Transcription
Good evening everybody. Thank you so much for joining us for what is about to be an exciting evening.
Just before we get into tonight's proceedings, just a tiny in-house rule from me. Please do resist the urge to take photographs or videos in your personal mobile devices. We have professional photographers in house and I would trust you to leave it to the professionals to take care of that. Thank you.
Okay. So, created in 2017, the Professor Steven Hawking Fellowship was established to recognize individuals who have made distinguished contributions to science and technology. Administered every year by the Hawking Fellowship Committee, the fellowship acts to memorialize Professor Hawking's legacy and to celebrate innovation in the field. Fellows from past years have included Bill Gates in 2019, Jane Goodall in 2020, and the Open AI team in 2023. Tonight's honoree makes a stunning addition to that list of inducted fellows. Jensen Huang is the founder, CEO, and president of Nvidia. Founded in 1993, Nvidia has pioneered accelerated computing. From the invention of the GPU in 1999 to major developments like the Grace Hopper Super Chips, this Stanford graduate and alumnus has been a trailblazer in innovation and tech pioneering since the very beginning. Please give me a warm Cambridge Union welcome in inviting the 2025 Professor Steven Hawking Fellow, Mr. Jensen Huang.
Heat. Heat. Thank you. Thank you.
And now I would like to take the opportunity to invite Lucy Hawking, Professor Hawking's daughter, to present this year's fellowship.
Thank you. So maybe I'll have my wife hold it.
Oh, perfect. Please.
I feel safer with her holding it. As far as... Thank you. Wow. This is quite a moment. You know, today when I arrived, I was so overwhelmed by being on Cambridge that I decided to, uh, bask in the moment and write a proper speech. And, and so I, I was in, is it the St. James Lodge? Is it, uh, where it was a beautiful place? The fireplace was, uh, was going, and, and, uh, I sat in a, must have been a 30,000-year-old chair, chair, and, um, it was, it was incredible. And, and here, here it is. I wrote it.
So Lucy, Ivan, members of the Hawking family, and everyone, everyone here. Incredible. Everyone here. I'm deeply honored to receive the Steven Hawking Fellowship. And to receive it here at Cambridge is profoundly humbling. Cambridge is a cathedral, a cathedral of world-changing ideas. Newton, redefining motion and gravity. Darwin, questioning creation. Maxwell, one of my favorites, uniting light and magnetism. Turing, another favorite, imagining a machine that could think. And Stephen Hawking, expanding our understanding of time and the universe. Professor Hawking's life showed that curiosity has no boundaries. By the way, that was one of his theories, as you know, no boundaries. Even when his body was confined, his mind traveled beyond the stars. He reminded us that discovery does not come only from intellect but from conviction and optimism. And through his life's work and through the way he lived his life, he inspired us to look beyond our limitations, to meet challenges with curiosity and humor. I can think of no higher compliment than to be associated with that spirit.
Nvidia is a story of nearly impossible odds. With the three of us, three friends in a townhouse in 1993, uh, set out with an idea that we invent a new way of doing computing to solve problems that normal computers cannot. Along the way, we invented a new product category, the GPU. We invented, in fact, a new form of computing called CUDA accelerated computing. We invented new strategies to take that technology and that architecture and proliferate it literally all over the world. And along the way, we created the instrument of scientists, artists, designers, dreamers, and most importantly, we sparked a new industrial revolution, the AI industrial revolution. We've come a long ways, and our discoveries have led to the most impactful technology of our time, and, uh, probably of all time, the ability to manufacture intelligence. We've now seen the technology advance incredibly in the last decade. It is now transforming every application, every field of science, every industry. Everyone will be impacted. Every company will use it. Every nation will build it. It will now be part of infrastructure, the intelligence infrastructure. And to realize that, like energy, like the internet, we will now be building AI infrastructure all over the world. That observation, that realization led to the company that you see today. The company helping every company, every industry, every country build out artificial intelligence as part of their social fabric.
And so here we are, 33 years later. I'm the longest-serving tech CEO the world's ever known. The way you achieve that, by the way, is, uh, don't get bored and don't get fired. And in a lot of ways, I feel Nvidia and I have been, uh, reborn. That the company is renewed. That the entire technology industry is being completely reinvented. Every single layer of the single most important instrument of humanity, computers, is being reinvented from chips to systems, the software, algorithms, applications, and the potential impact. In no time in history has this happened. Surely not in the last 100 years. And now the entire tech stack is being reinvented. The entire tech industry is being reinvented, and in fact, we do feel completely renewed. And so, in a lot of ways, I'm starting where you're starting. We're all newbies now. We're all looking at a future beaming with opportunity and equal amounts of concern. No technology of this capability can be advanced without thoughtfulness, without care, without scrutiny. And yet the opportunities ahead are incredible. And so I feel like a startup again. 33 years later, Nvidia is now the world's largest startup. And so I'm, I'm incredibly, incredibly proud to receive this. It's a great honor. And, um, I'm looking forward to spending time with all of you. Thank you.
