NVIDIA CEO Jensen Huang: What Quality Will Help You Succeed in the Age of AI?

Jensen, welcome to Kazakhstan! And it's such a pleasure to have you here. Let me first start by thanking yourself and the NVIDIA team for the partnership that we have. It spans from the first supercomputers, educational programs, and of course the Hackaleme Hackathon, which was set a Guinness World Record. And the local telecom operators are building their infrastructure powered by NVIDIA technology. And of course, our flagship project. In the AI data center valley, which is designed around tens of thousands advanced NVIDIA chips, and Kazakhstan is building one of the largest in the world AI gigafactory. So in Berlin, you challenged Europe how to build the AI economy. And today we would like to explore with this development of AI what the future looks like for Kazakhstan, for the people of Kazakhstan and for the talents and economy. So let's start. You're talking about the five-layer cake and Kazakhstan is also approaching the development of AI throughout the five-layer cake. The foundation and the most important one is the energy and Kazakhstan is investing into the energy infrastructure and we are planning to almost double our installed energy capacity by 2035 and make it above 40 gigawatts. So energy is intelligence Doesn't mean that the countries that have limited access to energy, they will have a limited access to intelligence? Let's start with that one. Well, first of all, I'm so happy to be here. Thank you for the great welcome. I want to congratulate your president and Kazakhstan for declaring that Kazakhstan will be an AI nation. It's incredibly visionary and it has... Yeah, that's right. You should be commended for your foresight and recognizing that this is a global technology reset and what an incredible opportunity it is for Kazakhstan. I want to congratulate you Thank you. for your leadership and helping guide Kazakhstan's AI journey. And in just a very short time, frankly, you've made incredible progress. While the rest of the world is Deciding how to engage and how to build the AI economy. Here in Kazakhstan, you're doing it already. I think, yeah, that's right. At the core, AI is a new technology that is reshaping every industry. Because this is a general purpose technology, and because of its transformative capabilities for every single industry, for every single citizen, for every company, for every country. This is a technology-driven industrial revolution. This industrial revolution is going to affect every single person, every country. However, it is also going to be a new industry in itself, not just the technology. It's a new industry in itself because AI is manufacturing. It has to be generated in real time. It cannot be pre-computed. And because it's generated in real time, For every single use, it needs to have a new type of factory. That's true. NVIDIA creates the computing systems that power AI factories. And these AI factories require natural resources. And the natural resources it needs is electricity. It needs electricity, it needs data, and it needs offtake. The offtake is coming from the inspiration of the leadership in Kazakhstan. You're inspiring all of your young people, students, universities, scientists of the industries. You're inspiring them to engage the AI industry, to engage the AI economy. That's where the offtake is going to come from. Kazakhstan has a fundamental advantage because you have so much natural resource already. You have so much energy. And instead of just exporting energy, in the future, you could transform the energy into intelligence and export intelligence. However, in order to do all that, you also need to transform your own data. And one of the important observations, and I'm very happy to see that there is Kazakh LLM, that you are already transforming, harnessing, processing your own data into your own intelligence. It's important to realize that your country's data and bodies and codes, your country's culture, history, your humanity, your values are captured in the data and that data belongs to you just as your natural resources belong to you. The data of your society is also part of your natural resource. Now you're able to combine the natural resources of energy, the natural resources of the Kazakhstan Data of your people in combination with a computation system from NVIDIA, you should be able to produce the intelligence necessary to drive all of the AI economy here in Kazakhstan, which is your vision. Your vision is to create the next economy, the digital economy, the AI economy for Kazakhstan. And so the five-layer cake is an excellent framework, and you have all of the assets, all of the natural resources to do so. Yeah, we use the five-layer cake as the main pillar of our digital Kazakhstan strategy for the next five to ten years. Because currently The total addressable market of AI exceeds 50 trillion US dollars and you know at which layer there will be the highest value captured. So therefore, maybe we're in the beginning of the process. We have to move in a systemic way with all five layers. Jensen, you mentioned about KazLLM. Let's speak about a bit of Sovereign AI. So, the people are questioning whether we need to have our own Kazakh GPT, so we want that AI understands and speaks Kazakh and captures cultural context and cultural specifics. So we built several models, Kazakh large-language models, based on open-source, open-weight foundation models. So at what point it makes sense or it's worthwhile to build our own foundation model from the scratch? Or does it really make sense at all? I think it makes complete sense. And in fact, I believe it to be a national security concern. And let me tell you