Nvidia CEO Jensen Huang Talks AI Golden Age in South Korea, New Naver Investment (Full Interview)

Jensen, I think we just start with the basics, like Korea is incredibly important to the AI build-out globally, we'll get into high-bandwidth memory, but just from this summit, from the president being here, what is the takeaway? What is it you're trying to achieve? Well, we're announcing a whole bunch of partnerships with them. This is the golden ages for Korea, as you know. Their semiconductor business is booming, their industrial business is booming. You know, this is a country that has the ability to help the world build out the AI infrastructure, that they're incredibly adept at adopting new technologies, and is a really technologically forward-leaning society. And they love using AI, AI has really diffused throughout their society and their industry. So this is a great time for them. We're announcing several things. We announced a big partnership with SK Group, where our companies are going to enter into a business partnership where we do over $500 billion of business with each other, whether it's consumption and purchasing of memories or selling AI supercomputers to them as they scale out 2 gigawatts of AI factories. There's a whole bunch of other announcements. We're investing $1 billion in Naver to help. They're Korea's leading AI cloud. They're going to scale up in Korea. They're going to scale up 200 megawatts, I think it is. And they're going to expand across the world. And so we have a whole bunch of announcements that we're making today. With the expanded SK relationship, there's also sort of more direct involvement with NVIDIA on the roadmap for HBM, future generations of HBM. Talk about that. I remember you being on stage earlier this year saying, five years ago, we told our supply chain what was going to happen. And it did happen. And you gave some credit to the memory makers in going with you on that journey. But clearly, you want to be involved in the direction of travel for future generations of HBM. Yeah, we're working together on, of course, we started with HBM2, worked on HBM3, 3E, 4, 4E, and then beyond. And so we've got a whole roadmap of memories that we're working on together. It is also the case that the semiconductor industry has really changed. And the reason for that, because we used to build computers for people to use, and we're going to still continue to build incredible computers. These are now processing AIs for humans to collaborate with. But in the future, we also have AI agents and robots, and they're going to be using computers. So instead of just a billion people using computers, we're going to have a hundred billion agents and billions of robots, all using computers. The computer industry that's built on top of the chip industry surely is not big enough. And so this is one of the realizations of the semiconductor industry, that now computers are built not just for people to use, but computers are being built for computers to use. My guess is that the semiconductor industry is probably going to have to be 10 times larger than it is today, over the next decade or so. And so working with our partners in Korea and around the world to scale up the supply chain of semiconductors so that we're prepared for this AI future is really important. I've had the opportunity to ask you about this more than once this year, but how much do you need the Korean economy to get going to increase the supply of HBM bits for NVIDIA-based systems, wherever they are? Well, we don't have enough bits. We're constrained in HBM memories, LPDDR memories. We're constrained in just about every part of the supply chain. We're even constrained now with land and power and construction workers to set up the data centers. I think this is one of the areas that is going to make sure that we continue to build out in a throttled way for a decade. And the reason for that is because these infrastructure, unlike electronic devices like PCs and phones and things like that, it's really, really hard to scale up land, power, and shell. And so all of the supply chain just really needs to get built out over the years. I think we have the ability as an industry to double each year, but we're going to have a hard time growing much faster than that. The $500 billion number is large. Could you just talk a little bit more about what it encompasses? We've gone over a lot on your commitment to the US in terms of spending. Is that NVIDIA spending in the Korean economy, or it's SK fronting capital expenditures, just a little bit more detail? We're going to be purchasing memories from them for many years to come. And as you know, we build a lot of computers. In order to build a trillion dollars worth of their Rubin systems, you're going to have to buy a lot of system memories to go with it. And so we have large purchase agreements and large purchase intentions with SK Hynix. Meanwhile, SK Telecom is going to become an AI cloud. We're starting to build already. They're intending to build up to two gigawatts in the near future. And in that agreement, we will be selling AI supercomputers to them. So between us, we're going to do half a trillion dollars worth of business over half a trillion dollars worth of business. I was able to sit down with SK Group Chairman Chae Taewon very recently for about 40 minutes. And at the end of the conversation, we got to what is the difference in approach, the academic difference in approach on AI between the United States and China? And his view on it was that China is very focused on lowering the dollar per token. In America, we're still focused on the quality of tokens. I wonder what you think of that. The goal of AI is to produce an intelligent, a smart answer. Now, you could approach it in a couple of different ways. You could, of course, make all of the tokens smarter and smarter and as a result, result in using less tokens to do so. You could also produce AIs that are much more efficient. And maybe you can think longer, explore more options. And as a result, produce a smart answer. There are many different ways to reach intelligence and deliver smart answers. In the end, really, I think you have to take a step back and just realize that both countries have extraordinary AI researchers. And whatever conditions and whatever resources that they have, amazing people will find great answers. My expectation is that China and the United States will continue to advance AI. The conditions are different. Their resources are