Nvidia's Huang, Michael Dell on Agentic AI, Memory Demand and China

Michael in the quartic on 1000 new clients for I server.

For AI factory is a hell of a jump. What is it that that those new finds 5000 total are actually building now difference one year ago.

I think that's probably a good place to start.

I think the change we see is it's kind of new from testing and evaluating into production. And we showed some great examples on

production. And we showed some great examples on stage, right, right. With Eli Lilly, with the thousand GP in the physical world in Samsung. And these are not things that are on the screen. Right.

screen. Right.

This is in the real world with the largest companies in the world.

And so it's propagating broadly across all customers, every industry in every country. And

country. And you know, you see the improvement in all the models.

And now we have the adventure capabilities.

And so while it is exciting, it has been a tremendous amount of growth.

I still think it's just the beginning of this wave.

Particularly when it comes to enterprise, which is really where, you know, we have an enormous opportunity. What's so fascinating, Jensen, is you spent four years telling me that we needed to change the definition of the computer in the context of accelerated computing, but the big focus was on the hyperscalers right? Cloud.

hyperscalers right? Cloud.

What I took from point was présentation where this is happening.

You said locally but on prem. What's the Nvidia interpretation of that part of this site. Intelligence has to be performed produced at the point of context. And so wherever the context is, wherever the action is, that's where you want to produce the intelligence.

For most of the early applications of AI tools in the cloud.

A lot of consumer services are in the cloud.

However, for Lily Samsung, the future manufacturing a lot of companies.

You want agents to be on prem because that's where all of your data is, where all your secure data is, your proprietary data and all of the skills associated with your companies. And so now we have agents that are here, AI, that can do work, right. ChatGPT was fantastic at launch generative AI, but it just hit content. That was it made.

Making content is very important, but doing work is really valuable.

And now we're doing productive work incredibly well.

That's what they call a genetic AI in this new era, what everyone is trying to work out is on all the GPUs locked up at the hyperscalers.

How is Michael Dell going to service these 1000 new clients with the GPUs to build their own on prem local AI factory?

Well, the supply chain that Jensen has built, we've built together is continuing to scale up. And while it's true that there's more demand than supply. There's more supply that's that's being added. And, you know, customers are figuring

added. And, you know, customers are figuring out how they start to scale these systems up.

So, uh, you know, I think what's also happening is companies are understanding that when they reimagine their workflows using this technology, they don't get 10 or 20 or 30% improvement. They get ten times or 20 times or 100 times. And that is really the speed that

times. And that is really the speed that matters to make a business successful. We're doing it ourselves.

Nvidia is doing it. And so it's not a secret anymore that these things are possible. And every company wants to capture that speed and translated into competitive advantage and outcomes.

Dell was the sales channel, right? Jensen Michael's company is very good to selling technology to America's biggest companies.

How is that going to change things for media going forward, like the makeup of the types of companies we're talking about are at scale.

But there's also that kind of middle market, the data center that's being built, different kinds of it in the industrial space, in healthcare.

Is that something that puts a veneer into new territory, away from the frontier labs, away from the hyperscalers?

Well, Nvidia is a technology company, right?

The hyperscalers have the ability to take our technology and integrate them, operate them into a service. Dell has the ability to take our technology, turn them into a solution that delivers impact to customers.

If you look at what has happened, a genetic AI has completely, as we were talking about earlier, reinvented computing.

We had to do several things together. The first thing, of course, we have to build the brain. This is the Grace Blackwell ambulance 72 the Vera Rubin, MD 72 giant large language models.

The second part now is the Vera CPU that we're now in the process of launching the the highest performance CPU in the world.

It's designed for energetic AI. And now this will be the harness running the agent itself using the tools.

What does this mean? Well, harness.

Harness is what, uh, puts out puts a harness around the large language model so that it can access memory, access the network, use tools, have local scratchpad memory, working memory access, long term memory.

And so that harness basically turns, if you will, the brain into an agent, okay.

Into a digital robot, if you will. That can do work.

