Business Book of the Year: Author Stephen Witt on Nvidia’s rise

0:00In 1898, John Nuveen had a vision to underwrite municipal bonds that would propel American infrastructure into the 20th century. More than 125 years later, that vision lives on. Today, Nuveen has grown into a leading global asset manager with expertise spanning public and private markets, still investing with a future in mind. Nuveen. Invest like the future is watching. Visit nuveen.com slash future to learn more. Investing involves risk. Principal loss is possible. Hey there, Behind the Money listeners. It's Michela. Last week, the Financial Times and Schroeders gave out their prize for the best business book of the year.

0:42And the winner was a book about the company that has dominated headlines this year, NVIDIA. The book is called The Thinking Machine, Jensen Huang, NVIDIA, and the world's most coveted microchip. And it was written by Stephen Witt. My colleague Andrew Hill, who's the FT's senior business writer, has been in charge of running the book awards since 2005. And last week, he interviewed Stephen from our London studio. I had the chance to sit in on this interview, and I found it completely fascinating. You're going to learn a lot about NVIDIA's CEO, Jensen Huang, and the guy beyond the black leather jacket.

1:19And what you're about to hear is an abridged version of that conversation. I hope you find it as interesting as I did. Enjoy!

1:56research into NVIDIA, the company of the moment, and particularly Jensen Wang, the founder and chief executive of NVIDIA. Just in your own words, tell us a little bit about the project that you undertook. Yeah, so it actually completely grew out of my own conviction that ChatGPT was going to replace me. I started using it in 2022 when it debuted and was just like, man, I am cooked. This thing can write almost as well as I can and very soon is going to eclipse me. Like, what do I do? I guess I'd better just pivot to start writing about AI as soon as possible, while I still have the opportunity to do so.

2:30You know, and I was looking for a story, and I was looking at OpenAI and other things, and they were all very crowded. And I was like, you know, the hardware side of this is not really well explored in press. And at this time, NVIDIA was not the most valuable company in the world, and most people had never heard of them. But if you were adjacent to AI and you were adjacent to tech, you knew that all of these systems were being developed on NVIDIA's hardware platform. And that, in fact, they basically had a monopoly on training modern AI and these modern AI infrastructure. So if AI was the gold rush, they really were the company that was selling the shovels.

3:04I mean, they were the ones who were making all the money. So I reached out to them saying, hey, you know, I think maybe the New Yorker would be interested in doing a profile on Jensen. And they said yes, and the New Yorker said yes. And so it just started out as a magazine. profile. And I was able to win Jensen's trust, temporarily at least. Leverage that into a much deeper kind of exploration of who Jensen was, the places that he came from, and the NVIDIA story, which I thought going in, there probably isn't enough here for there to be a very good book. I think it's just a hardware company. It's probably had a rotating cast of CEOs.

3:41It's probably had kind of a tortured and maybe somewhat boring corporate history. No, it was one guy pushing relentlessly for one vision as an absolute maverick on the absolute outside and then ultimately conquering his whole industry. And that was just so compelling. And his character, he was such an unusual person. I've never met anyone like him. And so I got very lucky that not only did this company end up becoming the most valuable company in the world while I was writing the story, but also the people who were driving it were just such compelling characters. Right. Now, you set Jensen very clearly at the center, which is where he is of the company, really the driving force.

4:22But in some ways, we have in our minds, I guess, a picture of Silicon Valley entrepreneurs and multimillionaire billionaires, which looks like a Jeff Bezos or an Elon Musk or somebody of that ilk. And Jensen Huang doesn't seem to be that kind of leader. Not quite, no. He's much more like almost a lead from, I wouldn't say lead from behind, but he's very driven, that much is for sure. He's very ambitious, but he's also very humble to the point where he's almost like self-effacing. Like he genuinely never thinks that he's doing a good job, No matter how much he succeeds, no matter how big NVIDIA grows, he's convinced that he's not doing enough.

5:08Fear, guilt, paranoia about competition, you know, kind of beating himself up each day. This is his fuel. This is what makes him go. So, and I've said this before, but he's just a very neurotic guy. And I don't mean that in a clinical sense. I just mean he's just totally driven by negative emotion in a way that I have never seen a CEO be before. But he's able to use this to kind of repurpose his energy into these very productive kind of use cases. But it's hard to be Jensen day to day. It's almost nightmarish. He's constantly, like, paranoid about competition. He's constantly paranoid about people taking NVIDIA down their competitive position.

