred to a statement they made after Cruis’s suspension that its “commitment to Cruise with the goal of commercialization remains steadfast.” Safety “has to be our top priority, and we fully support the actions that Cruise leadership is taking to ensure that it is putting safety first and building trust and credibility with government partners, regulators, and the broader community.” Cruise did not reply to a request for comment.
“You know, most of the vehicles cannot operate at highway speeds right and not [sic] really work very well in complex intersections,” says Singh, acknowledging the complaints. “There’s a lot more development as needed before it’s going to become a reality, before the widespread adoption.”
Still, the lack of interest in these vehicles could largely stem from the fact that they have only been used in select areas and cities around the world, making it harder for them to catch on with consumers.
Ultimately, the crux of the matter revolves around whether the long-term promise of self-driving cars can soothe the concerns of Nvidia investors, who are currently preoccupied with the immediate challenge presented by stricter U.S. regulations on chip exports to China. Unfortunately, early indications do not offer much reassurance.
“The success of fully autonomous vehicles hinges mostly on AI performance, and the technology is still not up to scratch,” says Pedro Pacheco, vice president of research, automotive and smart mobility at Gartner. “Making a car driving autonomously on [a] straight strip of tarmac is easy. The really hard part is when it needs to tackle the most challenging cities in the world and do it fully autonomously, in full safety, and at a great return on investment for operators.”
Still, others seem to think the increased foray into autonomous driving vehicles is just the beginning of Nvidia’s AI expansion, and more is yet to come. “There’s never been an everything cycle like this. There’s been a half a cycle, there’s been an internet cycle, there’s been a 4G cycle; not an everything cycle,” says Rosenblatt Securities analyst Hans Mosesmann. “Nvidia is going to be playing at this thing in virtually every single market they can think of.”
Nvidia and Foxconn did not reply to requests for comment on this story.