Text size
Courtesy of Mercedes-Benz.
If you’ve ever spent hours pushing buttons on a video game controller, I’d like to thank you for your service. I’m starting to think you helped invent driverless cars. You may have also participated in cancer treatments.
See Tuesday, Nvidia (ticker: NVDA), a powerhouse of video game chips, announced a partnership with Mercedes-Benz, the world’s largest luxury car maker, to redesign its entire fleet around a computer network defined by software with over-the-air updates and recurring revenue for applications. “This is the time for the automotive industry iPhone,” Nvidia CEO Jensen Huang told me.
For investors who have followed Nvidia’s transformation in the past five years, this makes perfect sense. It helps explain why the stock has multiplied 18 times in price during that period.
Huang co-founded Nvidia in 1993 to make a new type of computer chip ideal for powering three-dimensional video games, a few years before games needed such a thing. The most elegant games soon followed. This fiscal year, which ends in January, the games will have an estimated value of $ 6.1 billion in revenue for Nvidia. But there is much more to the company than that.
“People thought we were a video game company,” says Huang. “But we are an accelerated computing company where video games were our first killer application.”
To see the second killer app, just look at Wall Street’s forecasts for this year. Revenue in a category that Nvidia calls the Data Center is expected to double, to $ 6.5 billion. That will mean that games are no longer the company’s biggest money generator.
Nvidia owes its success in data centers to the increasing use of artificial intelligence, because video game chips are very suitable for this type of computing. There is a perfectly good explanation for why this is the case: I know, because sometimes people try to guide me through it, as I look at them like a dog looking at Cirque du Soleil.
The closest I came to understanding it was last week, speaking to Chris Rolland, analyst at Susquehanna Financial Group. He described how illuminating a pixel on the screen involves three “vectors”: an x and y coordinate, plus a color. Artificial intelligence also uses three vectors, he said, in an operation called a multiple accumulation matrix. Soon after, my brain felt warm, so I started thinking about wagging apples, which helped me.
Regardless, Rolland says many companies are trying to make the chips even better suited to artificial intelligence, but that Nvidia has created powerful software to program its chips, called CUDA, so its role is secure for now. . Nvidia leads in what is called parallel processing, or performs many tasks at the same time, while Intel (INTC) governs in serial processing, or performs individual tasks very quickly.
Right now, there is massive growth in parallel processing. Jefferies analyst Mark Lipacis calls it the fourth tectonic shift in computing, after the shift from mainframes to minicomputers in the 1960s, to personal computers in the 1980s, and to cell phones and data centers from the late 1990s. He attributes the rise in parallel processing to cheap memory, virtually unlimited data storage, and, yes, improvements in chip and software processing. In other words, Nvidia’s Huang made the product, and the market popped up for it, just like with video games.
Artificial intelligence is increasingly useful for turning raw customer data into sales tips, or automatically indicating which photos in a collection include Uncle Burt, or detecting cancer in a patient scan that a doctor might miss . It’s also what cars will one day use to drive. If someone asks how it works, tell them that there is a subset of AI called machine learning, and a subset of that called deep learning, which uses something called artificial neural networks. Next, pretend you have to answer a NASA emergency call to help with a very difficult space question.
According to Rolland in Susquehanna, Tesla (TSLA) leads autonomous driving, in part because it has developed its own hardware and software. Other automakers must spend large sums and spend on artificial intelligence and chip manufacturing, or find a partner if they don’t want to risk falling behind. Rolland says a company called Mobileye is in a good position to be one of those partners. Mobileye, however, was captured by Intel in 2017. The advantage of Nvidia, he says, is that its chips are already being used for the job of training autonomous driving algorithms.
Huang at Nvidia calls the Mercedes deal a transformative moment for the company, because it has moved from video game hardware, to AI hardware and software, and now to automotive hardware, software, and services. He says that makes Nvidia a platform company, not just a chipmaker. “The first vertical market we choose is autonomous vehicles because the scale is very good,” he says. “And the life of the car is so long that if it offers new capabilities to each new owner, the economy could be quite wonderful.”
Nvidia’s earnings per share is expected to increase 40% to $ 8.15 this fiscal year. The shares were recently trading at $ 370, or 45 times earnings. Data centers could be a $ 10 billion business for the company within three years. Automobiles are still less than $ 1 billion.
How long until the cars really drive themselves? I would have guessed more than a decade, but Rolland in Susquehanna says it could be 2025.
Huang says, “We are going to have humans on the circuit for a long time.” I asked him if it is true that cars will have to be programmed to prioritize who to hit and avoid in the event of an accident. A top chip boss told me that over dinner several years ago, and I’ve been trying since then to figure out how to get on the “avoid” list.
Huang says it is not true. “If a car gets in front of you, stop,” he says. Otherwise, stay in your lane. Simple.”
Write to Jack Hough at [email protected]. Follow him on Twitter and subscribe to his Barron’s Streetwise podcast.
.