5:46And he drives his staff incredibly hard. He's really demanding. And there's at least one episode in the book where the company achieves an extraordinary landmark. And literally the day after, he is saying, we've got to do better, buckle up for the next phase. This is only the start. There's no point at which you see him sort of celebrating. The company IPO'd in 1999. He's taken it from a company that was in a diner at Denny's six years earlier to being ringing the bell at the stock exchange. And the next day, he's writing emails like, we're way behind. The company is doomed. We're going bankrupt any day now.

6:21This is 24 hours after the IPO. He's just grossed$100 million. This is the email he's writing. Right. And I mean, he wasn't wrong in the sense that it did come very close. It's had at least two bankruptcy scares. Yeah. And two points at which its shares have dropped 90%. Right, right. And maybe a third one. We'll talk about that. Well, that may be looming. But yeah, so he has been through it. And his mindset is we have to act out of desperation at all times. We have to pretend or act like the company's on the verge of bankruptcy at all times. because that is the only way that we're going to continue to feel that desperation that we need to fuel innovative products and to shore up our competitive position.

7:00You know, the thing about Jensen is, you know, as NVIDIA became the most valuable company in the world, he became extremely, visibly stressed. It was hard. I mean, he knew if you're earning a 70 to 80 % gross margin on your product, it's just going to attract competition. Suddenly there's a target on your back. Everyone's looking at Jensen. They're all trying to knock off his products or make something similar or better if they can. And so it actually, I think, made his life even harder in some ways. That and then, you know, AI came along and I think he correctly recognized that he had gone from being an executive who was selling, you know, a profitable product, video game equipment to something else.

7:40Like now suddenly he's at the center of a world historical revolution. And I think he really embraced that role. That's when he brought out the leather jacket. That's when he brought out the Jensen we know, the sort of charismatic, visionary CEO swaggering around on stage. That happened with AI. And so he also really has grown and evolved in his role over 30 years to become the person that he is today. I mean, just to paint a little bit of the history in here and then bring this right up to date, but this was a company that started with chips for gamers. And then you've described it, I think, as a sort of field of dreams moment where he builds something empty, essentially saying, there's going to come along a use case for this.

8:22And then he sort of gets lucky with AI almost. You could say so. People go back and forth on how lucky Jensen was. Here's what happened. In the early 2000s, they had the leading video game controller in the world, but that's a niche market and it's saturated. So now where do we go? What started to happen was that scientists were looking at these video game controllers I was like, God, there's some really impressive circuitry in here. If only I could somehow hack this to run my quantum physics model or my model of the Big Bang, let's say, which is one of the things they did, I could get some really impressive results.

8:58And so they started hacking into the programmable shaders to get to the circuitry underneath. Jensen saw this and said, well, don't hack the hardware. I'll build you a whole scientific computing platform. And not only that, I'm going to ship it on every card we sell, including the ones I'm selling to the gamers. The ones with like a guy shooting a gun on the front in the hardware I'm selling, I'm going to put scientific computing tools. So who is this for? Well, it's not for top edge research scientists because they all have access to supercomputers. It's for kind of marginalized scientists. It's for mad scientists, right?

9:34It's for scientists whose research is so out of favor that they can't even get a$5 ,000 research grant to study what they want to do. And so Jensen kind of flooded the low end of the supercomputing market, which was a very tiny market, with this product and saturated it with the hope. And you can see them saying this in the time that they will unlock some use case that will be like a paradigm shift in computing. Did they know it was AI specifically? No, they did not know that. Did they hope that something like this could occur? Yes. And it's very clear, even from talking to people at the time, that they thought they could unlock some new branch of science that then they would become the dominant provider for.

10:14I think AI came as a total surprise to them, though. You talk in the book about the concept of the zero billion dollar market. Just explain what that is. This actually is Jensen's phrasing. It's a market with no customers at all. Right. So you're going to build a product that literally just basically doesn't have any customers. and you're doing it with the hope that someday someone will come along and start using your product or that even you can evolve your customers around you. And this is really what NVIDIA did. The AI hardware market did not exist. Well, there was no such market when Jensen entered it in 2012 or 2013.