And so now the agent runs on a CPU. We also worked with Dell to create a new type of long term memory for agents. He called the Dell AI data platform that's built on Nvidia. The networking to scale it out is built on Nvidia. So the agent, the brain, the long term

on Nvidia. So the agent, the brain, the long term memory, all of the networking necessary to scale it up, as well as the agent runtime itself, we call Nemo crawl, running in a secure and governed container called Open Shell. All of that has been put together.

And now the technology part of the technology parts, what Del has to do is turn it into a solution that people can use.

Dell will do for the world's enterprises what the clouds do for the clouds.

It makes perfect sense. What is the Dell story?

Michael around CPU and sort of like general purpose computing in the tech era. We've talked a lot about, uh, the AI

era. We've talked a lot about, uh, the AI factory offering the GPU, but actually there's potential for you in more general purpose workloads. The build outs happening either way.

It is. And the demand is a two supply there as well as, you know and look, as you move to these agent frameworks inside companies, you use a lot more CPUs.

Yeah. And, uh, you know, that's that's just the reality of what was happening. And

I think I think that's only going to increase.

So instead of humans using tools, it's now agents using tools and agencies you were talking about earlier on stage there.

We're going to have we have a billion people will have hundreds of billions of agents. People use tools every now and then.

agents. People use tools every now and then.

Agents are going to use tools all the time, and agents use tools very quickly.

And so we're going to need a lot more CPUs.

And those CPUs are connected to GPU brains so that the CPUs know how to think, how to reason, how to plan, and how to use those tools.

So that's basically how it works. Gentlemen, what is the biggest supply constraint that for you right now? Well, certainly, you know, memory is a challenge. I think it is memory.

challenge. I think it is memory.

The advanced node semiconductors are still challenging.

You know, it's it's really I mean, we we think about it from the things that we're producing in. Right.

Uh, the semiconductor supply chain is ramping, but the demand is growing faster than in the supply. In our case, we provide the technology integrated. And so the memory comes with our technology. We've been planning our supply chain for

technology. We've been planning our supply chain for a couple 2 or 3 years. We have the largest supply chain in the world. Our partners have done a great job

world. Our partners have done a great job securing supply for us. And so all of the pieces go together.

The cost is lined up with the HBM, which is lined up with the Grace Blackwell's in the CPUs. There's a Kawasaki, the colossal, the Colossus. All of it is all lined up.

Colossus. All of it is all lined up.

The silicon photonics is lined up. Everything is all lined up.

It's just that the demand is much greater than the overall capacity of the world. So the overall capacity.

world. So the overall capacity.

Jensen, should I put my textbook away? Because if I get my textbook out, it tells me that memory historically is cyclical.

It's boom and bust. And so you both kind of have to convince the memory makers of the permanency of this to build the capacity that what sort of fool the way is that the right way of looking at it, that this is not a boom and bust cycle, it's just a complete change in the structure of that market? Well, Michael, I do this all the time.

market? Well, Michael, I do this all the time.

We spent a lot of time with the supply chain.

I mean, if you ask Sanjay Mitra over a micron, they'll tell you.

Three years ago during a meeting, I explained the future to him exactly.

Is this happening right now? and I was really grateful that that microphone and video really lined up to line up all of our roadmap.

Tony will tell you over at SK that we did the same thing years before, and so it's our job to make sure that the vision of the future of the industry we convey upstream to our supply chain so that they are building for it.

We also have to convey it downstream to people who have power generators and land and financing and so on and so forth.

And so we have to make sure that the supply chain upstream and downstream are prepared for the future. It is true that the simple, the simple logic is this, that we have now reached a level of a genetic I useful, I productive, I capability. And the way to think about these agents is kind of like just digital workers, right?

We have hundreds of millions of digital workers in the world.

We're going to have billions of AI agents in the world, and they're going to be working 24 over seven. And so just as we give every digital worker a laptop and a small part of the data center, we're going to have to give every agent, essentially a computer and a little bit of storage and a data center to use. Think about this way.

You know, you you do individual work, you know, as a person a day and you send it on to somebody else. And, you know, there's interactions.

Well, now you might have hundreds or thousands of, you know, digital agents working for Ed, right? I supervise, you supervise.