It might be me. It might be that one of us has to go. Should I?
Perfect. I think we're fine. Wonderful. Well, thank you very much for joining us today. And I think that it would be a miss if I didn't begin by saying congratulations on receiving the 2025 Professor Steven Hawking Fellowship.
Thank you.
With such a rich career and legacy and sense of, of achievement in all that you've been able to do so far.
You're making me nervous. I promise you I'm on my best behavior today.
Taking that, that setup is too much.
You know, taking all of that into consideration, it's tricky to figure out where to start. But to quote The Sound of Music, let's start at the very beginning. It's a very good place to start. So, being born in Taipei and moving to America when you were nine. Um, what comes really apparent in your early years is a great sense of determination, drive, and personal discipline. How were you able to cultivate that through your education, through your time at Oregon State, through your time at Stanford to keep going and to continue being relentless even in your early education?
My mom taught me English, and she doesn't speak English. And that kind of tells you all, that tells you something about, about, um, the impression that parents could leave with their children. Uh, she told me when I was young that, that, uh, that I was special. That somehow, somehow I took tests and I was, I did well on tests, and, um, she encouraged me. Uh, often times, if people tell you that you're, you're better, greater, uh, more capable than you are, you might live up to that expectation. Uh, it reminds us to do the same with our companies, and it reminds us to do the same with each other. Uh, and, and she, she left with me an impression that, that nothing could be that hard. You know, to this day, and people have seen me adapt. And I think it was Professor Hawking that said, "Adapt in intellect is the ability to adapt." Wasn't, didn't he say that true intellect is the ability to adapt? Uh, in a lot of ways, that, that kind of defines Nvidia, kind of defines me. I, I approach almost everything from the perspective, you know, how hard can it be? Now, often times it turns out to be really hard, but you approach it with the attitude, how hard could it be? And, and if you look at all the things that we've done as a company and what I've done, I, I've never been CEO before. This is my first CEO gig. Okay? And so, I, I, I think I'm, I'm going to get it right someday. Um, but when we first started the company, and I never raised money, I never wrote a business plan. I still haven't written a business plan. Um, and, and I've never been a CEO, never even been a manager. In all cases, I think, I think her, the, the, the fact that she was able to teach me English, and she doesn't speak English. Uh, she can't read English. So, you got to ask yourself, how did, how did she do it? Um, turns out, piece of paper and a dictionary. And, uh, I approach almost everything like that. You know, how hard could it be? And break it down to first principles, and you learn along the way. And, and, and I meant it earlier. Uh, so long as you could stay in the game long enough to learn the sport, and staying in the game is, in fact, most of it. Uh, I, I was able to to do what I'm doing today because I didn't get bored and I didn't get fired. That, I think, was the magic, all, all of it. It's 100% of it.
And I think staying on that topic of Nvidia being the beginning of a lot of firsts for you. It's as if you've read my mind. This is flowing so beautifully. Let's talk about the decision that led to you becoming CEO of the company. So the dream and plan for Nvidia came together over many coffee meetings at the chain restaurant Denny's. Those meetings revealed to your co-business partners, Pri and Malahowski, that you were to be the CEO of Nvidia. What do you think made them unanimously decide that you were the best fit amongst the group to lead Nvidia as its CEO?
I think because they didn't want the job, and they were right. They didn't want the job. And, and, and, um, all three of us were engineers, and, and I think that's the answer. They didn't want the job. Um, in retrospect, I, I could have been smarter myself. And, and, uh, to be CEO is a, is a lifetime of sacrifice. And most people think that it's about leading and being in command and being on top, and, uh, you, none of that is true. You're, you're in service of the company. You're creating conditions for other people to do their life's work. You're, you're inspiring through example. Most of the examples are making difficult decisions during very difficult times. It's mostly about sacrifice. It's about strategy. And strategy, as you know, is not, choose, not just about choosing what to do. It's about choosing what not to do, which is sacrifice. And, and the, the determination, the, the conviction, um, the pain and suffering that goes along with with overcoming obstacles. That's all sacrifice. It's a, you know, being the CEO. I'm not trying to talk anybody out of the, the profession. Uh, if you were to do it, it, it is the greatest honor. But I, you, you have to realize that it's not about, you know, fame and glory. It's mostly about pain and suffering. And, and to a lot of that, um, you know, I, I, I attribute to me growing up. It wasn't an easy, easy, uh, uh, easy journey. Uh, you know, we came to, we went to, we went, we came to America. Went to America. Um, my parents wanted us to pursue the American dream. They didn't have very much. They didn't, they were, they were, uh, quite modest. Um, and, uh, uh, moving to the United States, you know, was quite difficult for us in 19, 1973. But somehow we found, we, we made our way through it. And I think the, the, the life, the life of struggle, endeavor, um, nothing for granted, having, having to earn everything. Uh, that, I think, was good CEO training. You know. And so, I, I think the answer is that they were smart. They didn't want the job, and they live very high-quality lives now.