10:47I mean, he really just invented it from scratch and then built an entire software platform around it and then just opened a sign like, say, you want to do AI? Come use this specialized AI hardware, but there was no one who was expecting such a thing to even exist. It's hard to remember, but AI was a career graveyard for decades. I mean, people did not go into AI to make money. It had no investment capital almost at all involved with it, say, as late as 2010, 2011. So it was really Jensen's platform that turned that around and made modern AI possible. And really, the thing to understand is there's two technologies here.

11:23One is Jensen's parallel computing suite. This is what powers the circuits. And that was a minefield, or it was a road littered with corpses, as Jensen described it. Every parallel computing company that had ever attempted to start had gone bankrupt. Similarly, every kind of pure play AI company that ever started had also gone bankrupt. So it was actually two dead fields that could only kind of work when they were united in tandem on Jensen's platform. So a stroke of genius, a stroke of good luck, possibly a combination of the two. I mean, it's very difficult to break that out. If you go out into the ocean with your net and you go stand in some part of the ocean that nobody stands in and you throw the net into the ocean every day for 10 or 11 years, and then at the end, you catch the biggest fish anyone's ever seen, did you get lucky?

12:11Well, maybe. But you also put yourself in a position where you could get lucky. Yeah. And you were the only one standing there. Yeah. So let's bring this right up to date because obviously with any business book, specifically one written about people who are still active, you're at the mercy of what happens after the book is published. The book came out in April and obviously was written and finished before that. So you've had, just to mention a few, the whole geopolitical uncertainty of a Trump second administration with tariffs and US-China rivalry coming into play. You've, of course, had even more growth from NVIDIA itself, But you've also had rivals emerging with open source foundation models in China and elsewhere that might challenge the American original LLMs.

13:03You've got the challenge from Google, which has a new, I think it's called Tensor Processing Unit, which some of the markets think is one to bring. To me right now, that is actually the biggest risk to NVIDIA because Google's Gemini is topping the kind of evaluation leaderboards now. They actually leapfrogged the GPT systems to become number one, and it was built outside of NVIDIA's ecosystem. So that's a huge threat to NVIDIA. There's now a credible alternative hardware platform that you can use, or at least that Google is using. It's not clear if they'll open that up to the public or to other vendors, but they've built something quite incredible outside of the NVIDIA ecosystem.

13:42The market hasn't really responded to that yet. The market still has a very positive view of NVIDIA's prospects, so perhaps I'm overrating it. But to me, it seemed like an almost existential threat that they were able to do that. So on the key question of the moment, is NVIDIA heading for another fall? Maybe not a 90 % fall, but it's come down already from the extraordinary$5 trillion plus that it surpassed briefly in October. I mean, I do think that we are living through certainly the most profound technological revolution of my lifetime. I think the hype is real regarding the technology. Whether all of the profits from it continue to flow to this single hardware vendor, I think, is a really big question.

14:21It's easy to imagine a world 10 years from now where NVIDIA is not the dominant provider and their market capitalization has dropped by a significant amount. Particularly, and there's a number of ways that could happen. We could plateau with AI, which could certainly happen. The premise of all of this is shoving more NVIDIA microchips into the data center will result in better AI. So far, that has proven true, but it's not some immutable law of the universe that that will always happen. So that could start to plateau, and that would make NVIDIA stock price go down a lot. Google, Huawei, Broadcom, any other provider might show up with a cheaper stack, and people start to switch away from NVIDIA's platform.

14:59And then, you know, there's kind of the geopolitical tension between the U.S. and China. If war broke out or China invaded Taiwan, that would be very bad for NVIDIA stock price. So there's any number of things that we could think of. The upside is Jensen is extraordinarily capable individual. And I think also there's a big bet on robotics shaping up right now. And I think that's starting to get priced in an NVIDIA stock price. I see that with Tesla, too. Yeah. I mean, how, from your knowledge of Jensen Huang, is he reacting to this? This is just fueling his paranoia, presumably. Even now, in a rant somewhere in the world to his colleagues saying, you know, we've got to buckle down now for the next.