why. Intelligence is foundational. Intelligence is general purpose. Intelligence affects every single citizen, every single company, every single country. A country's intelligence is produced from its data. And as we mentioned earlier, your country's data is your natural resource. It belongs to you. You cannot export something as fundamental as your people's culture and values And export that to import the intelligence from someone else. It makes no sense. So you want to control your own data. You want to control your natural resource. You also want to control the delivery of the service to your people. Just as you have Kazakh telecom, you have your own electric grid. In the future, you will have your own electric grid, your own internet, your own telecommunications. And your own intelligence grid, it makes no sense for the intelligence grid, which comes from, which powers every part of your country to be delivered from another country. It is okay to have another country and company supplement, but it cannot be completely. And so you have to have some ability to deliver the basic infrastructure services for your own country. The creation of a Kazakh... LLM, KazLLM is sensible because the invention of the technology is difficult, but the application, the adaptation of the technology is not so difficult. It's of course not easy, but Kazakhstan has a wealth of science and technology and mathematicians, and you have so many programmers. The developers here are rich. You're rich with expertise here. So you should be able to take open source technology, open model technology, apply it to your language, apply it to your culture so that you can run and operate the infrastructure of intelligence yourself here. Yeah, we also believe that open source allows us to be sovereign and use it on premise with the sensitive data. And we have our own, we built our own tokenizer, collected tokens, we know how to train to fine tune. So this talent development and human capital development is quite important. And I saw so many today already. Yes, exactly. And you've seen by the way, By the way, so many startups are here. Yeah. And you've seen some of them are in the physical AI space. So let's talk a little bit about this physical AI. As you call it, it's a next wave. People get used to feel AI through the screens, on the screens, on their laptops, computers, smartphones. But now with the next wave, it's moving into the... Physical AI and where the robots and other hardware systems interact with each other and work together. So Kazakhstan is rich on heavy industries. So we have factories, mines, agriculture, transport and logistics infrastructure. And of course, talented people with the deep expertise in those industries. And so what should Kazakhstan do in order to become a hub for training the robots? You've seen some initial work already, so our specialists know how to train the robots, how to train the models, and work with them. So what do you think, Kazakhstan to become a hub and maybe one of the main partners for NVIDIA to train the models for physical AI? It's a fantastic idea, and I think it's also an imperative for you to do so. You want Kazakhstan to be an AI nation. AI starts with cognitive AI, cognitive intelligence, soft intelligence. But AI includes everything, all the way to physical AI. Physical AI is the largest industry of all, of course, because most of the world includes interaction and understanding of the physical world. You also have Very important application here yourself, mining and natural resources and agriculture, all of your own indigenous industries today would benefit from the next phase with physical AI. Now, physical AI is not just robotics. Robotics is going to be very important. But physical AI includes everything from seismic understanding to mapping mines and wells To understanding geography. And geology, where to mine, how best to mine, understanding material science, the processing of chemicals, all of that is going to be part of physical AI and AI will be able to tremendously transform each one of those fields from the mining of the processing of the minerals to the transformation of the minerals into new materials in every single stage of that will benefit from physical AI. Now, we're at the beginning of the physical AI journey. This is the next wave. And so everybody is in the beginning. This is a great opportunity for you to participate, to help shape this future industry. In the future, just as we will have artificial general intelligence for cognitive intelligence, cognitive AI, we will have artificial general robotics. And I saw some today. Cause butt. was an example of the beginnings of artificial general robotics. And I have great, great hopes and ambition that this will be a very large industry. And we're delighted to partner with the robotics company here. I was so happy to see NVIDIA is already working with all of them. And so it's great to see. And we'll try to produce artificial general engineers. That's right. Artificial general... Farmers, artificial general, miners, artificial general scientists, every single field. Great. One of the main questions that concerns people even sitting around us is their jobs and how AI will convert their working lives. So everyone is curious about whether AI will help make their jobs better or will lead for them to leave behind. So what do you think, what should people care about and use their energy for developing them in five, maybe next five years? The only thing that matters is, do you have ambition? If you have ambition, then AI and robotics will lift you. It will empower you to do more. If you were a farmer and you have ambition, The robots will help you farm, but because you have ambition, you can