And building upon what you said about struggle and strategy and personal development, at a recent visit to Stanford, you told an audience of students, "Greatness comes from character. Character comes out of from people who have suffered." In the beginning years of building Nvidia, amongst many setbacks, how did you keep going on? Because you've said that you had never fundraised before at such a high level. You'd never pitched such a brilliant idea. You had never had a business plan, but there was still this idea of strategy and duty and personal sacrifice. I mean, it's one of those, it's, it's one thing to to say it, but to live it is often really difficult. How did you continue to keep the faith in those really beginning stages?
I say it to this day. We believe what we believe. And so, we reason about, we reason about about our perspective of the future. And you reason about from first principles that you learn, um, through education, and, you know, fundamental first principles. And, uh, you have to reason all the way back to either the first principles of computer science, or the first principles of physics, or, you know, whatever first principles you could, you could hang on to. And you try to, you try to reason as far back as you can. Now, once you do that, and you come to the conclusion that all of the, all of the environmental conditions and all the information that you have, uh, causes you to believe in what you believe, then at that point, you believe what you believe. And you got to decide, am I going to be somebody who does something with it, do something about it, or do I just become one of those people that, you know, that say something like, "Oh yeah, I knew that," or "I also knew that," or "I've said that before," but you did nothing about it. And so I tend to be somebody that, that would reason about these things and believe it so deeply that, frankly, I could see it in my head. And once I could see it in my head, as far as I'm concerned, it's might as well be real. Everything else is just details. And so you, you manifest your belief as deeply as you can. And after that, it's, you're kind of hard to, um, dissuade. Now, of course, every single day I gut-check all of the assumptions that I made. All the assumptions that I used to reason about the strategy, I'm constantly re-evaluating. If any of that changes, if any of the the assumptions or any of the the principles that I used were were flawed, um, I'm quick, quick to adapt, you know. So, I'm constantly learning through failure, and I'm quick to adapt. And by adapting, you get to stay in the game. And so, so I think it's, it's not a complicated equation. Often times, oftentimes I think, I think people have a hard time pivoting because they feel that their ego is somehow tied to some decision they have made or something that they said. And that's really hard for CEOs. You know, for CEOs, you, I, I stand up and, you know, as Nvidia today, we have 50,000 people, and I declare something about the future. You know, I, I, I describe what I, I think as a, as a direction the company ought to go, and, and, and you reason about why you do, do so. And you, I do, do that in front of 50,000 people. And I'm doing it constantly. I'm saying it constantly because you can't just say something once, you got to say it a thousand times. After you say something a thousand times, when you discover that in fact you were wrong, it's fairly difficult to pivot. But over time, I, I've earned the right to change my mind. And I, as soon as something is wrong, and I feel that it's, and, and the way we describe it is, we have, we have a phrase in the company, and I, I remind ourselves of that all the time. You, we have to be intellectually honest. You know, if I know that, if I know that, um, we had, we had to change our mind, and I don't, um, that's a, that's a character question, you know, it's an ego question. It, you know, somehow I'm preventing myself from allowing everybody else to, to do the right things. And so, so I, I, I'm quick to get out of that. And over time, I discovered that that leaders are not meant to be right. That's not our job. You know, we're not, our job isn't to be right. Our job is to help other people succeed. It's related, not the same. And so to the extent that they always believe that I want to help people succeed, then they want to help me succeed. When I change my mind, nobody thinks about it twice. It's almost like they never forgot what I said before. You know, this new thing that I say was, "Yeah, he's right." And I change it a little bit. "Yeah, he's still right." And so, you know, people just want you to succeed. And, and, um, you want to create the conditions where you allow yourself to be vulnerable, uh, to change your mind, uh, to be wrong, uh, but because people know that that, um, I always have their best interest, and I want the company to succeed, and I want us to realize the future, you know, they, they'll go along with it.
Wonderful. Let's talk about the GPU. So in 1999, Nvidia creates the light bulb moment in the tech industry when the company manages to offload graphic rendering from the CPU to the GPU. In 2006, you followed up with the CUDA microservices with over 400 libraries. When you're creating and innovating at NVIDIA, how do you strike the balance between building and developing on pre-existing technologies and deciding, like you said, when to pivot into a new, into a new lane? How do you keep that balance?