15:42No, he told his colleagues, like, look, you have to understand if you're working at this company, there's a team inside Google whose job is to kill us. There's a team inside Amazon. And they're smart. They're great. They're highly capable. They're some of the best engineers. And all they want to do is destroy us. And Jensen, the thing is, in the early days, okay, back in the graphics controller days, it was a knife fight. There were like 60 or 70 companies trying to build this stuff. And Jensen won that fight. It was like the movie Battle Royale where all the kids are trapped on the island and only one person can emerge.

16:15He really lived that. I mean, and he gutted and destroyed 40 or 50 competitors to be the leading graphics controller guy. The way he did that is with his lieutenant, David Kirk, and a couple other guys, they would just go on a whiteboard and list out the competition. They would list out the best engineers who worked for the competition, and then they would attempt to poach those people in what they call the brain extraction strategy to kind of hollow those companies out from the side. It was as brutal as you make it sound. It was absolutely ruthless. So Jensen has buried two dozen rivals, just buried them and salted the earth.

16:52And he knows that his rivals want to do that to him as well. So now inside Google, inside Amazon, they have those lists and they have these lists. If we can poach this guy from NVIDIA, get him to come work over here, maybe we can start to get a competitive advantage over these. Will that work? Because one of the things that really interested me is that for all that you describe, Jensen Huang, and we'll come to the point where you got a taste of his temper, for all you describe him as delivering these two-hour rants when things go wrong. He has engendered this extraordinary loyalty among staff, and he is himself loyal to them.

17:29He rants to somebody in one of the passages, he gets two hours of Jensen full on, but then he keeps him on. Yeah, he never fires anybody. I mean, he will demote you for underperformance, like you will have responsibilities reduced, but he basically, except for very egregious cases, never fires anybody. He really wants you to improve and thrive and to buy into the company. The other thing is he's extraordinarily generous with stock grants. And this keeps people around, especially now as the stock has gone up a ton. NVIDIA's policy for a long time because the stock was so volatile. If you work there, you could buy this company stock at a 15 % discount to the prevailing market price at any point in the last two years.

18:13So as the stock started to go up, you know, four or 500 percent a year, employees could still purchase stock like it was two years ago. They could purchase NVIDIA at 2022 or 2023 prices. So that was just free money, right? It was just free money. I think the statistic that I saw was 50 percent of employees in NVIDIA have a net worth above$25 million, which is crazy to think about. And they still show up to work every day. And this is, I think, the thing Jensen can do. He has them convinced, and he's correct. You can't leave. This is the once-in-a-lifetime opportunity to work on this engineering project.

18:48If you leave, you will spend the rest of your life missing out and regretting that you weren't here when we built AI, when we built the foundation for AI. It's more than money. There's a sense of mission, and I usually would not believe this kind of hype about a corporation. Like, I don't think this is true at Amazon. But I think in NVIDIA, they genuinely believe it. like we are working on the most innovative technological product of all time. We can't leave.

19:37seek today. Nuveen. Invest like the future is watching. Visit nuveen.com slash future to learn more. Investing involves risk. Principal loss is possible. I was going to ask you about mission because I had the impression from your book that Jensen himself, unlike perhaps a Musk or a Bezos, isn't thinking about, I'm going to change the world. He's thinking about in that paranoid neurotic way, how do I keep being first here? And the consequence of changing the world is not exactly a byproduct, but it happens because we're the number one. Yeah, this is the thing. Jensen is ultimately a businessman, right?

20:17So if somehow NVIDIA could make the most money selling orange juice, they would sell orange juice, right? He's not ever going to just quixotically pursue some bizarre technology just because he thinks it's interesting. He'll have a strong business case for doing anything he wants to do. But what he has discovered and what has happened, going all the way back in his career, even before he founded NVIDIA, he was always working on the most technologically complex product. Even sometimes when in the short or even intermediate term, that product lost money. His point of view is, you know, we're building circuit architecture.

20:53We're building microchips for a very long kind of duration use case where we start the architecture and eventually our customers will come to us. And there is a strong business case to be made for doing this. It looks counterintuitive. It looks like we're being crazy. It looks like we're losing money. But if we can stick with this for long enough, whole entire industries will grow around our platform and then we will be the only provider selling these services. So that's the business case. He's doing that now with robots. How much of that is vision? How much of it is kind of like sci-fi technological world building?

