POWERING THE AI REVOLUTION: A CONVERSATION WITH NVIDIA'S JENSEN HUANG
AHMAD RASHAD: Welcome back. One of the biggest stories in tech is the explosive growth of NVIDIA. And today we have the opportunity to dig deeper into what the company's breakthroughs mean for all of us.
Please help me welcome its founder and CEO, Jensen Huang.
(Applause.)
Who will be interviewed by one of the most respected financial journalists in the world, known for insightful conversations with leading investors and innovators. Please welcome CNBC's Becky Quick.
(Applause.)
BECKY QUICK: Thanks, Ahmad. Good morning, everybody. It's really great to be here with all of you.
Most of you know Jensen Huang. I don't know if you know his whole story, but let's set this up.
Jensen Huang is the head of a $3.3 trillion company. It is the second most valuable company on the planet. It's got a market capitalization that is greater than that of Boeing, Caterpillar, Goldman Sachs, Home Depot, IBM, Johnson & Johnson, Coca-Cola, McDonald's, Pfizer, Disney, and Walmart combined. It sits at the center of the AI boom.
And while Jensen isn't the father of AI, you could consider him the general contractor. NVIDIA's chips power everything when it comes to machine learning. Companies like Tesla, Oracle, Microsoft, and Amazon are all vying for their share of NVIDIA's limited supplies of really cool semiconductors. In fact, Larry Ellison described a dinner at Nobu last fall between him, Jensen, and Elon Musk as basically, in his words, "Me and Elon just begging him for GPUs."
It is a long way from the Silicon Valley Denny's, where, just over 30 years ago, Jensen and two of his buddies sat down over Grand Slam breakfasts and bottomless cups of coffee, and they drew up these plans for a company that they were hoping would build the best PC graphics chips on the market.
They thought about naming the company Envision, but that had already been trademarked by a toilet paper company, so they went with NVIDIA instead.
What you may not realize is that the semiconductor business is a second act for Jensen. His first was as an athlete, and he has something in common with LeBron James. They both graced the pages of Sports Illustrated when they were teenagers. For Jensen, the year was 1978, and the sport was table tennis.
Jensen, I think I would love to ask you what you learned playing table tennis that translates to business.
JENSEN HUANG: Well, first of all, that's such a nice introduction; I would like to leave right now.
(Laughter.)
That's probably all I had to say. But before we start, happy Valentine's Day.
BECKY QUICK: Oh, thank you, Jensen.
(Applause.)
JENSEN HUANG: Hey, come on, who doesn't have a crush on Becky? We watch her on TV every day.
(Laughter.)
Okay, all right.
BECKY QUICK: So what did you learn playing table tennis that actually --
JENSEN HUANG: I should probably leave right now.
BECKY QUICK: Yeah, we both should, because it's not gonna get better.
(Laughter.)
JENSEN HUANG: What did I learn? What did I learn? What did I learn?
I love the life teachings of athletics. Now, being good at sport, being good at anything is some talent, enormous amounts of hard work, and it's all character. It's a hundred percent character.
And I think the best athletes, the best characters, the best winners, the best companies, the best employees, they all have exactly this exact same thing. They have some amount of talent, but they work incredibly hard.
And somehow they're resilient, and they take setbacks incredibly well. They're willing to take -- learn from mistakes, and they love coaching, and they love to -- they're self-critical, and they're thoughtful about the way they practice. And it's not just hard work. It's, you know, smart hard work. And somehow the combination of all that creates champions.
BECKY QUICK: I was thinking about that, too, a lot of the things that you mentioned -- competitiveness, hard work, resiliency, grit -- all of those things are things that you see; that's what it takes to rise to the top of the NBA.
And since we're here, I just wonder if there are any NBA coaches, any NBA players that you kind of associate with.
JENSEN HUANG: Well, I don't know enough about -- you know, I'll just be -- I'll confess now. I haven't watched a game in a long time. However, the last time I saw a game, it was, of course, the Golden State Warriors.
And my first observation is they should probably change the game because Steph Curry, the three-pointers is so routine now, you know. I think there needs to be a four-pointer, and I think the game would be a lot more interesting if -- don't you guys think? Just, you know, from half-court, it's four points.