Uh, it's really hard, and the reason for that is this. When you, uh, when you reinvent something, for example, let's, let's use a phone for example. When the iPhone came along, it's still a phone, and initially it didn't do much more than a phone. Uh, it had a browser, and most people didn't need a browser all the time, and, uh, it had a map that was nice, and it played some music that was nice, but it cost five times more. And so the challenge is, at the time of a new product category, the cost of the technology is much higher than the value that it provides. And the same thing with the GPU. We invented the GPU because we wanted computer graphics to be a medium that was expressible through software. Prior to us, prior to the GPU, graphics accelerators were fixed function. Grow shading with gross shading. Fun shading was function shading. The way it, you know, whatever specular highlight you decide, that's exactly the way it would render it. Well, we decided that that computer graphics should be an artistic storytelling medium, and the software, uh, it should be programmable by software using what is called a programmable shader. So we invented this idea called a real-time programmable shader. And so that, that led to a bunch of future opportunities. But on the day we announced it, there were no applications, but it cost twice as much. And so here's something that you have no need for, and as the current customer would prefer to have something that is, quite frankly, half the, half the price than to have something that has future promise. And so, uh, there's, there's no easy answer for that aside from you've got to believe what you believe. And when you create it, you know, the rest of it is about ecosystem development and, uh, inspiring the, the developers to create applications that realize the potential. Um, a whole bunch of software, you know, that kind of stuff. And that's all, that's all kind of mechanical work. Um, but in order to go do that, you have to believe in a future. And then crossing that, crossing that chasm is extremely painful. It's life-threatening. Most companies don't make it. Most companies don't make it from classical phone to become a smartphone. Notice none of them did. No feature. It's back then we call our phones feature phones. I, I don't know phones. And so they were called phones. And then now they're called smart mo, smart phones. No phone company made it through to the smartphone. Nvidia is the only company that made it from one generation to another generation to another generation to another generation. And we constantly reinvented ourselves. We've now reinvented ourselves through six computing eras. And I, the mechanics of it is fairly mundane. It's easy to explain. Um, you know, I'll teach a course here one of these days. It's a, it's a five, it's a five-step program, and it's not, it's not hard. But the part that is hard, it's the courage to do it because when you leap across to that next thing, at the moment when you're in the middle of that canyon, your, your cost is incredibly high, your value is incredibly non-existent, and very, very rarely do people cross to the other side. And just, that's 100% courage, the ability to endure pain and suffering. Um, the rest of it is, you know, skills.
One of the beautiful things about Nvidia's work is the way in which its discoveries have reverberated across the world and have spurred on lots of different technologies. This year, Bristol University unveiled the Isambard AI, the 11th fastest supercomputer in the world, powered by 5,448 GH200 Grace Hopper superchips. Isambard AI is being lauded for being able to power cutting-edge medical and sustainable research. How does it feel to see the work of the Grace Hopper superchips be extrapolated upon to build what is essentially really crucial technology that is shifting medical research at the highest levels?
Yeah. Um, before, is, so, so that we get the record straight, the fastest supercomputer in the UK built, um, I built, and it was called Cambridge 1.
Did you guys know that? Okay. I never got, I was, I did it because I thought, um, Nvidia was going to have a headquarters here in the UK. We, uh, did you guys, did you guys ever hear about the story? I almost bought a, a, UK company. Uh, it was blocked by the UK, just to, I know, it breaks my heart to this day. Uh, so, so anyways, we almost bought ARM. You guys know that, right? We almost bought ARM, and, uh, I worked on it for a long time. I thought it would have been a great idea. I still think it would have been a great idea. I don't think it's too late. No, I'm just kidding. Anyways, ARM's a great company. Turned out, turned out. Uh, so, so anyhow, you know, what we build is the, is the most crucial instrument of knowledge discovery. And for the very first time, I, we've built a computer that can understand the meaning of the information that it is processing. It's not just processing data. It understands the meaning of the data it's processing. So, for example, it's not processing letters, it's processing words, and it understands the words. Um, it's not just processing, uh, a whole bunch of numbers. It understands that these numbers represent fluid flow. Um, or that it, it understands that this sequence of numbers actually represents a protein, uh, or a small molecule chemical. Um, and it understands the meaning of it, its vocabulary, its meaning. It understands its functionality, for example, you know, understand the, the semantics. It understands the context that it's in, and therefore, uh, it understands how we would react in that context. I've just used a bunch of words that, when you apply the, uh, ChatGPT, is very, very obvious. But remember, the protein and those English words to the computer is the same. To the, to the, to the extent that we can under, we can help the computer understand the nature of proteins, the meaning of proteins, the structure of it, its dynamics. Um, it should be able to understand how it interacts with other chemicals and other proteins in various contexts. We should be able to talk to our protein in the future. What are you? How would you behave? Are you soluble? How do you, how do you behave in temperature, high temperature? How do you behave in different, different types of liquids, in different contexts? Um, how would you react to this particular chemical? How would you bind to it? You could just literally talk to a protein in the future. Now, what I just described sounds a little ridiculous right now, but as you know, you could talk to an image today. Just go up to an image and, you know, what are you? I'm a picture of a cat. What kind of cat are you? Can you move? And, and all of a sudden, the video, the, the image turns into a video. And so, uh, notice, notice this is the world that we're in now. Not only were you processing data, we understand the data that we're processing, and the implications of that in the field of, of course, drug discovery, or, you know, material sciences, or any other, any form of science is really quite profound. So, you know, we, we've created, um, what Turing, you know, imagined, right? Artificial intelligence.