21:27And how much of it is just strictly dollars and cents? If you talk to Jensen, he'll tell you it's all strictly dollars and cents. He told me he's never read a science fiction book in his life, which I found kind of fascinating and bizarre. But I think it's true. Interesting. Almost the anti-Elon in that sense. I mean, and I say this in the book, Elon's going to start with a vision of what he wants to do. He's standing on the surface of Mars. He's going to build backwards to build technologies to get him to the surface of the Mars. Jensen's going to do the opposite. He's going to look at what's directly in front of him, and then he's going to project forward as far as reason and logic will allow.

22:00But no further. He's not going to speculate on what the sci-fi future of humanity looks like. He's going to speculate correctly, logical reasoning and deduction, seven to eight years in the future, which is one to two years further than anyone else can go. But it's all logical. It's all coming from like this happens and then this happens and then this happens. Does that make him in some respects more dangerous in that he's in a sense the neutral party? He's just going to say, I'll supply whatever needs supplying for whatever needs doing. And then at the very end point when I don't know we're surrounded by gun-toting robots, he's going to say, what have I created?

22:36You know, I asked him this question multiple times. He didn't like it. He does not like this question. I imagine. You know, and I was like, well, what are we building here, Jensen, in the long term? We're just, you know, where does this go in 20 years? And it just wasn't the kind of question he was willing to entertain. You know, there's an empty part of the map. You can't just draw a dragon on it, basically, was his point of view. We don't know what's out there, but I'm not afraid just because we don't know. And it's foolish to entertain these kind of what he called ridiculous sci-fi scenarios.

23:05But what made it complex for me was I was like, OK, maybe I'm just being ridiculous. Maybe none of this fears about AI are real. But then I would go talk to people like Joshua Bengio and Jeffrey Hinton, and these are the pioneers who built the software side of this, built the neural network software side. And they saw further than anyone where this was going to go at a time when it wasn't popular to study this stuff. And now they're terrified of what kind of the future looks like. And so to me, it wasn't just me asking this question. I was asking it on behalf of guys like Hinton and Bengio. And I was really shocked at how dismissive Jensen was, especially of Hinton.

23:40He felt Hinton was tarnishing his scientific legacy by bringing this stuff up. And, you know, Hinton won the Nobel Prize. And I was like, wow, really? You think so? But I think for Jensen, he just doesn't want to hear it for whatever reason. And you quite forcefully let me know that in our final interview. Yeah, tell us a bit about that because you did trigger his famous anger. Well, I just kept, you know, I'll tell you how I, I'll tell you, this started out as a New Yorker profile of Jensen. But what I did was I was like, okay, I have to make Jensen like me. Like, how do I do that? And my theory was that Jensen doesn't like having his time wasted, okay?

24:14You know, Jensen's allocating a trillion dollars worth of capital each year. So, like one minute of his time is roughly worth like a million dollars, right? So, what I did was I went and I watched 40 or 50 prior interviews with Jensen. I made a list of every question every journalist had ever asked Jensen, ever. And then I made my list of questions and I made sure that none of those questions were any questions that he'd ever been asked before. And this worked really well because it's also the way Jensen interacts with his executives, as it turns out. I got lucky. When you are an executive working in NVIDIA, you cannot revisit a past question.

24:47You will get screamed at. You're wasting Jensen's time. It all has to be new, and it all has to be really well-researched, and you have to be able to answer four or five deep questions about the subject you're talking about. So the amount of preparation that these guys do in anticipation of, say, 10 minutes of FaceTime with Jensen is just unreal. And I fortunately had done the same thing, and so this gained his trust. But I did keep revisiting the risk question. It was the one I always went back to. And so I think his point of view was now I was wasting his time because he already answered that question.

25:16And that's what really made him explode at me more than anything else. But even so, you point out that it was unnecessary and disproportionate. It's disproportionate. It's self-indulgent. I believe his displays of anger have a degree of self-indulgence to them. I mean, he has a theory about why they work. and I think they do motivate people at his company in the manner that, you know, I think probably a general in a war, or I often think of Sir Alex Ferguson, the manager of Manchester United, what they call the hairdryer. People just scream at these players. It seemed to motivate them in some ways.