different. Their constraints are different. But these amazing researchers will find answers. I think that in the case of China, they're producing more AI researchers than probably all of the world has in any given year. Manufacturing intelligence is important. They manufacture the most important version of it, which is the researchers. So this is an area, a country that's going to produce excellent AI technology. We ought to continue to learn from them, work with them. As you know, you're here in Silicon Valley, right here in San Francisco. The number of AI researchers here that came from China that are Chinese is really quite significant. We're really fortunate to have them here. We just got to keep on racing. You made your first post on X. I did. And you did so by sharing a letter signed by many of your peers, American companies, to talk about the importance of open models to America, to the industry, to the development of AI. And in the letter, it's pretty well explained, your rationale. But what was the catalyst for now? Why did you and Satya Nadella and others need to do that in this moment? Well, we sense that there's a growing sentiment and the wrong sentiment for open models. It's really important to realize that open models is essential for safety. Open models is essential for security, for cybersecurity. Open models are essential for innovation. It's necessary for startups. It's necessary for sovereignty, company sovereignty. I see a future where the world uses tons of closed models. And I encourage everybody, including my company, to use open AI and cloud and cursor and cognition and perplexity. Use everything that you can out of the cloud because it's just easier. And you build only what you must. And so in order to build what you must, you need to have open models to do that with. And the areas where we must, maybe it's because we have expertise that we simply cannot afford to share. This is our company's alpha, our company's intelligence. And we have to make sure we keep that proprietary. Maybe it's because our company works in an industry that's regulated. And therefore, we simply can't pass along the service level agreement. And we have to make sure that we can deliver fully on the service and the promise that we sign up for. Maybe it's something to do with sovereignty that you simply, in a particular country, you have to have your own AI. You have to control your own AI. Whatever those reasons are, they could be cost reasons. But I think that largely, I would recommend people build their own AIs, especially when they need to control it for whatever reason. And so I think the future is going to have lots and lots of use of AI. That's closed and AI that's open that you can build your own AI. Now, one of the things that people misunderstand about these open models is, yes, you can host it yourself. But you can build your own computer. But most people use computers in the cloud. Frankly, I think closed models are cheaper. You know, if you don't have to build it yourself, if you don't have to train it yourself, it costs a lot of expertise to fine tune and maintain and guardrail and keep it safe. And of course, even build computers to host it. So there's nothing cheap about doing that. The reason why you need open models is because you need to have control, because you need to adapt something for your own very specialized use cases. And so I think there's a lot of misunderstanding about closed versus open. We felt that it was important for people to understand that there's a world for both. Open weighted versus open source as well. There is a distinction. Open weighted, as much as open as you can. The more open it is, in the way that we work, we put the weights out there. We also teach people how to train the model from the data that we also open source. And the reason for that is we want to enable you to completely reproduce the AI model that we've open weight. And so that ability, by us teaching you how to do that, you can then do it for yourself. You know, I think the idea that the world is going to be one or the other is just completely wrong. And the idea that open models is somehow unsafe is also fundamentally wrong. And so we just want to make sure that people understand. To finish our conversation, you know, the two big case studies were the release of KimiK3, which on an open weighted basis releases for the July 27th. And then the case study of two open AI models mistakenly accessing HuggingFace's systems and HuggingFace trying to use an open model in its defense where the guardrails were a factor. Would you just reflect on those two? I know that you've been asked about them, but they seem to be like really big moments in AI overall. Those are perfectly perfect canonical examples. Just because something is closed doesn't necessarily therefore make it safe or secure. It is possible for a model to be jailbroken. It's possible for a model to be, if you will, stolen. It could be possible that somehow it's leaked from the inside. It's possible that the guardrails or the sandboxes of an AI closed AI model wasn't properly engineered. And as a result, it was able to attack another company in some way. And so just because something is closed and just because something is proprietary doesn't necessarily make it secure and safe. Of course, thank goodness we have two companies, well, I guess more than that, several companies that build closed AI models. And these are extraordinary technology companies, and they're doing their best to keep it safe and keep it secure. But it is also the canonical case that single points of failure is where we have the greatest vulnerability. We cannot have single points of failure. As an industry, as a world, we should have distributed, massively distributed self-defense. And so in the case of the example you just mentioned, Hugging Face, thankfully, was able to access an open model. And I think they used GLM 5.2 as my understanding. They couldn't get a proprietary model. They could not get a closed model to help them figure out what happened. But this is exactly the reason why you want to have open models. Because in that case, they used GLM 5.2 to identify where the vulnerability was, where the penetration was, and were able to quickly identify them and patch it up. And so this is a perfect example of self-defense that's necessary. It's a perfect example of diversity of AI technology being necessary. And it's a perfect example of why open models and open capabilities for self-defense is really important.