Barron's Streetwise - Nvidia Takes on Tesla

Episode Date: June 26, 2020

Nvidia CEO Jensen Huang on how the chipmaker has parlayed video game success into supercomputing dominance. Next up: Autonomous driving. Learn more about your ad choices. Visit megaphone.fm/adchoices

With record levels of dry powder available for investment, find out what's in store for private markets in 2025 and beyond. Listen to Crafting Capital in partnership with UBS at partners.wsj.com slash UBS, Spotify and Apple Podcasts. We now have multiple growth drivers, video games, high-performance computing, artificial intelligence in the cloud, our data center business is doing great. And then today we announced what I believe is potentially one of the largest business opportunities in our company's history. Welcome to the Barron Streetwise podcast. I'm Jack Howe. The voice you just heard,

that's Jensen Wong. He's the CEO of NVIDIA. And when he says today, he means the other day when he and I spoke. I think a lot of people who know about NVIDIA think of it primarily as a maker of chips for video gaming. But increasingly, it's an artificial intelligence company. And now it's taking on Tesla for the future of autonomous driving. Listening in, keeping me on track is our audio producer Meta. Hi Meta. Hi. Let's talk about video games. Do you play them and have you ever been totally immersed in one from start to finish? I played them a little bit when I was a kid, but it wasn't something I was very much into.

I am not the biggest video gamer, but about once every 10 years or so, I get into a game. And I'm always shocked by how far the graphics have come. I think I played my first games back around 1980 when I was in middle school. My brother and I had some epic battles playing something called Combat on the Atari 2600. There were tanks that looked like blocks, and you shot one it would fly off the edge of the screen and then reappear on the other side of the screen spinning around. A decade later I was playing Ghouls and Ghosts on my college roommate's Sega Genesis. And that was what you call a side scroller. The background changed as the knight named Arthur hurdled tombstones and hacked at zombies.

Ten years after that, around 2000, for the first and only time in my life, I assembled a home computer, which is a lot less complicated than it sounds. I basically ordered a case and some components and followed some online instructions. The main processor came from Intel, and I also learned that if I wanted to play any of the new video games at the time, I'd have to have a separate graphics processor. There was a company called NVIDIA that made those, and I got one. I ended up playing a shoot-em-up game called Unreal Tournament, where the players ran around in three-dimensional virtual worlds. Now, it turns out that in 1993, closer to my ghouls and ghosts side-scrolling period than my Unreal Tournament 3D days, an electrical engineer named Jensen Wong had co-founded NVIDIA.

Hi, Jensen. Thanks for making a few minutes to talk with me. Hi, Jack. Jensen explained how NVIDIA back in the early 90s began designing chips for modern video games before modern video games really existed. Video games were 2D and so we had to go and build a 3D ecosystem, develop tools, evangelize it, and then eventually one day, you you know a young genius came along named John Carmack who put Quake on the PC and it changed everything. I know Quake is a landmark game one of the first real 3D video games I don't remember it it was a non-video game year for me might have been busy reading classic literature I mean probably not but maybe. Video games today are still 3D, but the detail is incredible. You can see

every wrinkle on players' faces, and it's an enormous market for NVIDIA. Wall Street expects the company's gaming unit to bring in more than $6 billion in revenue this fiscal year. But if you look closely, NVIDIA has gotten into something much bigger than gaming. There's a business unit labeled Data Center that's expected to jump from just under $3 billion last year to around $6.5 billion this year, passing gaming. NVIDIA attributes that to the growth of cloud computing and artificial intelligence. But how is it that a gaming company is growing so fast in those other categories? I asked Jensen. It's just two sides of the same coin. People thought we were a video game company, but we are an accelerated computing company where video games

was our first killer app. Jensen says the search for applications that need vast computing power took the company to scientific computing and to supercomputers and supercomputer data centers. It also led to other opportunities. Our work there led us to researchers all around the world. And one day, deep learning found us and we found deep learning. And the combination of this algorithm and this processor that we were making that sped up that algorithm from months to days, it made it possible for this approach to even be viable. We kind of created the big bang of artificial intelligence. And that artificial intelligence early days allowed us to think about, hold on a second, what can we do with these skills? I feel like I need to do some deep learning about artificial

intelligence, but let's come back to that because I want to get to NVIDIA's announcement this past week about cars. It's partnering with Mercedes-Benz to provide chip software and more to make cars drive themselves. This is the autonomous vehicle computer and all of the software and the infrastructure that goes along with them. And so the first thing is that we announced a partnership that has never been quite this broad and deep before. This relates to every single car of Mercedes-Benz starting in 2024. We're going to build one architecture that spans the entire fleet. We are going to develop the AI applications, the autonomous vehicle applications, and then we're going to offer it to their customers