25:49And so maybe there is power coming from these kind of negative emotions. You know, it's not the product of therapy culture. I'll tell you that much. Like it's a kind of older way of doing things. But if you grew up, you know, playing for a tough soccer coach, or maybe if you were in the military, you have some experience with this kind of leadership style. And it works at NVIDIA. Some people can't hack it. Some people are like, yeah, I'm just, I just, I talked to Clement Farabett, who's building Google's engineering stack now, and he was formerly in NVIDIA. He's like, yeah, Jensen just started screaming at me one day and I resigned on the spot.

26:20Right. I just, and this guy's really talented, skilled engineer. He's like, I wasn't going to put up with it. Yeah. And so some people can't hack it, but I think if you're willing to kind of take the heat, the rewards are there for you. Even Clement was like, look, NVIDIA, he motivates people more than anyone else I've ever seen. So what part did the sort of childhood and childhood obsessions of Jensen Wong play in his rise? There's a lot about table tennis in the book, to my enormous surprise. Yeah, well, Jensen, I remember talking to his friend Ben Bays, who had been in sixth grade with Jensen and then never saw him again for 40 years.

26:54And I was like, were you surprised that he became this kind of world conquering billionaire? and Bam was like, no, not at all. I just, I always knew he was special. I always knew he was different. So Jensen had this kind of like star quality, even when he was young, he started playing table tennis when he was, I think, 14 or 15. And within six months was like a nationally ranked player. He has this incredible capability of just learning stuff at accelerated rate, just through practice diligence and being very bright. And he was actually naturally quite athletic too. So he became obsessed with table tennis, slept in the table tennis facility sometimes.

27:31And his coach wrote a letter to Sports Illustrated, this US magazine in the 70s, saying, Jensen Huang, one of my students who's 16 years old, is like the greatest table tennis junior I've ever seen. And I dug this up. I found it along with like, you know, I talked to all his old table tennis buddies. But what it showed was, I think, two things about Jensen. One, he can learn so fast. His IQ is so high. Like his ability to just absorb, process new information and get to like a world-class level very, very quickly is just unparalleled. He just has an incredible ability to learn new stuff. And then two, he has that competitive athletes killer instinct.

28:13He absolutely hates to lose more than he even likes winning. He just absolutely hates to lose and is so driven by ideas of not losing. I mean, that's what really pushes Jensen. Right. You made me think of another question, which is, you know, what is the apre Jensen? Obviously, at some point, there's got to be a post-Jensen one. Yeah, as far as I can tell, there is no succession plan in NVIDIA. So key man risk is huge. If a satellite falls and hits Jensen on the head, like, I don't know what's going to happen over there. He does not have like a second in command. There's Jensen, and then there's like 60 executives that answer to Jensen.

28:51But there's no like right-hand man or woman or anything like that. You know, his kids both work for the company, but they don't have technical backgrounds. They're not real successors to Jensen. The two names I heard the most were a guy, Jay Puri, who runs sales at NVIDIA, and a guy named Ian Buck who developed the CUDA platform, which is the kind of supercomputing platform. I think probably they'll give it to Puri. Jensen's secret in some ways is that he always interfaces with his toughest customer. He always goes, and this is even before NVIDIA, he would always go to the hardest computing and engineering customer that he could find and lose money to build them a product.

29:33And Jay Puri does that now. So Jay Puri is out there talking to the AI guys, yeah, also quantum physics guys, quantum computing guys, medical imaging guys saying, what's the thorniest problem in your entire field and how could we fix it if you had a million times more computing power? And then they come back and they prioritize what to build and they will absolutely lose money just to have those customers close to them. Stephen Witt, congratulations on the prize, author of The Thinking Machine, and thank you very much for joining us. Thank you so much. Thanks for listening to this week's show.

30:09If you'd like to learn more about Stephen's book or any of the other shortlisted authors in the 2025 Awards, check out the links in our show notes. Behind the Money is hosted by me, Michaela Tendera. This episode was produced by me and Misha Frankel Duvall. Fact-checking by Simon Greaves. Sound design and mixing by Alexander Higgins. Original music is by Hannes Brown. Topher Forges is the FT's acting co-head of audio. See you next week.

31:02Nuveen. Invest like the future is watching. Visit nuveen.com slash future to learn more. Investing involves risk. Principal loss is possible.