One of the key, um, visions at NVIDIA is about being able to turn biology into engineering. And NVIDIA is, um, alongside other AI companies doing...
How hard could it be?
Turns out, super hard.
We've been at it for 10 years. I'm still working on it. Nvidia, alongside other, um, AI companies, are doing a lot of really good work at studying, um, molecular biology, particularly as it pertains to things like amino acids, um, and trying to not only keep the work that you do, um, limited to computers, but also to do cutting-edge medical research. At what point did you realize that Nvidia possessed the capability to switch into helping with the, with the, with the healthcare industry? And what are your, um, views on the progress that's being made in molecular biology research using AI?
40 years ago, something really amazing happened. Uh, 43 years ago, I was the first generation of engineers that designed a computer inside a computer. Before then, all of the engineers did it practically by hand, prototyped our systems into existence. My generation was the generation of what is called computer-aided design. Well, 40 years later, 100% of what we build exists completely as a digital twin inside a computer before we make it. To the point where, in the old days, we would, we would tape out a chip. We'd go fab a chip, and we'd hope it works. Today, these chips are a billion times more complicated. Tens of thousands of engineers work on it at, at one time. We send it to the fab, and when it comes back, I know it works. And the reason why I know it works is because it's been living inside another computer for a long time. Well, that's 40 years later. I believe the time is about now where we have the ability to represent the various hierarchies of biology so that we can have computer-aided drug design. The idea of drug discovery. I, the, even the word is wrong. Drug discovery. It's like, "Hey honey, uh, I'm going to go look for mushrooms now." And some days you come back empty-handed, most of the time. Some, some days you come back with some amazing truffles. Look what I discovered today. I found it today. Now, of course, drug discovery is a little bit like that. It's much more science and less engineering. If you look at the way we do computer design, design today, it's 100% engineering. And so we have every single year, we get better and better and better at it. Drug discovery is very, very hard. Every single drug for every single disease, almost feels like a brand new discovery, like a brand new journey. We've, we have to create the computer tools and the, the, the information representation, transistors, and logical gates, and, you know, large functions and large chips. We have different languages and different tools to represent the different hierarchies of electronic design. We need to discover the hierarchy of information representation for biological design. Does that make sense? And so once we have that, once we have tools that understand that representation, and we can manipulate it, then the world becomes drug design. And every single year that goes by, we become better and better and better at it. We stand on the top, you know, on the shoulders of giants. And one of these days, you know, who knows? And so that was, that was, I just reasoned, I literally just did just now what I did 10 years ago at Nvidia. And I explained this to a bunch of computer scientists, and then everybody goes, "Okay, well, let's go give it a try. How hard can it be?" And so it's been 10 years, and I got to tell you, it's super hard, but it's okay. But I still believe it. I, I have 100% certainty that we will discover the representation of biology and the tools necessary to go design it. And the reason for that is this. Although we, although we are, we're all a little different, we're largely the same. So there's obviously structure in biology, and obviously we, it's, it's a repeatable thing. Not only is Nvidia crossing over in terms of its legacies into different fields, it's...
By the way, I just wrote the entire business plan of Nvidia going into drug discovery, and this business is now for us billions of dollars, and I just wrote the entire business plan just now. No number, no, no spreadsheets were used. No numbers were applied. No calculus was necessary. Just like that. All simple reasoning. I did the same with autonomous vehicles. I did the same with robotics. I did exactly the same for artificial intelligence. Just like that. We sat there and reasoned about it step by step by step. Um, usually we have the benefit of a whiteboard. It makes it easier to to communicate. After that, somebody takes a picture of it, the company goes off, and 10 years later, in the case of artificial intelligence, 15 years later, here we are.
Well, I mean, let's talk a little bit more about artificial intelligence. So, not only is the work that Nvidia is doing crossing over different fields of science, but it's crossing across borders. In September 2025, Nvidia announced a $2 billion pound investment in the UK AI ecosystem alongside firms like Baldin, Phoenix Cord, Excel.
I did it right in front of the Prime Minister.
Kier Starmer.
Yes.
I mean, Kier Starmer championed this.
I'm investing your company. I'll invest in your company. Their $2 billion went really fast.
Not only did Kier Starmer champion it, it was also recognized by a bunch of other business leaders who thanked Nvidia for its bold leadership. How does it feel to be able to support burgeoning, um, leadership in AI and creativity and innovation, particularly in the UK, and also across the world?