And so, anyways, as I was watching it, I think they need to change the game. And they did that for table tennis. And so table tennis now, the nature of table tennis has changed over the years because people have gotten a lot more athletic.
BECKY QUICK: What did they change? I didn't even know that.
JENSEN HUANG: The ball is bigger; make it harder for you to win right off the bat.
BECKY QUICK: Oh, does it?
JENSEN HUANG: Yeah, and they also -- the bat now has two colors because -- I don't want to get -- I don't think anybody here wants to learn a lot about table tennis.
BECKY QUICK: Let's talk about resiliency because you came to that the hard way. I don't think you ever had anything handed to you. You were born in Taiwan. You moved here with your brother when you were a child, before your parents came over.
And your parents sent you to what they thought was a pretty fancy boarding school. They thought it was an elite boarding school. It turned out that it was a reform school, and you had some hard knocks along the way. What did you learn there and that message of resilience?
I mean, you went to Stanford, I think it was a year or two ago, and you told the kids there that you wish them -- I hope a lot of pain and suffering happens to you.
It's kind of an odd blessing, and I just wonder if you could explain that.
JENSEN HUANG: Well, so first of all, that school had excellent marketing.
(Laughter.)
BECKY QUICK: The Kentucky school.
JENSEN HUANG: The mission, the school says, this is a school where we welcome everyone, the troubled kids, the broken homes.
My roommate, I won't say his name just in case he's still alive, but my roommate, I was 9, he was 16. And somehow a 9-year-old and 16-year-old were bunking together. The dorm room has a door, but there are no drawers, no closet doors, because you don't want things to be hidden. It's kind of like a prison.
And the two of us were getting ready for bed, and it was the first night that we met each other, and he took off his clothes, and he had all these stab wounds all over him. He had just gotten there, and he was in a fight. And he was 16, so he couldn't -- you know, they couldn't put him anywhere else.
And so this is a school where they welcome everyone, including international students. It was the most affordable private school, boarding school that my parents could find, and it happened to be in Clark County, Kentucky, and Oneida Baptist Institute.
I still exchange emails with the president of Oneida Baptist Institute, and I still get emails from the students that graduated from there. You know, they tell me that I'm one of the most successful students that graduated from --
(Laughter.)
BECKY QUICK: You're one of the most successful students that ever graduated from anywhere.
JENSEN HUANG: -- Oneida Baptist Institute.
(Laughter.)
JENSEN HUANG: And so, anyways, I'm a very proud graduate of Oneida Baptist Institute.
And so when we were there, I was 9, my older brother was 10, and everybody had chores. There were 120 boys, 120 girls, or something like that, and the boys' chores after school was to go work in the tobacco farm. And my job, since I was the youngest kid, my job, I didn't have to go to tobacco farm, but I had to clean all the bathrooms.
And just so you know, 120 kids could leave quite a mess. And so I've seen a lot of things that no one should see, and you can't get it out of your head.
But anyhow, I did that for two years. And I didn't know any better. I thought that that was my job, and other kids' job is to go tobacco farming, and so I just did that. I think a lot of -- and I tried to do my best, and I, you know, did a good job.
But I think that it was a tough school. It was 1973. So the world, you know -- Kentucky, Clark County, Kentucky, the epicenter of the opioid crisis today. And so it's a pretty tough, tough place, and kids were tough.
There were town kids. They carried pocketknives. Everybody had a pocketknife. Everybody smoked. 100 percent of the kids smoked. I smoked for a week. I had better use of the money at a quarter. And so I saved it for popsicles and things like that.
But it was hard work. And so, anyways, I think I've always been used to hard work, and I don't mind it.
BECKY QUICK: But when you wished that to the kids at Stanford, you think that that's what it takes to really --
JENSEN HUANG: I think that kids at Stanford have -- they're all talented. They're already rich. They obviously are there for a good reason. They're top of their class. And so they're at the top of their game when they're 17 years old. They're at the top of their game at 17, 19 years old.
And so the question is: How do you become better? How do you go from being top of your game at 17, and you're talented, you're smart, you're the best top of your class, how do you go from that to great?
And to go from that to great is character, suffering. It's pain and suffering. And as all the athletes know, you don't get to the top 0.01 percent of your field without extraordinary amounts of sacrifice. And so you have to learn how to sacrifice, learn how to endure pain and suffering, and still love the sport and still love your work and still have a positive attitude.
