Transcription
I think for Jensen, you know, he's not going to do anything that violates US law ever. I mean, he needs America and he needs the support of America. America is where his company's based and he is an American, you know, and before that he was from Taiwan, you know, he's not ethnically from the mainland of China.
So, I think his point of view is more that his mentor was Morris Chang of TSMC. And Morris's point of view was always like, let's make Taiwan strategically, geopolitically important. Let's build this TSMC company that sells all of these microchips throughout the world in Taiwan. That actually reduces the likelihood of geopolitical tensions or engagement between China and Taiwan if they're integrated. I think that's also Jensen's point of view. I think he has studied from Morris Chang. Uh, you know, Morris and Jensen were close friends. In fact, they spent a lot of time at each other's houses. So I think Jensen's point of view is if our economies are integrated in this way, it actually reduces the likelihood of geopolitical tension.
Conversely, if they build their own stack and they don't need us, then maybe there's actually increasing the tense uh the the the the sense of like geopolitical tension or maybe even war. And so I think Jensen's point of view is that it would be better if our economies were integrated. That makes us need to cooperate. That forces us to cooperate. That was always Morris Chang's point of view and I think probably Jensen is doing that as well.
>> Okay. Now here's a follow-on to the issue of whether a cheaper, better mousetrap can be built using a slightly different technology. So I'm really saying TPU versus GPU for those of you who are nerds, which I am not. Now my question, which was originally about Google versus Nvidia, is now about China versus Nvidia. Why can't China go the way that Google did? They can do the TPUs or they can't do the TPUs, and therefore is China in a position where they can catch up and where they can rival uh the almost US monopoly on on higher-end chips in a very short period of time?
>> Yeah, the bottleneck there is actually the light printing technology that we use to make microchips anywhere. Uh, this is called photolithography, and they use this extreme ultraviolet light to print the chips. Um, China has not yet built those machines. They're extremely finicky. They're actually made by a third company out of Holland called ASML. Um, the US has been very aggressive about blocking China's ability to uh access these light printing machines. And without those, you can't build world-class microchips. And so, actually, this is the sticking point. They don't have the printers they need.
Now, can they develop them internally? Probably. Yeah, actually, given what they've done with high-speed trains, given what they've done with cars, it seems quite likely to me they could also build a a world-class light printing machine. But it's a long technology cycle. It's going to take them three to four years to catch up. So the Chinese bet, if we can call this the gamble, is can we afford not to use world-class equipment to train our AI until our semiconductor infrastructure is up to the global standard? And it is a big gamble. I mean, you really can't afford, as Google has shown, to even be a few months behind right now. Like, OpenAI just fell behind Google, and that caused Sam Altman to issue a a code red because he's a couple months behind. So, being years behind in this race is very, very costly. It's a tricky calculus for China. It's a tough game. My guess is that while sort of officially discouraging the use of Nvidia equipment, they will continue to do so one way or the other, whether that means going outside of China to do the training or getting chips in through Malaysia or some other channel uh until their own industry has caught up. Just a guess, but that would be my guess as they're thinking.
M Stephen, um, I mean, just like any other industry uh in this uh particular sector, the product matters, the timing matters, and you actually recently mentioned how AI sort of um happened like at, you know, it was lucky that it happened and for Nvidia to really take off, and also the leadership matters. And I believe that's the reason why we decid you decided to write about Jensen Huang and Nvidia. What is it about Jensen Huang that you understand that you want the rest of the world to understand to determine whether he is really ready for the next wave of AI adoption?
>> Jen is one of the most extraordinarily capable people I've ever met. He's might be the single most adaptable individual I've ever met. You know, he has really had to adapt himself and his company multiple times to to meet the demands of the moment. Uh, we've talked about his friendship and closeness with Trump. They've appeared in public seven times this year. Well, before that, Jensen had basically no political experience at all. I mean, as far as I could tell, he's never made any political statements. He never made any political donations of any time, any kind. That was completely new territory for him. But he did quite well at sort of managing Trump and getting what he wanted out of the US government. So again, that's his adaptability. Uh, you know, one of his executives told me that stick uh interacting with Jensen is like sticking your finger in the electric socket, and I found that to be true in my own interactions with Jensen. He's extremely intense. He's very driven.
And then his true secret, and this is a thing that people really need to understand that makes him so hard to replace. On top of all of these qualities, at his core, he is a world-class computer scientist. Jensen could design these microchips himself. That's actually how he got how he got started designing microchips on paper 40 years ago. So he is completely immersed in this industry. He is an absolute master of this industry. He is the best at what he does. Okay? And so it's very hard to find all of those qualities in one person. The managerial brilliance, the adaptability, but also the deep, deep domain expertise. You cannot just go and get some other man, uh, you know, MBA to run this company. They have to know computers as well or better than anybody. It's hard to find all those things in one person.
>> I I really hate this question, but I'm going to ask it anyways. Do you think >> What do you think keeps Jensen Huang up at night? Uh, you often use the word neurotic uh to describe the Jensen Huang, not in a negative way. Uh, but what do you think that he is most worried about in December of 2025?