Nvidia succeeds when other companies succeed. Remember, we're a platform company. We're a tools company. We're a platform computing platform company. Nobody wakes up in the morning, okay? Nobody in your family wakes up in the morning said this morning and go, "Guess what we need? We need to buy a computing platform." Nobody says that. And so the only reason why we succeed is because companies and developers who use our platform create something incredible. That's how we succeed. And so, so in a lot of ways, I've got one of the, the really, the one of the greatest jobs ever, which is to have a company whose mission is to desire other people to succeed. And this is a phrase that is always inside the company. I'm always reminding everybody in the company that that we want and we need and we desire other people to succeed, and it's through their success we get to ride their, their coattails. And so that's, that's it, you know. And so when I, when I saw that the UK, um, really is a, has a Goldilocks moment, you know, this is a, you, you have incredible researchers here. This is the home of computer science. If, if there's a country that represents computer science, if that's got to be the UK, you've got, you've got, uh, a rich, uh, ecosystem of entrepreneurs, and the only thing that you don't have, as it turns out, is the instrument of knowledge, the instrument of science, the instrument necessary to do that. And I, I know how to do that. And so I felt that that, um, with our involvement, we might be able to create that spark. And, um, and then maybe the UK becomes, you know, one of Nvidia's headquarters after all, you know. So, but, but is this, it is, uh, this, you, you really have an extraordinary moment. But for some reason, I do feel that that the, the culture of, of the UK, you're so modest, you know, you know, in, in Silicon Valley, what, however great we are, we describe it as a hundred times more than that. And however, but here in the UK, however great you are, you're one-tenth that. And so, so, um, I'm here to tell you, you, you're pretty extraordinary. And, and look at the list of, of, uh, of the inventors and the scientists and discoverers that have come before you and that have, that have inspired us and, uh, has, uh, has made it possible for us to do what, what we do. And then just remember, the industrial revolution was invented here, was discovered, it was created here, and there's a new industrial revolution now. So take advantage of it. It's got all the, it's got everything that, that you're, you're part of, and everything that you're great at. And so, this is your moment. You've got to, you've got to capitalize on it.
And you've spoken earlier on in this interview about what makes the UK such a promising, um, field for AI development. Um, let's talk a little bit about South Korea. So Nvidia has just committed over $260,000 of its most advanced chips to South Korea in what is being called a $10 billion mega deal in AI. What would you say is what makes South Korea an ideal partner for this level of investment, and what are some of the promising projects that you hope that Nvidia might be able to achieve in the, in the area region?
South Korea has a, uh, is one of the major regions, major countries in the world that has industrialized deeply. Uh, South Korea, as you know, manufactures chips. They manufacture ships, chips, ships, cars, electronics. Uh, this is a country that is incredibly good at industrialization. In, in just a few short decades, um, they reinvented the whole nation and became a global industrial giant. Uh, South Korea also has, has a unique capability of doing software. And so the technology ecosystem is able to deal with, um, hard hardware and manufacturing on the one hand, but software and the artistry of of developing software on the other hand. And so I, I think the, the, it's a real opportunity to see them, uh, take advantage of artificial intelligence and reinvent how industrialization works. You know, this is the next era of AI. Currently, AI understands language and numbers and images and videos. But in the future, in the near future, we need AI to understand the laws of physics, understand causality and object permanence and, you know, inertia and, you know, that needs to understand gravity and things like that, friction and things like that. And so, so I think the, the, um, when, when that happens, and when we create embodied AI, AI that could operate, manipulate, and operate in the world, then, then we're going to reinvent, reinvent how industry works. And, and as you know, the world has a ve, quite a severe shortage of labor. The world's GDP would be much higher today if not for the fact that we had a shortage of labor. We have shortage of labor in just about every single company, every single field. And so although we talk about about jobs being lost, it is very likely that all jobs will be transformed, and it's very likely that AI is going to drive a product, productivity gains and, and GDP growth like we've never seen before. And that's, that's certainly the hope, and I, and I believe in it.
Just the last question for you before we hand over to the floor in the live audience Q&A. Many of the young people sat here today, like you, have a passion for science. Cambridge is a university much like your alma mater Stanford that has produced 126 Nobel Prize laureates with an exciting AI development scheme. The city has produced a culture of many eager to turn their ideas into businesses and to top firms, as you have successfully done. What is one piece of advice you would share to many of the young people who've got an idea, who've got a bullet point, who've got a bit of code that they're excited about? How would you encourage them today?