and share the economics. Tesla already makes autonomous driving systems for its own cars. NVIDIA will sell an autonomous platform that other car makers can use if they don't want to start the transition to autonomous driving from scratch. The Mercedes deal doesn't stop NVIDIA from signing on other customers. The deal also allows NVIDIA to roll out new features for years as software upgrades with recurring revenue, much like Apple does with its iPhone and App Store. If you download new apps for your phone, you get a better phone without changing your hardware, and Apple gets a cut. Or as Jensen puts it, This is the iPhone moment of the car industry. Jensen says Mercedes and NVIDIA will work together on design. What we're going to do going

forward is use one architecture across the entire fleet. And not only that, put plenty of extra horsepower inside the car so that we could sell applications into that car and for that customer for the entire life of the car. Horsepower in this case refers to computing power. Jensen sees this as a transformative moment because NVIDIA has gone from hardware for video games to hardware and software for data centers to hardware, software and services, an entire platform really for autonomous driving. is autonomous vehicles because the scale is so great. The opportunity to bring automation to it for safety and benefits and comfort is so great. And the life of the car is so long that if you have the benefit to offer new capabilities to each new owner of the car or the existing owner over the life of the car,

the economics could be quite wonderful. I should note that Barron's magazine named Jensen one of its top CEOs for 2020, based in part on his efforts for workers and customers during the pandemic. You can read more about that in this week's issue of Barron's. But the boss is going to be so proud of me for plugging the magazine like that. Are we going to get a raise? Probably you will. That's fine. I wanted to hear from someone outside of NVIDIA about how the company's technology stacks up. And since I barely know the difference between deep learning and deep dish pizza,

I wanted someone who could connect a few more of the dots for me between video games and artificial intelligence. Hi, Christopher. How are you? Doing all right. You go by Chris or Christopher? Chris is fine. I called Chris Rowland, a semiconductor analyst at Susquehanna Financial Group, and he explained the whole thing to me.

There was something about vectors. Video cards use three of them to light up pixels for video games. And it started to come together for me when he described how artificial intelligence also uses three vectors. When you're looking at artificial intelligence, what you're also doing is multiplying three vectors together. It's called matrix multiply accumulate. And that's the magic to AI. Matt, that reminds me, I have to rewatch The Matrix before the end of June because it's leaving Netflix. I can't remember anything that happens in the movie. Did you see it? Oh, I loved it. I even bought like a long leather jacket after watching it. Oh my, what?

Like a long leather jacket after washing it. Oh my, what? Wait a second. Slow down. After watching The Matrix, you bought a long leather jacket. Yeah, I think it was pleather, but yeah. Did you also begin learning any martial arts? I don't remember that, no. And I didn't have the sunglasses either.

Oh, I love it. More to the point, Chris says chips for video games just happen to be close to perfect for AI, and that a lot of companies are working on even better chips, but that NVIDIA has developed a key advantage. It was very difficult to program these GPUs. advantage. It was very difficult to program these GPUs. NVIDIA realized that people were still doing it anyway and thought that they would make this programming easier. And so they created a language called CUDA, C-U-D-A. And CUDA is the best language for programming artificial intelligence today. By the way, if you're not sure what artificial intelligence does, Chris has some examples. Think about recommender engines like Netflix telling you what video you should watch next.

Or think about conversational AI or chatbots, you know, where you're not even sure if you're actually talking with a customer service representative or an AI bot. All of those are powered by NVIDIA today. Okay, chatbots. What else? And when we talk about autonomous driving, make no mistake about it, we're talking about artificial intelligence. That's what it is. Now we're talking. I asked where NVIDIA stands in the race toward autonomous driving. So Tesla is leading the pack here. Tesla has talked about their full self-driving update coming by the end of this year. I think that's wildly optimistic,

this idea that the car will be able to drive itself point to point fully autonomously. But that is Elon's stated expectation. By Elon here, Chris means Elon Musk. He's the guy who runs Tesla. Three, two, one, zero. Ignition. Liftoff. Launches rockets, tweets a lot the tesla ceo tweeting today tesla stock price is too high

imo in my opinion and events unique baby names ah a12 musk smokes pot on podcasts i mean it's legal right it's totally legal anyhow ashow, as Chris said, Elon Musk expects his autonomous cars to be ready by the end of the year. Chris is a little skeptical. I think it will take considerably longer, but to have a fully autonomous vehicle, likely a Tesla first, by 2025 is not outrageous in my opinion. first by 2025 is not outrageous in my opinion. That's shocking to think about. My driver's license will expire in 2025. If cars drive themselves by then, will I have to get it renewed? More importantly, where does NVIDIA fit in? Chris says the early leaders in autonomous driving include Tesla and a company called Mobileye, which was bought by Intel three years ago, but that the market is large and there's plenty of room for NVIDIA. He's bullish

on NVIDIA shares. They trade at more than 40 times earnings, but earnings are expected to double in three years. I think there is still a good size role for NVIDIA to play. So it is a massive endeavor to hire your own chip engineers, to design your own chip, to go to TSMC or Samsung to do the foundry work for your chip. It is a massive endeavor. Most companies don't have the budget. A lot of them have scale, but not necessarily the profitability to fund something like this, a project like this. For NVIDIA, NVIDIA will be able to supply those other companies with these capabilities, with the hardware, in order to do this. Matt, how about a listener comment i know we have someone who disagrees with me on gold we do yeah we have uh peter um he sent us his file but he didn't give a town no town does that mean i get to make up a town for him sure can i make up a whole backstory? Yeah. I mean, I've got time. I mean, St. Augustine, Florida, right? Home of the Golf Hall of Fame. Has a mustache.