And that's going to help you be way more productive, get way more things done, and expand your creativity. Now, it does require a lot more computing and memory and storage and working and all of the things that we're doing together. Last one on this, like Jensen outlined

doing together. Last one on this, like Jensen outlined the micron microRNA, the example three years ago, you gave him the heads up.

Did they believe you? Are they sort of acting on their investing. I mean it's we're managing through it.

investing. I mean it's we're managing through it.

But these things are very hard to predict right.

If you tried to predict, you know, in 2023 what the demand was going to be in 2027. You would have a hard time doing that.

2027. You would have a hard time doing that.

So it does take a long time to build these factories.

But we've got great relationships with these partners.

We have for decades that's helping us. And they see that we're winning.

And so they want to work with us even more.

And it's really a great long term partnership, even though we'd like more right now, we're in the beginning of the I build out.

This is literally the very beginning of the agency that I build out.

We're going to be building this out for a decade, maybe more, because after this, digital agents will be physical agents.

Then we go to the physical eye. We haven't even started that.

I mean, you saw some examples of that, you know, in the keynote, but that is a way bigger market and it will require all sorts of new infrastructure capabilities for the very first time, bring it to the world's 90 trillion other industry. And so there's a giant industry ahead of

other industry. And so there's a giant industry ahead of us to build towards now. Meanwhile, the supply chain is more than doubling every year. I mean, it's probably quadrupling every year, but we'll still have a hard time keeping up with the buildout for at least the decade since China. Jenson, you just returned from China on

Friday on Air Force One, the president said that H 200 came up, but that China's position is it wants to support its own industry.

Can I just ask what the net outcome was that your trip to China and your understanding of what is not or is allowed with 80, 200 and the customers who you have or do not have in China. The president wants America to win everywhere, right? The president wants America to lead the

everywhere, right? The president wants America to lead the I Revolution. And so, uh h two hundreds are licensed

I Revolution. And so, uh h two hundreds are licensed to sell to China. Uh, the Chinese, the Chinese government has to decide, uh, how much of their local market do they want to protect and how much of their local market do they want to expand with more a more I capacity a my senses that the demand in China is so incredible, just like it is

here. A genetic AI is also making enormous

here. A genetic AI is also making enormous progress there. My sense is that over time, uh, the

progress there. My sense is that over time, uh, the market will open. Uh, President XI was very clear that he wants China to be an even wider open market.

Uh. Uh, Premier Li Chung was very straightforward and and to explain very eloquently that that, uh, China will be an open market. So I'm looking forward to trying to be in a more open and clarify, as you were able to meet with those officials directly to discuss whether or not you can sell to those Chinese tech companies. I didn't I didn't discuss directly with

companies. I didn't I didn't discuss directly with them about 200. Right.

I was there to represent the United States, and I was honored to do so.

I was there to support President Trump. And, uh, really glad to do so.

Uh, but that was really the focus of my trip.

Uh, President Trump had some conversations with with the leaders, and, um, I'm looking forward to. To what they decide.

Michael, you did not go to China. But I think what was interesting is you are a member of the President's Council of Advisors for Science and Technology, as is Jensen. Your your net conclusion on whether or not China will become open to American technology companies to do business there. You know, we have a business in China.

there. You know, we have a business in China.

Obviously we comply with all the restrictions and, you know, various controls that are in place. But I hope that there's more economic collaboration between the United States and China that ultimately is will lead to greater outcomes and prosperity for everyone.

And, you know, a greater likelihood of, you know, a successful relationship between the countries and, you know, around the world.

The final question on that trip, Jensen, is the sharpest rhetoric was probably on Taiwan. We've talked about the supply chain But

Taiwan. We've talked about the supply chain But what did you take from those comments from, from President XI on on the issue of Taiwan? Of course, from a manufacturing capacity

of Taiwan? Of course, from a manufacturing capacity standpoint, TSMC is a critical partner. You and I have discussed it in the past, but at this moment in time, how top of mind is it for the security of supply for Taiwan? None of us were involved in any of those

for Taiwan? None of us were involved in any of those conversations except for President Trump with respect to Taiwan.