There are some skills. There are some skills in determining whether an idea is a good one or not. If you want to do something, just go do it. Um, and so long as you're, you're, you're intellectually honest, you're contextually alert, um, you're environmentally alert, understanding what's going on around the world and, and your industry, uh, and you're willing to pivot, then, then that's all fine. Okay? But nonetheless, there are skills in determining, uh, upfront whether an idea is worth pursuing or not. But I, I would say that the singular best piece of advice for entrepreneurs, and it's served me incredibly, is to have this childlike view of the future, which is optimistic. You're curious, and you ask yourself, you know, "How hard could it be?" And don't let anybody tell you that in fact it's really hard. You're going to find out for yourself that in fact it's really hard. Um, but you have plenty of time to do that. You have plenty, plenty of time to do that. And so people have asked me, "Would I start the company again if I knew everything today back then?" The answer is, of course not. It's too scary. It's too painful. Too much sacrifice. How could you take all of the, you know, all of the feelings and everything that I've learned, all bottled up at this moment, transport it back into a 29-year-old body and go, "Hey, with that, get at it." You're never going to do it. You're never going to do it. And so, don't, don't be afraid of ignorance, and don't be afraid of being naive, and don't be afraid of, you know, all, all of those things. You're going to learn everything you need to learn along the way. And just, you know, if you, if you're really passionate about something, go do it. You know, just tell yourself, "How hard can it be?"
Thank you so much, Jensen. And on that note, we are going to open the questions up to the floor. So, if you do have a question, please stick your hand up nice and high. And then when called upon to... My word. And then when called upon to do that, we will get an EM to bring you a microphone. Um, let's start right here at the front with the gentleman who's got his hand up. And then afterwards, we will go upstairs to the man in the dark blue jumper. So, we'll start here and then go upstairs.
Jensen, your point about, uh, cooperating with each other and not competing really resonated. Um, but here at the university, we spend a lot of our time training students for exams, uh, marking them, ranking them, and so on. What would your advice be to Cambridge University? Should this 800-and-something-year-old institution abolish exams and have a more cooperative way of building up our student base so that we can adapt to the future which is coming with AI?
I, I guess, I guess if, if there's no, um, ranking and rating, uh, it would be hard for you to know whether the problems and the curriculum is even hard enough, that is sufficiently challenging to to push the students. I get that. Um, a long time ago, when I first started the company, there was a concept in management called ranking and rating. It was created by another Silicon Valley company, and, and because, because that company was, was famed for being incredibly good at management, um, I took a lot of the, the learnings that were being, being talked about, and I applied it at Nvidia. I can tell you today, I've dropped 100% of them. There's no benefit in ranking people. Uh, there are ideas called, uh, even back then, 360 peer review. There's no benefit in that. There's no benefit in asking people to to rate and review you. Um, I, I found that most of those techniques don't really work. Um, but in the context of academia, frankly, with AI, I think the idea of information being somehow, uh, unaccessible so that it's even worthwhile to compete anymore to get the right answer. I think those days are kind of gone. Um, I, I do wonder, I do wonder how, I wonder if the way you teach courses are going to be very similar to the way we renew ourselves at a corporate level, continuously learning, continuously renewing ourselves. I'm, I'm quite famed for saying I don't fire anybody. And, and the reason for that is very, is, is grounded in in wisdom. We want to encourage our employees to innovate, which requires them to take risk, which requires them to be vulnerable. They're going to do something that puts themselves out, and they'll likely fail. And if they're taking sufficient risk that is proportional to the type of company that we want to be, then they should fail often. But if, if we let go of people that we rank to the bottom, the bottom 5% rule, you might have heard of that management technique as well. Every single year, remove the bottom 5%. Because if you do that, then what's left in the soup, you know, is consummate. Okay? But that's complete nonsense because we love stew. We want, we want the mess of a stew, not the purity of a consummate. And so in fact, we don't want that 5% to be lost because they, they happen to be precisely the people who had just taken a risk, failed, and they learned something from it, and they might tomorrow be the outlier that reinvented something to save your company. And so I think the, the old cultures, the old cultures, the old systems deserve some reevaluation. And certainly, because intelligence is now a commodity, we got to say that kind of out loud. Intelligence is about to be a commodity. Then what's left? And what's left is a lot of the things that we were just talking about. It's courage. It's intellectual honesty. It's the absence of ego. The ability to be vulnerable in public so that you can create. You know, as you know, artists and inventors and creators, they're humiliated, laughed at often because what they do isn't always perfect. And so, you need to have that, that humility, the vulnerability, and yet the courage to put yourself out there and be laughed at. And so, so, um, some of those things I, I think are going to become more important.
Can we take a question, um, to in the greenest jumper right there?
Yeah, indeed.
Hi Jensen.
Hi.
Um, thank you for coming to Cambridge and congratulations again on your fellowship.
Thank you.
My name is Josh. I'm in Trinity College.
Um, you mentioned earlier that you think that the future of jobs are going to change, um, as a result of AI, and a lot of young people, aspiring lawyers, accountants, consultants, bankers, they don't really share that optimism. I think they're more pessimistic about it. So, I was hoping you could just maybe shed some more light onto what you think that job transformation will look like and what the benefits will be that will come from it.