Long one. We're talking handlebars. What else? Wellington boots, but only on certain occasions. All right. I was going to say plays the ukulele, but I feel like those go together naturally. Good. So in an earlier podcast episode, I said that I think gold is fine as a speculative instrument, but I don't think it's ideal as a hedge against inflation or as a safe haven,

including for people who are financial doomsayers. One thing I said was that gold doesn't sit around thinking about ways to make itself more valuable. It just sits there, and it has relatively few industrial uses. Let's hear Peter's take. The U.S. Geological Survey suggests that while U.S. gold consumption has declined 10%

in the last 10 years, gold consumption for electronics production has actually increased fourfold. So it is benefiting from innovation and the expansion of the tech and space industries. 10 years, gold consumption for electronics production has actually increased fourfold. So it is benefiting from innovation and the expansion of the tech and space industries. Peter also says that gold's density means you can store a lot of value in a small space, and that it doesn't require much care or maintenance. Is gold the world's greatest investment? Probably not. As a safe harbor, however,

it makes a lot of sense. But ironically, that's only if we don't end up in a post-apocalyptic doomsday scenario. Thanks again. Keep it up. Thank you for that, Peter. And the mustache is looking sharp. Let's turn to a listener question now. Madda, who do we have? We've got a question from John from Chicago, and his question is about some stocks that he bought. Let's hear it. Six weeks into quarantine, I bought a few different stocks. It included Spy, IWM, USO, Disney, and Starbucks. I know that long term, I probably want to on to the SPY and IWM but as for like USO, Starbucks and Disney I'm kind of trying to figure out how long I should hold on to these stocks

before they recover after Corona Great question John Now much as I like Jim Cramer on CNBC I'm not sure I want to start doing a lightning round for individual stock holdings. For one thing, I don't have the sound effects. I'm not sure I have Jim's charisma either. But let me say a few things about the holdings you mentioned. First of all, when you talk about the recovery after the coronavirus pandemic,

what is striking about this period is that the stock market has recovered so much faster than the economy. I'm not sure I would still be holding anything waiting for a further stock market recovery. I'd only want to hold things for here that I like long term. Now, the first two things you mentioned, one is an S&P 500 index fund and the other one is a small cap index fund. Those are ideal long term holdings. The fees are low. You can set them and forget them and continue to save money in them. Then you mentioned USO.

That's an exchange-traded fund that holds oil futures. Frankly, I don't get it on that one. I don't have a passionate view about what's next for the short-term price of oil. And I'm not sure I view oil futures as a long-term holding. I'd rather have money in a company that can capitalize on oil by turning it into useful consumer products or by burning it for transportation or things like that. But you already have exposure to companies like that in your S&P 500 fund. And then you mentioned Disney and Starbucks. Now, I'm going to assume what you're doing here is you're keeping a core of low-cost index funds, and then you mentioned Disney and Starbucks. Now, I'm going to assume what you're doing here

is you're keeping a core of low-cost index funds, and then you're picking a couple stocks on the side because those companies seem exciting to you. I think it's a good approach, frankly. You don't have to take too much risk by relying on your stock picking for the bulk of your portfolio, but you can still get exposure to companies you like. I think Disney probably owns the most valuable entertainment assets on planet Earth. I realize this year has not been kind to Disney's theme park business. It's been difficult for movies. It's put a halt on television production, but I think Disney's growth prospects remain bright. Starbucks is similar in that it's a big company that can still muster pretty peppy

growth. So if you like those stocks for the pandemic recovery, I see no reason why you wouldn't want to continue to hold them long term. Madeline, I'm not sure if I brought enough lightning to that lightning round. What do you think? Let me find some lightning sound effects for you. Hang on. Yeah, really lay it on. Thank you, Peter, for that comment. And thanks to John for sending in your question. And everyone, please keep the questions and comments coming.

Just tape on your phone. Use the voice memo app. Send an email to jack.how. That's H-O-U-G-H at barons.com. Thank you for listening. Matt Alutsoft is our producer. Subscribe to the podcast on Apple Podcasts, Spotify, or wherever you listen to podcasts. And if you listen on Apple, write us a review. Follow me on Twitter to find out about stories and new podcast episodes. That's at jackhow, H-O-U-G-H. See you next week. about stories and new podcast episodes. That's at Jack Howe, H-O-U-G-H. See you next week.