>> Great question. Uh, there's a lot of competition. There's Google, right? We talked about that. There's geopolitical tension with China, right? There's that. There's Trump. There's Broadcom and AMD at home. But I will tell you what I believe Jensen is most worried about is not any of those things, but missing the next wave. Where is the next three to four trillion dollars of market capitalization in Nvidia going to come from? I think Jensen believes it's going to come from a coming wave of robotics, from autonomization of all sorts of devices. Jensen has said that anything that moves in the future will be autonomous. That means cars, but it also means vacuum cleaners. It means walls in buildings, even. It means home robots. Jensen is very much enthusiastic about participating in that wave, that coming wave of robotics, which will be a a trillion-dollar industry, if not more. But I guarantee you, it also keeps him up at night. Scenarios where he misses that wave. And this is the genius of how Jensen thinks. He's not just looking at the competition today. He's looking at the competition tomorrow, two years, four years. How can Nvidia be centrally positioned to be maybe even the only player in a new wave of technology? That's what made Nvidia so big before, and it's what's going to make it grow if he succeeds again.
>> You know, Stephen, the next wave being robotics. I mean, Jensen's not the only guy. You've got Elon Musk and uh and uh Tesla going down that same road as well. So, we know, and I have to confess, I haven't read the whole book, but I flipped through parts of it. Right. Uh, you're sort of describing Jensen as as a lot of people already know him. Right, a visionary, someone who basically uh uh uh uh premised his career on basically a a field of dreams bet. You know, if you build it, they will come, because when he started down this road uh going away from uh moving on from graphics, the market didn't actually exist. He actually created it. Right.
>> We know that. And you mentioned he's a terrific computer software uh computer engineer. That's not in dispute either.
>> But is he thinking about the end state? What happens when everything becomes uh automated? What happens when you know robots are able to think for themselves? What happens when they start picking up guns?
>> Yeah. So Jensen rejects science fiction narratives and very forcefully, as I learned when I brought one up with him. He does not like to let science fiction scenarios entertain his own thinking. He always thinks from what's in front of him, the microchips in front of him, and then he moves forward as far as logic will allow into what industries are going to look like in the future. But if you look at someone like Elon Musk, it's almost the opposite. Musk is going to start with a vision of standing on the surface of Mars and then he's going to move backwards to the technology that he needs to develop today to get there. Jensen is really the opposite. He He's not projecting towards some science fiction scenario. He's just going to look at what the circuits can do today and reason critically about what they're going to do tomorrow, maybe two years, maybe four or five years, but he's not going to think 20 or 30 years in the future because he thinks that's impossible.
Now, having said that, he is extraordinarily good at understanding the implications of what a microchip can do today and what new frontiers of computing may be unlocked. Um, again, robots is a big part of it. Jensen, I think, sees a ton of potential in medicine. I think that's going to be a very big deal for Nvidia. They're investing a lot of resources there. Um, and then, of course, just maintaining their basically central position in AI. Uh, much more so than anyone anticipated, maybe even Jensen anticipated. AI has turned out to be heavy industry. I mean, we are brute forcing our way to smarter machines with these gigantic data centers that are filled with hundreds of thousands and sometimes millions of NVIDIA chips. You know, Jensen has described those as AI factories where data goes in and intelligence comes out, and that's the world we live in now. Um, whether Jensen can maintain that central position with all of these competitors coming online, I guess is a question we'll answer in the next couple of years.
>> Yeah. So there there's the company Nvidia and there's also the man, right? And to Sherry's point a couple of seconds ago about leadership. I mean uh I think the job of most great leaders is to start thinking and planning for succession. That doesn't seem to be happening uh at Nvidia. I mean, okay, granted, the guy just he's just past middle age. He's not that old, right? But I mean, there is the whole issue of of keyman risk. Were you able to talk to him about that and and what did he say?
>> I was I was like, what's your succession plan? He's like, "I don't have one." Simple as that. There's no designated successor. You know, if you look at a typical company, you've got one guy on top or one woman on top, and then you've got maybe 10 or 12 kind of like people answering to the chief executive. At at Nvidia, it's like 60 people. Okay, Jensen's at the top. There's no second layer of management. And then the third layer is just like 60 or 70 people all answering to him. So, how he manages to do that, it's sort of extraordinary. It's not what you expect from sort of classic modes of management, but he somehow is capable of just answering, you know, 10 or 15 times as many emails, responding to 10 or 15 times as as initiatives, managing that many more people as anyone else.
Now, part of the secret is that those executives are themselves extraordinarily capable. Jensen is surrounded by brilliant, brilliant people, and he drives them very, very hard. So he can go to one of his lieutenants and give them a task and say, "Come back in six months when you've completed this. I'm delegating all of this authority to you. Go do it and come back." And because he's been so careful about who he's hired, that can actually happen, which I think is rare at other companies and harder to do. But, you know, I talked to the members of the board a little bit, like, is there a succession plan? And one board member said, basically, look, we have a couple names in mind. We've thought about this issue a little, but I'm going to be real with you. Jensen is irreplaceable.