Okay. Um, one is, is, um, uh, optimism in humanity. One part of the answer, one part of the answer is, uh, pragmatism, and then one part of the answer is evidence. Okay, I'll start with evidence. Radiology was going to be the first industry to be completely destroyed by artificial intelligence. And in fact, almost every radiologist today uses AI. And yet, the number of radiologists being hired has increased. Why is that? Because they can now do more things. There are so many cases that went undiagnosed using radiology, and there's so many cases that were not deeply diagnosed because there's just a bottleneck of how many radiologists could study all those, all those images. And so now the basic stuff is done very, very quickly. Now, the number of cases that they're getting, the depth of the cases that they're getting, um, is much, much more, much more interesting as a result. More, more people getting hired. Uh, the, uh, pragmatic part of it. Uh, let's say that you have a job, and, uh, within your job, part of it is doing some task that, however you describe that task, that task became infinitely fast. Okay? And, and it used to take you a week to do something. Now it takes you a second to do it. Well, my question to you is, based on that, what is more likely to happen? That you now have more time to enjoy coffee, or you just became busier? It is likely you just became busier. It's just another example of the radiology. The reason for that is because
That thing that you were doing is causing you to not be able to do all the other stuff later, be past that pipeline, you know, past that workflow or past that task. Now that task is done infinitely fast, all of a sudden the answer came back to you. You're now the critical path again.
It's no different than if I were to issue an instruction to my my my group of people that I'm working with and they're supposed to go off and do a research or do some analysis or simulation and bring that answer back to me so that I can make the next decision. Today, it'll take a day or a week for them to do that. Meanwhile, I'm doing something else. But if it takes a second for them to come back, I'm back in the critical path again. I'm busier than ever.
Now, the reason for that is because I have so many ideas. We have so many ideas to go pursue that when the tasks become super fast, it turns out we could do more things. We become busier. There's some evidence that smartphones made us busier, not less busy. There's some evidence that computers made us busier, not less busy. And so, it's because we're we are now increasingly the critical path and because we have so many ideas, we can pursue more things.
Now, I just hope in humanity, we always find a way to discover new things to do, new things to be busy. And and I'm hoping that that as a result, um we'll go off and work on the problems that are that are the most meaningful, the most valuable in all of our work, which is the ones that are poorly defined. The poorly defined work is the most valuable of all work, discovery work, creativity work, the original creativity work. I'm not talking about change the cat, you know, to a cat with shag r shag rag, you know, shag shag rugs, you know, and is I'm talking about really create creating something out of the box. And and so the the ability for us to use AI to solve problems that are fairly easy to des well not it doesn't have to be even easy that this describable um so that we could go and work on things that are very difficult to describe I think is is really quite powerful.
>> Thank you. Shall we take the question from the member in the gallery?
>> Hi um hi Jason my name is
>> Every job will change. You will not lose your job to AI. You will lose your job to someone who uses AI.
Thanks for that. Uh, hi Jason. My name is Lily. I'm a EMBA student at the Jaji Business School. I also run my own venture builder called FunX. Um, it's very interesting that you just mentioned the UK's position almost having this Goldilock moment because we commonly believe the AI game is actually happening in China or happening in the US. So what do you think the UK or the UK startup community can do more to leverage its current position and benefiting more from the AI game?
>> Regulate less.
>> And I I mean that I mean that deeply and and uh uh genuinely for the best interest of the UK. Uh you I know I know that I know that regulators um and because so much so much of regulators are are lawyers which is good um but because they want to protect us they want to pro protect society um they could regulate too early especially on technology like this where it's hard to predict the future. You could watch sci-fi movies but that's not the future. That's called the sci-fi movie. And and to use the the science fiction movies projected into reality um and and um through some words of some people uh cause cause social panic um to the point that you cause you you overregulate um you're stifling the UK's ability to innovate. Uh the fact of the matter is China as you know is a under technology is underregulated and the reason for that is because the leaders in China mostly are engineers. The leaders in United States are mo mostly lawyers and so you can kind of tell what's going on. Uh and the the rate of of technology evolution and industry evolution in China is just running incredibly fast because they regulate late. They wait until the problems show up and they regulate the problems. Uh they solve they create regulations the way engineers solve solve problems. You know let's not dream it up. Let's let's observe the problem. Let's understand the root cause of the problem and then solve the problem. And um and so anyhow I would say under regulate less.
>> Jensen thank you very much for this evening. I know that we have so many questions but in the interest of time I think it's fitting for us to conclude there. But I want to say a huge thank you from us um from all of the members gathered here today and on behalf of the selection committee. Congratulations again on being the twins.
>> Thank you very very much.
>> Thank you.