And so somehow, you know, it takes a special -- I think a special kind to simultaneously suffer all day long and somehow enjoy it.
BECKY QUICK: I think there's a word for that.
JENSEN HUANG: Enjoy suffering?
BECKY QUICK: Yeah.
JENSEN HUANG: Yeah. But I think it's because you believe it so much and that you know it's part of the journey.
And so I think athletes learn that pain and suffering is part of the journey. Most students don't know that. Computer scientists don't know that. They just don't know that because they haven't endured going from a good computer scientist to a great computer scientist.
BECKY QUICK: Let's talk a little bit just about failure because a lot of companies, a lot of CEOs won't admit if they've made a mistake, if they failed at something. You not only admit it, but you embrace it. And it's not just for you, it's for everybody who works at NVIDIA. It's an interesting management style to embrace failure.
You want to describe how you use that as a management style and why?
JENSEN HUANG: The single most important part of leadership is to empower others to make good decisions. NVIDIA is not only one of the largest companies in the world, we're also one of the smallest companies in the world. Only 36,000 employees. We're smaller than every single company you mentioned just now.
BECKY QUICK: Except by market cap.
JENSEN HUANG: Yeah. But in terms of head-count. And the reason for that, I think, is because I spend all of my time reasoning. I spend all my time. Every single meeting is about reasoning. Now, what did we observe? What did we do? How could we have done it differently? It's about reasoning, reasoning, reasoning.
I start from first principles every single day. Every single major decision, I start from first principles. I reason my way through it, and then whatever the answer comes out to be is what the answer comes out to be. A lot of people start with the answer. I'll just start from first principles, and I reason through it.
The reason why you do that is because you're trying to enable them to see the world through your lens and to demystify the work that we do. And by helping people watch you go step by step -- and my meetings are never one-on-ones. I don't do one-on-ones ever. And so this is the only one-on-one --
BECKY QUICK: Yeah, you have 50 or 60 direct reports?
JENSEN HUANG: I have 60 direct reports. And constantly we're together reasoning through things.
And the reason we want new college grads, different organizations, they're all sitting in my -- I don't have an office, I just sit in the conference room. The conference room holds about 20, 30 people. And so we're constantly reasoning through things.
And the best thing you want to reason through is a mistake. And so you say, okay, this is what happened. I did this, I suggested this. Let me tell you why I was wrong.
BECKY QUICK: Give us an example.
JENSEN HUANG: Let's see. Gosh, Becky. That's kind of personal. There are investors out here.
BECKY QUICK: We're among friends.
JENSEN HUANG: How would I give you an example? How would I give you an example? Let me think of one. Let me think of one.
But mostly my meetings are about what should we do, let me reason through it. Mistake, what do we learn from it, and how do I reason through it. And so that process of reasoning and explaining empowers everybody else. And so as a result, the company becomes really agile.
And you want people to make decisions. You know, the company, even 36,000 people, is making a million decisions a day. And statistically you want those million decisions to point in the direction of good. You can't guarantee every one of them will, but you want to statistically, for the vast majority of them, point in the direction of good.
And so culture matters. The skills of reasoning matters. Understanding where the company wants to go matters.
And so I spend all of my time being transparent, reasoning about things, explaining things to as many groups as possible. So during the course of a day, I might have 10 meetings, and each one of the 10 meetings have 20, 30 people, and just do that. And they go off and they do 20, 30 meetings, boom.
BECKY QUICK: So it's creating a company culture.
JENSEN HUANG: Really, really incredible. Information moves incredibly fast.
BECKY QUICK: All right. So let me ask you, could that happen at NVIDIA without you?
JENSEN HUANG: Sure. Now it can. Now because it's just part of the culture. You know, our culture is not hierarchical, it's transparent. And it's a reasoning culture. It always goes back to first principles. It's not command and control. It's heavily influenced by, of course, the reasoning and the culture and things like that.
And we have amazing people that are new college grads, working in different offices around the world. We just empower everybody to do smart things.
BECKY QUICK: I'm going to circle back to the mistake at some point, so you're not off the hook on that.