expand your operation, you can expand your horizon, you can expand your business. If you are a scientist and you have ambition, then your AI assistant will expand the universe that you can explore. If you are an engineer and you have ambition, Then your AI engineer will allow you to explore better designs. And so the question is simply, do you have ambition? If you have ambition, and I believe you do, this is the first country that I've been to where the president declared the country to be an AI nation. In that case, AI will lift your nation. AI will lift your people. AI will expand your company so that you can compete and expand globally. AI will transform your society. That's how I would think about AI. Ambition and the talent and creativity should be taken through the since the childhood. So you've seen all these kids today, the TUMO program that spans from 12 to 18 years old. You've seen the Tomorrow School in partnership with the Call42, the peer-to-peer education, which is the 18 plus. Those are all people with ambition, children with ambition. Yeah. Yeah. So we embed AI into education, into schools, universities, and the concern that people expressing right now, how do we make sure that AI helps children to become better thinkers rather than just relying purely on AI to make it harmful for them? Again, the question is whether that child has ambition and aspiration. If the child wants to learn more, AI will help them learn more. If the child wants to do less, then the AI, of course, can help them do less. And so it is our job as parents to imbue our children, instill our children with hope, ambition, aspiration to help them dream bigger. And when you dream big, then AI will lift you to achieve that. Same exact idea for children, as well as for adults, for companies, as well as countries and nations. If you have aspiration, if you have ambition, and it requires the leadership to instill ambition, just as President Tokayev instilled ambition for the nation to be an AI nation, that's ambition. All the companies that I've met so far, you have ambition to be great. You have ambition to be global. You have ambition to be international. That ambition will be powered by AI. And so the most important thing we need to do as country leaders like yourself, as CEOs like myself and my friends, as parents, is to instill upon our people, our families, our companies, our nation. Instill upon them ambition. Let me give you a real example of ambition then. So we have, we talk about this youth, ambition, and the real case of ambition. So let me turn to our AI unicorn, the Hixfield AI. So, uh... I think this is the greatest example of ambition of young people. It's our first unicorn. The average age of the team is 25. And by the way, we have main co-founders here with us today. And we have Alex Mashrabov and Yerzad Bulat along with their broader team here. So please stand. The company is valued at the multi-billion valuation right now and I have a good update, good news for the people around us and those watching online. NVIDIA has invested into Hicksfield AI. Jensen, just a few words about Hicksfield AI. So what really made NVIDIA to decide in order to invest into Hicksfield? So just for the other followers, to be what should be a northern star for them in order to be in compliant to the investment decisions as yourselves? Well, the first thing is this, the first thing and the last thing. You invest in people, not companies. And the founders of Hicksfield are visionary in the work that they do. They had a wonderful idea. It's not easy to do. And they have global competitors. They have global competitors. Many of them are in San Francisco, in the United States. Incredibly well-funded and formidable expertise. But these are some of the brightest people in the world. They have an... 600 people here in Kazakhstan, and you have an average age of 25 years old. Many of them are deep tech engineers and deep tech computer scientists already. With offices here in Kazakhstan and in the United States, you're able to bridge the culture and the expertise from two different nations, and they're competing globally. And so, you know, when you're investing in a company, You're really investing in people and so you have to select the best people. I've been to their offices once. They start their work at 5 p.m. They're sitting, they're looking for a new office. They're expanding. They used to be like 150 people. You start at 5 p.m. I want my money back. But they work overnight. I start at 5 a.m. Today I start at 4 a.m. So, they're expanding right now, looking for an office, like 20,000 square meter office, and it's a great ecosystem. So, talented engineers. Well, Kazakhstan has a lot of real estate. It also has a lot of energy, and the DC Valley has one gigawatt plant, and so you're going to have a lot of compute right here in Kazakhstan, right? And so, it's going to be a hub, a computing hub, an AI factory hub for all of Europe. The telecommunications interconnection that you created, completely visionary. Now it put Kazakhstan right in the middle of the European AI ecosystem, and then you could feel the AI economy throughout. Fantastic. Let's talk a little bit about the security issues. So we see, I mean, quite recent news on some security hacks, jailbreaks, and so on. Doesn't NVIDIA plan to develop a technology that's embedded into the hardware whenever there are models got hacked or something happening wrong with the models? Will there be a technology on the hardware of NVIDIA that helps to mitigate those risks? Well, in order for us to enjoy, for the world to enjoy the positive benefits, And the incredible opportunities that will be created by AI, we have to develop it and advance it and