저자: JensenHuang · 외국인
요약 보기
한국은 반도체·산업 호황기로 AI 인프라 확산의 적기. 엔비디아-SK그룹 간 5000억달러 이상 규모 사업 제휴 발표(SK하이닉스 메모리 구매, SK텔레콤 AI 클라우드 최대 2GW 구축 및 슈퍼컴퓨터 판매). 향후 10년간 반도체 산업이 현재보다 약 10배 커질 것으로 전망하며 HBM 등 공급 병목(전력·토지·건설인력)이 지속됨을 언급. 오픈모델의 안전·보안·혁신 중요성 강조.
다듬은 글
I think we just start with the basics. Like career is incredibly important to to the I build out globally. We'll get into high bandwidth memory. But just from this summit, uh, 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, uh, adopting new technologies and is a really technologically, uh, forward leaning society.
And they they love using AI. AI has has really diffused throughout their, uh, society and, and their industry. And so, so this is a great time for them. We're announcing several things. We announce, um, a big partnership with uh, 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 of and purchasing of memories or selling supercomputers to them as they scale out to gigawatts of factories. There's a whole bunch of other announcements. We're investing $1 billion in never to help. They're the, uh, Korea's leading a cloud. They're going to scale up in Korea up to two.
They're going to scale up 200MW, I think it is. And, uh, they're going to expand, uh, across the world. And so we have a whole bunch of announcements that we're making today with the, the, the expanded relationship. There's also sort of more direct involvement with Nvidia on the roadmap for HBM, future generations of HBM.
Talk about that. You know, I remember 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 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 HP and two worked on HP and three three E4, four E and then beyond. And so we've got, we've got a whole roadmap of memories that we're working on together. Uh, it is also the case that the the semiconductor industry has really changed. And the reason for that.
Um, because we used to build computers for people to use and we're going to still continue to build incredible computers. These are now eight processing eyes for, uh, humans to collaborate with. But in the future, we also have eye 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 100 billion agents and billions of robots all using computers.
The computer industry, the chip 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 that the semiconductor industry is probably going to have to be ten times larger than it is today.
Over the next decade or so and, and, um, uh, 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, it's 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 kind of get going to increase the supply of HBM bits?
Uh, for Nvidia based systems, wherever they are? Well, we we don't have enough bits. Uh, we're constrained in HBM memories, Lpddr3 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's that's, um, is going to make sure that we continue to build out in a throttled way. You know, for a decade. And the reason for that is because these infrastructure, unlike electronics, 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, 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 going much faster than that. The $500 billion number is large. Would 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's spending in the Korean economy or it's SK fronting capital expenditures? Just a little bit more detail. We're going to we're going to be purchasing memories from them for many years to come. And as you know, we buy we build a lot of computers in order to build $1 trillion worth of.
There are room in systems. You're going to have to buy a lot of system memories to go with it. And so we have we have large purchase agreements and large purchase intentions with SK Hynix. Meanwhile, SK Telecom is going to become an AI cloud there We're starting to build already. They. They're intending to build up to two gigawatts in the near future.
And, um, and in that, in that agreement, we will be selling supercomputers to them. So between us, we're going to we're going to we're going to do half $1 trillion worth of business, over half $1 trillion worth of business. I was able to sit down with SK Group chairman one 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. Um, the goal of the goal of AI is to produce an intelligent, 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, uh, result in using less tokens to do so. You could also, um, uh, produce eyes that are much more efficient and maybe you can think longer, explore more options and as a result, as a result, produce a smart answer. There are a couple.
There are many different ways to 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 has extraordinary AI researchers and whatever, whatever, um, conditions and whatever resources that they have, amazing people will find great answers.
And so you're going to find you're going to you know, my expectation is that China and the United States will continue to advance AI, um, the conditions are different, the resources are different, their constraints are different. But they're all there. You know, these amazing researchers will find answers.
And and I think that that, um, in the case of China, they're producing more AI researchers than probably all of the world has, you know, in any given year. And so they're producing it. Manufacturing intelligence is important. They manufacture the most the most important version of it, which is the researchers.
And so. So this is a this is an area a country that is going to produce excellent AI technology. And we are to keep an keep uh, continue to learn from them, work with them. As you know, you're here in Silicon Valley right here in San Francisco. But the number of AI researchers here that came from China that are Chinese is really quite significant.
And so, uh, you know, we're really fortunate to have them here. And, and, um, you know, we just got to keep on racing. You made your first post on X. I did, and you did. So, uh, by sharing a letter signed by many of your peers, American companies, to talk about the importance of open models to America, to to the industry, to the development of AI.
And in the letter, it's pretty well explained, you know, your rationale. But what what was the catalyst for now? What why did you and Satya Nadella and others need to do that in this moment? What we sense we sense that there's, there's, um, uh, a growing, uh, um, sentiment that that, um, and the consummate 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. We I see a future where the world uses tons of closed models.
And I, I encourage everybody, including my company, to use OpenAI and Clyde and cursor and, and, uh, cognition and perplexity use everything that you can because uh, of of the cloud. Because it's just easier and and you build only what you must. Um, 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 we simply cannot afford to share. This is our companies alpha, our companies intelligence, and we have to make sure we keep that proprietary. Maybe it's because our company works in a 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 that, um, you simply in a particular country, you have to have your own eye, you have to control your own eye. Whatever those reasons are, there could be cost reasons.
But but I think I think that largely I would recommend people build their own eyes, 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 eye that's closed and an eye that's open that you can build your own AI. Now, one of the things that that people, you know, misunderstand about, uh, these open models is.
Yes. Uh, you can host it yourself, but you can build your own computer. But most people use used computers in cloud. Frankly, I think close models are cheaper. You know, if you don't have to build yourself, if you don't have the training yourself cost a lot of expertise to fine tune and maintain and guardrail and keep it safe and evaluated and, and of course, even build computers to host it.
So there's nothing cheap about doing that. The reason why you need open, open models is because you need to have control, because you need to adapt something for for your own very specialized use cases. And so I, I think there's a lot of misunderstanding about about closed versus open. We felt that it was important for people to understand that there's a world for both open weighted versus, uh, open source as well.
There is a distinction, um, uh, open, open waited as much, as much as open as that you can. The more open it is in the way that we work. We put the weights, um, 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 eye model that we've open.
Wait. And so that a built by by us teaching you how to do that, you can then do it for yourself. You know, I think I think the um, the idea that the world is going to be one or the other is just completely wrong. And, and, um, the idea that, uh, 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 where the release of communique three, which on an open way to basis releases for the July 27th, and then the case study of two open AI models mistakenly accessing hugging faces systems and hugging face, trying to use an open model in its defense, where the guardrails were a factor, which you just reflect on those two.
I know that you've been asked about them, but they were they seemed to be like really big moment today. 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. Is it possible for a model to be, if you will, stolen?
It could be possible that that somehow is leaked from the inside. It's possible that the guardrails or the sandboxes of an eye closed AI model wasn't properly engineered, and as a result, it was able to attack another and another company in some way. And so. So just because something is closed and just because something is proprietary doesn't necessarily make it secure and safe.
Of course. Thank thank goodness we have two companies. Well, I guess more than that. Several companies have built 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 point of failure as an industry, as a world, we should have distributed massively distributed self defense. And so in the case of in the case of the example you just mentioned, hugging face thankfully was able to access an open model and I think they use the GLM 5.2 as my understanding.
Um, they, they couldn't get a proprietary model. They could not get a close model, uh, to help them figure out what happened. And but this is exactly the reason why you want to have open models. Because in that case, they use 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 so this is a perfect example of self-defense. That's necessary is 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.
Comments 0