But this is a crowd that likes statistics and records, and you have a lot of them. For example, you have the biggest market cap gain in one day ever. That was $327 billion in July, I think July 31st of last year. You've also got the second biggest market cap in one day ever, too. That was $218 billion in May of last year. And you've got the biggest one-day market cap loss ever, too.
JENSEN HUANG: Come on, nobody has ever lost a trillion dollars.
BECKY QUICK: It was like $600 billion last month. That came about because of DeepSeek, this Chinese AI model that hit out there, where they said that they had come up with this much more efficiently, using far fewer chips.
It rattled all these investors who said, Oh, no, the market cap -- or the market, the CapEx is going to stop, people won't be spending as much on all of these things. So your stock and a lot of other stocks fell pretty precipitously on that one day.
You have lived through a lot of tectonic shifts. And I know you can't talk about the financials or the quarter, because you're in a quiet period right now, but let's talk about these broader tectonic shifts. Do you think that this is another tectonic shift in the industry?
JENSEN HUANG: Yes. And it's an exciting one.
There was a misunderstanding that the way AI works is you take all of the data from the internet. Well, first of all, that's kind of a miracle.
The invention that the company made, it took about 30 years, we created a whole new way of doing computing. It used to be hand coding, running on CPUs. Now it's machine learning, running on GPUs. And the software that machines write are very different than the software that humans write.
And so this transformation took us about 30 years to do. Now it's very clear this is the way computing is going to work in the future, which is the reason why NVIDIA has done so well and so many industries have now jumped on and so many applications have jumped on.
But one of the important applications of machine learning, of course, and the big breakthrough is we made computers go so fast. This is just an extraordinary thing. We made computers go so fast that one day a researcher said why don't we just take all of the data in the world -- could you imagine that thought? Just let's just take all of the data in the world and let's give it to this computer and let the computer figure out what human intelligence is.
And that's the big breakthrough. And that observation, that you could literally take all of the data in the world, is called pre-training.
Well, the observation, of course, is that, you all know, pre-training is a little bit like reading every single book that you've ever found, watching every single video, listen to every podcast. Well, that's the first part. That's called pre-training.
The really interesting part of intelligence is post-training. It's when you take all of that, now you understand vocabulary and syntax and you understand grammar, you understand relationship between images and words, like image of a cat and the word C-A-T. You understand basic math, you understand algebra, you understand geometry. You understand the basics of a lot of things.
Now the question is are you intelligent? No, you're knowledgeable. Intelligence is solving problems.
And so this is post-training. This is the big idea of where we are today. The big idea -- so just let me give you a statistic. So pre-training could be something like we'll train on, we call them tokens, maybe 20 trillion tokens.
Well, post-training is like your brain doing thought experiments. It's like you doing -- practicing algebra or geometry or trigonometry or calculus before you go to take the test.
It's no different than athletes. You're taking free-throw shots continuously and you're practicing. And so the practice is actually where most of your calories go, not the game. And so the practice, that's called post-training for AI. We have this AI going, we use reinforcement learning, we use technologies like search and beam search and best-of-n.
Anyway, so there's a whole bunch of technologies that we use for basically practicing. That's called post-training.
Well, during pre-training we could generate, I just told you, about 20 trillion parameters, 20 trillion tokens of bits of information. During post-training it could be 100 trillion. So the type of computers we need for post-training is actually even larger than pre-training. That's the part that most people don't realize.
And so when the first thing that somebody said was we ran out of data on the internet, it's true. We've taken a lot of data on it. We've run out. But post-training is where a lot of the AI is going to go.
And then what DeepSeek did was they said, We made an AI model a lot more efficiently. And so, anyways, there's a fair amount of controversy about what was counted. They did create a really extraordinary model. It did use a lot of computers.
But the big breakthrough of DeepSeek is that it took something -- it's a reasoning model. It's fairly state-of-the-art. It's kind of, you know, ChatGPT-4o level. And they open sourced it. That is the big breakthrough.
As a result, every single company, every startup, every computer company got so super jazzed about it, we got super jazzed about it, because now you could use a reasoning model of that caliber in so many different applications, from healthcare to, you know, scientific discovery, all the way to robotics and self-driving cars.
And so these applications have now been activated because of it. That's the energy. Wall Street reacted in a way that was surprising, frankly, to all of us. But, you know, those kind of things get sorted out over time.