Obviously, Taiwan is still epicenter of the world's technology, manufacturing and technology development. The supply chain is rich in Taiwan.

We're also, of course, we industrializing the United States, bring manufacturing back to the United States. And we're doing so at a time when demand for AI in the beginning of this new computer revolution is happening.

And so demand is extraordinary. So as a result, we're building more factories here in the United States, chip factories, uh, packaging computer factories, AI factories, of course. So we're building factories of all kinds here. They're also ramping up capacity.

here. They're also ramping up capacity.

And the reason for that is because the demand is just so great across the board. I think that I think the answer is, is

board. I think that I think the answer is, is that we want to have it is possible to have a supply chain diversity and resilience. And we everybody should be seeking to

resilience. And we everybody should be seeking to improve that. Um, it is also very true that Taiwan

improve that. Um, it is also very true that Taiwan will continue to be, uh, one of the epicenters of the world's technology hub. Like, I grew up using Italian pizza.

hub. Like, I grew up using Italian pizza.

You know that we discussed it in the past.

Desktop laptop. You and I never talk about computers in that context. We were always talking about

that context. We were always talking about supercomputers, accelerated computing. But, you know, you and I think you should upgrade. I mean, we have the new XPS 14 or 16.

should upgrade. I mean, we have the new XPS 14 or 16.

That would be my choice for you. So what is the story?

These are the best notebooks we've ever had.

Talk about APC, but we're gonna get chances to finish.

But but what is the role of the PC in this is a case like I'm using a computer at my desk to do what? Yeah, well, look, I mean, it's still the device that is the center of productivity for knowledge workers, and it is right there in front of everyone. And, you know, we have a great business there. And those devices are evolving, too.

there. And those devices are evolving, too.

You saw on stage how we're, you know, embedding the ability to run the small models and local models inside the PC. And,

you know, what's happening is customers are wanting more powerful PCs.

Yeah, because they want to be able to do all this, this, this great hybrid AI.

And so it's yeah, it's a it's it's a great business.

It's still very much alive. And it also gives us incredible scale and strength in our supply chain, which helps us secure all the, you know, needed ingredients that we need. So you you spent 31 years working on the service design to get that accelerated computing.

That's the scale we're talking about. Let me just be brief.

We started with the PC, but. Why don't you just team up?

I was, I was trying to sell them a gaming GPU.

So what's going to happen between the two of you?

A PC with a powerful GPU inside it. Why doesn't you have that?

Yeah. And what's the plan going forward for that? Well, we can't tell you the plan right

that? Well, we can't tell you the plan right now. Tell me very, very soon.

now. Tell me very, very soon.

We like to tell you there's there's a. Well, let's think about it.

Think about the arc. Think about the.

I'm. I'm interested in computing, no doubt.

Think about the Ark of computing. When Michael and I came into the industry, it was that it was kind of at the tail end of mainframes.

Not that it was a tail end of mainframes because mainframes go away.

It was tail end of of its growth, and it was the beginning of personal computers.

Um, we're now seeing the beginning of, of course, AI in the cloud.

And that's going to continue to grow. But we're also going to see personal AI instead of personal computers when personal AI.

So the question is, and the reason for that is just we were talking about earlier, AI needs to be where the context is.

If all the information that I have is on my laptop and I need.

I need help. I need AI to help me do work on my laptop. Then I need I to run kind of locally.

laptop. Then I need I to run kind of locally.

And if I have, uh, if I have a factory, then I need agents to run in the factory. If I have if I have a hospital, I need

factory. If I have if I have a hospital, I need agents to run the house, because that's where the operating room.

And you can't can't be running somewhere else, right?

Because that's where the context is. That's where the action is.

Yeah. If you got an autonomous vehicle, right, the it has to be run in the car inside the vehicle.

Yeah. And so this idea of distributed intelligence and unmetered intelligence, right, where you can generate as many tokens as you want. And under new XPS 16, you just have to get Bloomberg to get you one, you know? Sure we can.

I know, I know. Michael Dell, chairman and CEO of Dell Technologies and was CEO of media.