NVIDIA CEO Jensen Huang Reveals How China is Outpacing the World in Robotics and AI | AI1G

that it is the case we're because of this new industrial revolution that we've started, we are creating new factories in

America. We're creating new jobs in

America. We're creating new jobs in America. And somebody recently told me

America. And somebody recently told me that we contributed more to economic growth singularly than just about any company in the world today to the American economy. And I believe that

American economy. And I believe that that's probably true. And the reason for that is Nvidia is a multiundred billion dollar company supporting multiundred

billion dollar companies going after trillion trillion dollars of industry.

>> And so the economic prosperity the technology leadership unquestionable the technology the economic prosperity that

we can contribute to unquestionable.

Now this the question then becomes how do we think through the diffusion of the export of the proliferation of American technologies

and standards.

We should of course number one safeguard our national security the little little and little s to ensure that adversaries don't have access to

sensitive technology or advanced technology that we don't need them to have access to >> right >> we should number two ensure that American companies American technology

companies through partnership with us have the benefit of the best and first but then after that After number one and number two, we

should also proliferate American technology standards, compete around the world, fuel this flywheel of funding,

>> yeah, >> our R&D so that we can continue to be to be the mightiest technology industry in the world so that we can fund the tech

the mightiest military in the world and all of that I think goes hand in hand.

several times. Uh, you know, Chairman Dang, you've talked about energy as being a pacing problem here. We, uh,

you know, when we invented LED lighting, we just, uh, lost the demand signal and half of our uh, of our electricity u

grid is merchant supplier and so they don't buy ahead of need. And so we're now really way behind. And you pointed out that China

>> has built out twice the capacity of electricity than has the United States.

How big a constraint do you see that as being for our buildout for this revolution that you're trying to create?

Deeply serious.

I think at this point we have to use every form of energy we can. I believe

we can't rely on the power grid. We got

to build behind the meter. We obviously

need power generation systems. There's no question we should try to encourage and try to accelerate nuclear.

We need to have energy growth very, very shortly.

In the meantime, we're advancing our technology so quickly. No company our scale has ever

quickly. No company our scale has ever introduced new generations every year.

And when I say we just ship a new chip every year, people, you know, when people because there's so many gamers in the world and they've known me for so long and when they think about what

Nvidia builds, they think it's a a a module that looks like a gaming graphics card, our GeForce graphics card, and

they plug it into their PC. Well, a GPU for AI centers, AI AI data centers. That

GPU weighs two tons.

It has one and a half million parts.

It consumes $200,000 watts.

>> It costs $3 million.

>> Every so often, somebody says, you know, these GPUs are being smuggled. I really

would love to see it. Yeah.

>> Not to mention, you have to smuggle enough of them to fill a football field >> full of these things so that you could

run it as an AI data center. And so

anyhow, the technology that we make each year allows us to increase the performance at about the same power by

many times. And let me just pick a

many times. And let me just pick a number. Say five times or 10 times each

number. Say five times or 10 times each year. As a result, our energy efficiency

year. As a result, our energy efficiency improves by five times or 10 times each year.

>> But the problem is this. We're at the beginning of this technology buildout.

I'm improving the performance by a factor of 10 times each year, but demand is going up by a factor of 10,000 a million times each year.

>> AI is getting more computensive. The

adoption is going way up. I've got all these exponentials. And so we're going

these exponentials. And so we're going to keep chasing this. Uh we're going to be completely dedicated to advancing the technology as fast as we can. But the

bottom line is we need energy.

>> Yeah. And I I you know, forgive me for interjecting myself. I do think we have

interjecting myself. I do think we have to overcome the nimi constraints. You

know, we're going to have to find some structure of federal preeemption so we can overcome the the barriers. That's my

comment. That's not your comment. I

don't want you to get in trouble for for my saying that. Let me ask you, I mean, last year, >> thank you for that. [laughter]

>> Last year, um, the world installed two million robots.

Half of them were in China, which is really astounding when you think about it. Tell me how robots fit in with AI.

it. Tell me how robots fit in with AI.

Um, you know, let me just give you one example of why it's around the corner.

You know, these days you could describe, you could describe in text and you give it to um a video AI

and it generates a video. You guys know this, right? It actually from words you

this, right? It actually from words you can generate a video. Okay? And let's

say the video is uh Jensen reaches over, picks up a cup.

So I take a picture of this screenshot, give it to the AI. That's the starting starting condition. And I say, "Now

starting condition. And I say, "Now cause Jensen to reach over and pick up the cup."

the cup." The AI creates pixel by pixel, token by token, my arm picking up the cup. And

that everybody knows is possible today.

You guys have seen it. Well, the AI can't tell the difference between it manu manipulating pixels versus it's manipulating a bunch of motors.

So, the idea that I can tell the robot pick up the cup is clearly just around the corner. We just have to take that AI

the corner. We just have to take that AI which currently sits in the cloud and we have to put it into otherwise called embody it into a physical

>> mechanical system which is called robotics. So the AI is around the

robotics. So the AI is around the corner. We can see early evidence that

corner. We can see early evidence that the technology must be possible. Now

China is going to be very very good at this for several reasons. They have

great demand. They have a natural indigenous demand for more workers.

Manufacturing is core part of what they do. We, by the way, because we're now

do. We, by the way, because we're now re-industrializing reshoring manufacturing, we now again also have significant demand

for factory automation. And there's no question we have a shortage of labor. We

have right we all know that our industries would be would be larger more profitable more vibrant if we just had more workers >> and so they have the same challenge they

have worker shortage coming up very severe worker shortage coming up so they have a a a national strategic imperative to make sure that robotics happens number one number two they have the AI

technology and number three this is where they have the big advantage they're really very good at electronics and mechanical intersections

otherwise known as mechatronics. This

entire area is they have the harmony of demand and supply side capability.

>> Now many other countries Japan has surely demand side. They have

the megatronics but Japan needs to have much better AI technology. Germany great

demand, extraordinary meatronics.

They need to have great AI technology.

United States we have if we reshort industrial indust re-industrialize our nation we will have great demand. We

have great sa software technology but we really at the moment need to improve our mechanical electronics.

>> Yeah. I mean, you know, using AI to find a better, you know, vegan recipe for foyer gr, you know, maybe something my wife will look up, but we need to make this >> that would be a miracle indeed.

>> It would be good, but we we need to turn this into productive machinery and the way in which it's going to change the the landscape. Let me ask you, we're

the landscape. Let me ask you, we're because we're running out of time. Um,

you know, I was talking to a friend of mine who's a dean of a of a major research institute, and I asked him, I said, "How is your faculty dealing with

uh with AI?" And he said, "Well, you know, the engineering faculty is excited. They really think this is

excited. They really think this is fabulous." You said the science faculty

fabulous." You said the science faculty is really curious and they think it potentially opens up real opportunity and the humanities faculty thinks it's

the end of the world.

>> Um so it is a shorthand for the anxiety that people feel about the dark side

>> of AI. How do you how do you talk to us about that?

Um let me start from the from the end.

There's no question that everyone's jobs profession will be affected by AI

because the tasks within our jobs are going to be dramatically enhanced by AI.

>> Yeah.

>> Some jobs will become obsolete.

New jobs are going to be created >> and every job will be changed.

>> So that let me just I used two words just now and it's really important we think about these two words very differently. One is task the other one

differently. One is task the other one is job. Now, it turns out

is job. Now, it turns out I think it was something like seven, eight years ago