BECKY QUICK: Well, we have since seen that all of these big companies, whether it be Microsoft or Alphabet or Amazon, down the list, Meta, all of them saying that they're going to spend even more.
JENSEN HUANG: Yeah, we need bigger -- we need bigger computers.
BECKY QUICK: I think Ruth Porat just said the same thing in the green room.
JENSEN HUANG: Because of post-training, thought experiments, training, practicing, smart practicing, reinforcement learning, all of coaching. You know, it turns out, surprisingly, AI is trained -- we call it reinforcement learning, human feedback. That human feedback is basically a coach mentoring and giving feedback to the AI.
BECKY QUICK: I mean, you touched on this just now, but there were a lot of tech leaders at this AI summit in France this week who said that they don't think DeepSeek did things as efficiently as maybe they said, you know, whether it was borrowing from what other AI models have already done or whether it's having more chips than they said they were using on these things. What do you think about it?
JENSEN HUANG: They did a lot of great things. And so I don't want to take anything away from all the smart ideas that they implemented. And it is really a very nice piece of engineering. And so I think that that's probably where to start.
I think we probably, you know -- some of the text and some of the lore was probably exaggerated. But I hate for that part --
BECKY QUICK: To be the headline?
JENSEN HUANG: -- to take away from the great engineering that they did.
And I wish that, when it's all said and done, what we recognized is a really fantastic team of researchers did some really, really extraordinary work. And we should be proud of them, we should -- yeah, we should celebrate that.
You know, one of the aspects of -- we were talking about leadership earlier in culture, you know, one of the things that I'm really proud of is a genuine --
(Phone ringing.)
BECKY QUICK: Hello?
Somebody's phone is by itself.
JENSEN HUANG: It's okay, it's okay.
-- a genuine joy in watching people succeed. And this is inside our company. We teach that all the time. You know how when somebody gets promoted, it kind of sucks the life out of you?
(Laughter.)
BECKY QUICK: You mean if you're competing with them?
JENSEN HUANG: Inside a company.
(Phone ringing.)
JENSEN HUANG: It's okay. He just wants to know how many shares to buy.
(Laughter.)
Just somebody just text it back to them: As many as you can.
(Laughter.)
And and so somebody said to me: Every time something good happens to my neighbor, you know, a piece of me dies.
(Laughter.)
Is that a phrase from one culture? I think one of the cultures I'm looking at, it actually is one. And we teach exactly the opposite. And I go out of my way, like some of my leaders in the room, almost every single company meeting I'm helping people understand that when something happens to our colleagues or a supplier of ours -- you might have heard this before too. Some CEOs will monitor the stock price of all of their suppliers, and when something good happens to the supplier, they go and beat them back down.
I genuinely am excited to see my suppliers do well. And I'm genuinely excited when I give a keynote -- and somebody actually observed this. When I give a keynote, everybody's stock price goes up. When somebody else gives a keynote, my stock price goes down.
And and I'm super excited about helping other people succeed and celebrate their success. And this is really important inside cultures. And I believe this to be true for teams.
When something good happens to a colleague, something good happens to a teammate, you teach all of them to be happy for them. And yet inside companies, when somebody gets promoted, it sucks a little life out of you.
And so we got to take -- you know, inside our company, we turn all of that stuff around. And the joy of seeing other people succeed is great for culture. It's great for being a company. It's great for an ecosystem. It's great for building success.
BECKY QUICK: So you are a cheerleader, or at least Switzerland, when it comes to the wars and the battles that are taking place over open AI. But your customers don't always all get along that well. And I'm thinking of Elon Musk and Sam Altman right now and the battle over open AI.
You know both of them well. They're both important customers. How do you think that's going to play out?
JENSEN HUANG: I like both of them incredibly.
BECKY QUICK: So how do you see that playing out between them?
JENSEN HUANG: I think that intelligence is one of the largest industries ever. We're in the process of automating intelligence. It's one of the hardest problems. It's one of the largest industries. And it is always better to think of the world as not a zero-sum game. And it's proven not to be.
And, however, the problem of competition is that when you win, somebody loses. But that's just about that game. In fact, what is really great about the NBA is that all of teams want NBA to succeed. Because when NBA succeeds, it's great for all the teams.
