Romantic Piano - Wedding Background Music For Videos (Royalty Free Music) - by AShamaluevMusic

Romantic Rhythms
126 followers
AI Generated·May contain errors
Welcome back. As AI continues to transform industries across the globe, the world's
largest semiconductor company is racing to keep up. I sat down with the CEO of NVIDIA,
Jensen Fong, and he said demand for his chips is exploding.
It's for sure now we're going to have a very large 2026.
You said that you had a backlog, a big backlog, $500 billion in backlog.
Yes.
Tell us about that. Where are you on that?
Well, since I said that a few months ago, that backlog has gotten bigger than that.
And so I think we're just seeing AI has gone from these language models to now going into
healthcare, into drug discovery, into manufacturing. This last year we made huge
breakthroughs in this area called physical AI, where AI understands the world. It understands
causality and the laws of physics, and you can use it for self-driving cars. And of course,
we're going to see huge breakthroughs in self-driving cars. It's going to be all over
robotic systems of all kinds.
Okay. So the Rubin, what is the most advanced chip?
It's 10 times more energy efficient than Blackwell. It's 10 times lower cost for generating
tokens than Blackwell, 10 times. And Blackwell was 10 times that of Hopper. So every single year,
we're driving up energy efficiency and we're driving down the token costs. This way AI can
continue to be affordable, more clever, more smart, while being more affordable for everybody
to use.
There's a debate on whether or not China should actually be getting our chips.
What's your take? You want to obviously sell to that market,
but are there risks with regard to national security?
First of all, H200 is not used by their military because their military, like our military,
builds on our own chips. And they build plenty of great chips, Huawei and many new startups.
They've all gone public this year, hundreds of billions of dollars of market value
now in building AI chips. So they have lots of AI chips for themselves.
With respect to H200, remember the AI stack is five layers deep and we want to win the AI race.
And we need to win the AI race at every single layer. We want to win the energy layer. We want
to win the chip layer, the infrastructure layer, the AI model layer, and also the application layer.
President Trump wants us to go win the marketplace, compete around the world, bring home jobs,
create revenue for the United States, and continue to lead the technology around the world.
We can't concede any market. And so having the ability to keep the United States at the forefront
with Blackwell today and very soon Rubin, but allowing American chip companies to compete
around the world, we have the benefit of having a thoughtful and balanced strategy.
I want to ask you about the question that you hear all the time. And that is with all of this
money being spent on AI, are we inching toward bubble territory?
Future R&D budget is going to drive enormous amount of infrastructure. We're going to be,
you know, it's very likely this is the largest infrastructure buildout in human history.
And over the course of the next 15 years, a lot of estimates come in at, you know,
$85 trillion. And so we're going to build a lot of infrastructure.
So if you're talking about a couple of trillion dollars every year going into this,
what do you say to those people who say that's absolutely not going to happen? It's a bubble,
and we're going to have some companies holding the bag of banks who are lending out this money?
Because remember, some of these companies who are investing in AI, they're borrowing the money.
Is it a bubble?
Well, during the time when countries, when companies build infrastructure,
it is not unusual to borrow money. However, we're nowhere near a bubble right now.
And the reason why we know that is because every single GPU that NVIDIA has in the cloud,
and they're in Azure, they're in AWS, they're in, you know, their GCP,
they're all rented out. They're in CoreWeave, they're all rented out.
And in fact, the demand is so high for our GPUs that the spot price of GPUs we sold
a couple of years ago are going up. And so that tells you something.
And so demand for Blackwell is incredible. Demand for our next generation product,
Rubin, is incredible. Demand for our past generations in the cloud is incredible.
My thanks to Jensen Huang. I've got one more thing you need to know about ahead of next week.
That's next.