deploy it responsibly. Just as when we're building self-driving cars, safety is vital from the very beginning. Just as when you're mining and you want to have a mining industry, safety is vital from the very beginning. And so if we want society to benefit from AI, It is important, and it's the responsibility of the creators, the creators like NVIDIA, creators like Higgsfield, that we advance this technology in a responsible way. The technology to keep AI safe is foundational technology as well. The technology has to be invented. The technology has to be engineered. We created a technology called Open Shell. OpenShell contains this autonomous system. These agents are essentially self-driving cars or robots inside a computer. And if you give them access to files and access to tools, they will freely go about to access the files to do their job. And so we have to contain them to make sure that whatever access is exactly the access we want it to have, whatever privileges are limited to what it needs in order to do its job, and we need to monitor it at all time. And so the technology for deploying, for even creating the AI safely, is called Open Shell. And I advise everybody to use it. And definitely, there are lots of new ideas that are coming up, especially in this fast-changing environment. Has there been any particular situation where you changed your mind about AI in the past year? No, I'm more enthusiastic today about AI than I was a year ago, and I was very enthusiastic a year ago. The big breakthroughs in AI this last year, especially with agentic systems, we've made AI useful now. In fact, if you look at the Higgsfield generated videos, in just one year's time, it has improved tremendously. In the beginning, Higgs field was able to generate short videos. Now you can generate very long videos that is temporally and physically consistent. It's consistent. It tells a story for a very long period of time. It's not possible without agentic reasoning systems. And so the technology has become very useful in video generation. Of course, word generation and software coding and robotic systems and drug discovery in so many different fields. These systems have become very, very useful. And we haven't mentioned you and we didn't have a chance to present you that a month ago we were running the first AI film festival. So which... It's a contest with more than 8,000 applications, videos submitted, of more than 100,000 minutes in total, with more than 120 countries participated, with a $2 million prize. With how much? $2 million prize pool. Wow. And for the jury, it was very difficult to determine which video is better. Kazakhstan is going to be the... Home of the Cannes Festival of AI films. It's a decentralization of Hollywood. Uh-huh. Okay. Yeah, that's terrific. So for the sake of time, yeah, we have to wrap it up. And just the final one. Number one. I have a question from the audience here. And so if you were 18 years old and living in Kazakhstan today, what would you spend the next five years learning and building? Well, I... I envy being 18 years old today. And the reason for that is because the world is changing so rapidly and there's a complete reset because of technology. You know, I was fortunate because when I was 18 years old, the personal computer was just becoming possible. And my generation was... Was... I... I've been working for 43 years. My generation was defined by the creation of the computer. And all of our four decades of endeavors has led to now artificial intelligence, which is a complete reset of everything that we've done. Instead of programming the computer, you have to learn how to program AI. And so that young man here in Kazakhstan, you're 18 years old. AI is now new for you. It's new for everybody else. I would dedicate myself to be the world's best programmer of AI. You know, AI has the ability to automate. It has the ability to partner with us. It has the ability to assist us. But just as you partner with some human and you have assistance from somebody, you have to learn how to interact. You have to learn how to program each other. Do they have to learn technical skills or just social sciences in order to do better prompting? You need both. You need to understand the best programmer of AI is someone who understands how AI works. The best user of a machine is someone who understands how the machine works. The best driver of a race car is someone who understands how the car works. And so the programmer of AI should understand how the machinery of AI works, what its fundamental principles are, what its strengths are, weaknesses are, how it's developing, where it will be in the future. And so all of these properties about the technology must be understood if you want to harness the capability of AI to the fullest. And so you have a lot of technical reasons to engage AI, and you have extraordinary potential when you learn how to program AI, how to use AI. And probably you've seen that the children and the youth that we have here at LM AI, they're probably on the right track. Well, they were upstairs already using AI, programming AI. Yeah. Well, thank you, Jensen. So it was such a pleasure to have you here on the Fireside Chat. It's your first visit to Kazakhstan and hopefully not the last one. And we look forward to expanding our partnership with NVIDIA. Well, thank you. Let's get a round of applause to Jensen. And we have... Thank you. And we have a small Kazakh tradition to our respected guests. We usually give our present, which is called Shapan. Oh, wow. For you to feel like a king of AI. Wow. Yeah, this is the Kazakh traditional robe of honor given to our most esteemed guests. So yes, wear it with joy. It fits you very well. Thank you.