And so it's good to compete, but not hate each other. It's good to realize that the competition is just that moment. Ultimately the competition is NBA versus some other form of entertainment. Does that make sense?
And to realize that the NBA could be a lot, lot larger than each one of the teams in that moment will --
BECKY QUICK: So you're Adam Silver.
JENSEN HUANG: And so in a lot of ways we need all of them to realize -- and, you know, we're the only AI company in the world that works with every AI company in the world. And I like that position. And I help everybody the best I can because -- you know, even sometimes at our peril in the near term. Because in the long term I'm kind of thinking exactly the same thing as I said about the NBA, that the world is much, much larger than that.
And even sometimes I've used a phrase called rough justice. And many of my partners will say that sometimes I get a little bit better part of the deal. Sometimes I get a little worse part of the deal. And to me, so long as we're going to be partners for a long time, we're going to work together for a long time, over the course of our arc of our experience, there's going to be some rough justice along the way. And so long as we both enjoy being partners, it's good enough for me.
And so I think that that attitude would be great to pervade all of the competitors; that in some ways we're kind of in this together.
BECKY QUICK: You're right that generally when you give a keynote, it makes everybody else's stock go up. But there was something a little different last month, I think when you were at CES, and you kind of made an off-the-hand comment about what was taking place with quantum computing.
JENSEN HUANG: Oh, God.
BECKY QUICK: You said something like, you know, it'll probably be practically useful in five to ten years.
JENSEN HUANG: Did you guys see what happened?
BECKY QUICK: Yeah, the stocks tanked.
JENSEN HUANG: The whole industry's stock price went down 60 percent.
BECKY QUICK: Yeah.
JENSEN HUANG: Because of something I said.
BECKY QUICK: Well, I'm bringing it up only because overnight in Dubai, Sundar Pichai kind of echoed those comments. He said something. He said: The quantum moment reminds me of where we were with AI in the 2010s, when we were working on Google Brain and the early process.
And I just wonder, he's kind of agreeing with you, you saw what the industry did with your comments on this. Have you rethought any of that?
JENSEN HUANG: Okay, so this is what happened. So I made a comment about quantum computing. It turns out -- I'm not going to make a prediction. Their stock price is going to go down again. But here's what I'm going to say: I didn't realize that any of them were public.
(Laughter.)
It was my bad. I mean, we work with all of them. And I didn't realize they all went public, using these SPACs or something like that. And so I didn't realize it. So it was my fault. And so I thought I was just talking about all of our work together. And it's very complicated. It's really important work. It's incredibly, incredibly worth it. And we work with all of them.
And so this is what I did. Shortly after, shortly after, during GTC -- GTC is our annual conference. Please come. It's in San Jose. It's a few weeks away. A lot of big news.
And one of the most important things is I'm going to have a fireside chat, and I invited all of my quantum friends. And they're -- I don't know what they're going to bring along with them when they come. But I announced that we're going to have a fireside chat together, and we're going to talk about how I was wrong. And so that was -- and shortly after that came out, everybody's stock price went back up.
And so, anyways, please don't make me say anything.
(Laughter.)
BECKY QUICK: Okay, I won't. I will accept that.
JENSEN HUANG: It's not -- I'm involved in the industry, but it's their industry, and I'm very happy for them. I wish for their success, and I'm doing everything I can to help them.
BECKY QUICK: Well done. You're a nice guy.
The Biden administration's chip bans that were put in place that landed squarely on NVIDIA's head, just in terms of --
JENSEN HUANG: It was written for us.
BECKY QUICK: Yeah. It was written for you to try and keep you from giving the best AI chips to China. They were worried about the military and what would happen with that. It was pretty controversial within the industry. People like Larry Ellison came out against it. Obviously NVIDIA was opposed to it.
What's your problem with it in terms of the long term and what it might mean? And I know you were just at the White House a couple of weeks ago with President Trump. Do you think he sees those chip bans the same way that President Biden did?
JENSEN HUANG: Well, first of all, we support and we agree with several simultaneous things at the same time.
One, national security, of course. Of course, we don't know as much about national security, and we have to defer that to the administration, and we do. Whatever they decide, we'll support the best we can. Well, fully.
Secondarily, we believe in technology leadership, that American technology should be the standard by which the world is built on. And so you can't have leadership if you don't show up.
And so we simultaneously believe in these two thoughts. The question then becomes how's the balance and how do we think about the world beyond our direct relationship with China? And so the matter is, on the one hand, complicated; on the other hand, it's not.
If we don't take American technology into the rest of the world and take this opportunity to have the rest of the world build on American technology, it'll build on somebody else's.
And so we simultaneously believe in these two ideals, national security and American technology leadership. And somewhere in the middle is going to be some balance.
The policy came out of the blue, and we just didn't have a chance to engage the administration so late in the game on something that was 200 pages and was very, very complicated. In fact, it would have, in many ways, created some regulatory capture where the regulation itself would advantage certain companies at the expense of the rest of the industry.
And so we thought that that required some amount of dialogue and some amount of back-and-forth. And so now we have a new administration, and we have an opportunity to engage the administration from the beginning and talk through some of these complicated challenges.
BECKY QUICK: And you think they get it?
JENSEN HUANG: Yeah, I think so. I mean, there are a lot of technologies in the administration.
And I think we all want the same thing. We all want America to be great. We all want national -- we want our country to be secure. But we also want our country to be a world leader.
And those simultaneous, those two goalposts, are important to keep in mind. It's not -- it's not -- I understand that these are complicated matters.
BECKY QUICK: Let's talk about the future because you are something of a future seer. Probably everybody in this room has had the opportunity to play around with OpenAI or ChatGPT or something where they're doing things like finding recipes or writing letters or, you know, helping their kids with a term paper, something along those lines.
JENSEN HUANG: Oh, your kids don't need your help with term papers anymore.
BECKY QUICK: Yeah. I was being nice about it. They're using it without telling us.
JENSEN HUANG: Kyle's got it.
BECKY QUICK: Yeah, he does.
Where do you see this room and AI, let's say, three years down the road? How are they going to be using it?
JENSEN HUANG: Well, all the players are going to have AI coaches. I'm certain of it. And not only is your AI coach going to watch your videos, it's going to watch your videos, analyze your videos, and maybe even stick-figure your videos, turn your videos into 3D and show you all the things that you -- lost opportunities, you know, all the things you could do to be higher performing.
And so I think just as we have AI coaches, you and I both have AI coaches now, we're in the information world. And so I have a tutor, and I use both Gemini Pro and ChatGPT, and they're really terrific. And I use Perplexity as well, and they're really terrific. And so simultaneously I have all these different AI coaches and tutors that are helping me do a better job. And it's helped me a lot, learn new things.
Which, in the case of ballplayers, learn new skills. And so it could watch a game and compare to previous games and provide the analysis of it completely broken down. Lost opportunities, mistakes. It would highlight it all, you know, right there on the spot.
And so AIs will be watching the game from all the different angles in ways that you can't. And so there are a lot of different ways. Of course, AI is going to help you enjoy the game.
You know, one of the things that's really cool, as you know, today, when I -- I read a lot of papers, research papers and such. And so in the past you read the paper. Well, these days, I just give the paper to an AI, and I talk to the AI.
BECKY QUICK: Really.
JENSEN HUANG: And so in a sense, I'm talking to that paper. Do you guys understand? I'm talking to that paper.
So I give it the PDF, I give it a whole bunch of PDFs, and I'm talking to a whole section, a whole group --
BECKY QUICK: Asking questions of it?
JENSEN HUANG: Asking questions of it. First of all, summarize it for me, give me the high-level points, how does it compare and contrast with this and that.
And so I'm interacting with the researcher. That's the best way to think about that. And so whoever wrote that paper, I'm now talking to. All right, so what's going to happen in your world?
In your world, in the future, when you watch a game, if you want to interact with that game, you just talk to the game. And the reason for that is because that game has been -- we call it embedded. It's been embedded into an AI.
So that game has an AI. That game is going to become an AI. All of the intelligence, all of what happened about that game --
BECKY QUICK: Why did the Golden State Warriors just go on a 13-point run? Explain to me what happened.
JENSEN HUANG: What was the event that led to that? What was the critical moment? That's right. And so you could go back, you could interact with that game, ask it all the questions you want. And that game will talk back to you.
Just like the video, just like the PDF, my papers talk back to me, I interact with it, you'll interact with your game.
BECKY QUICK: Instead of our friends?
JENSEN HUANG: Well, both of you could be.
BECKY QUICK: You have talked in the past, there have been a lot of people who have been worried about AI, that it's going to lead to the end of industries, to the end of businesses, that people are going to lose their jobs. You've said the only people who are going to lose their jobs are the ones who don't understand AI.
We've all seen The Terminator. Explain to me why that scenario doesn't happen, or maybe why it would happen.
JENSEN HUANG: Well, time travel is still hard.
(Laughter.)
There's no evidence that time travel has been solved. And so --
BECKY QUICK: But otherwise?
JENSEN HUANG: John Connor is not coming back.
BECKY QUICK: Okay.
JENSEN HUANG: Yeah, he's not. Yeah, yeah.
And so, anyhow, I completely believe that, for the foreseeable future, you're not going to lose your job to an AI. You're going to lose your job to somebody who uses AI. And anybody who doesn't use AI has got a head in the sand so deep.
And the reason for that is it's so easy. This is the easy -- remember what AI is. AI is a new computer programming interface. It's a new computer interaction interface. It used to be C++ or C or Python.
And so now you program it with this new language called human. And so if you want -- if anybody who felt that the technology divide benefited a small population around the world, call it 100 million people or so, for the very first time, we've created a computer where everybody can use it.
You could be a farmer. You could be, whatever, a schoolteacher. You don't need to know how to program a computer. And for the first time, you can program the computer to do exactly what you want.
So we've closed the technology divide. And everybody should jump on as soon as possible, engage the AI. Because the way you ask the AI to do something for you is very much like how a leader inspires people to do things for them. You still have to provide context. You can't just go, "Do this." You have to provide context.
And so the way you prompt an AI, the way I prompt an AI, first I describe what is the background, what is the context, what is it that I want. I'm very clear about my mission. And then I might even provide some: Use this particular information as you think about what you're doing for me.
And so, for example, I give it specific papers. I might even give it papers that I think are counter to each other. And I make the AI read both of them so that I can have a contrasting view, if you will.
And so even programming an AI requires some amount of skill. It's not just type in something, some magic comes out. You have to provide -- it's called prompt engineering. And prompting the AI requires some amount of skill. Just like talking to people, asking questions and engaging your team to get them to do things.
And then one last thought on AI. People say as soon as you have superhuman AIs, then what are you going to do? Well, I'm surrounded by 60 superhumans as far as I'm concerned. And they're all individually better at what they do than I am. That's why they're on the management team.
And so I've got 60 superhumans that I work with, and somehow I still feel they need me. And the reason for that is because I do exactly what I just said. For them, I provide context, background, purpose, direction -- does that make sense? -- so that we unite everyone to go do something extraordinary.
You're going to do that with AI. And so it's a lot of it. Anybody who's great at leadership, you're going to be great at using AI.
BECKY QUICK: You are a superhuman yourself, Jensen. I want to thank you very much.
JENSEN HUANG: Thank you.
(Applause.)
BECKY QUICK: Thank you, folks.
JENSEN HUANG: Thank you very, very much.
BECKY QUICK: And we'll hand it back over to Ahmad.
AHMAD RASHAD: And thank you very much, Becky.
Jensen -- I'll ask you when we get back.
Anyway, we're going to take another break. Before we do, for all of you out there --
JENSEN HUANG: He sure takes a lot of breaks.
(Laughter.)
AHMAD RASHAD: What about the big tennis ball? Does that make it go slower, or does it go faster? Or is it easier to play table tennis with a bigger ball?
JENSEN HUANG: It's gotten bigger over the years.
AHMAD RASHAD: Both.
JENSEN HUANG: Yeah, yeah. Athletes are stronger and faster, so we've got to make the ball go slower.
AHMAD RASHAD: Oh, is that personal? Because I'm still strong and fast.
(Laughter.)
AHMAD RASHAD: All right.
All right, we're going to take another break. Before we do, for all of you out there, let me know if any of these dogs have been bothering you because we can't find one in the back.
(Applause.)
Ten-minute break. Go grab some coffee. We'll see you back here in ten minutes.
FastScripts Transcript by ASAP Sports