← 返回视频库

NVIDIA’s Jensen Huang on Securing American Leadership on AI

中级 ⏱ 52:08 职场

On Wednesday, December 3, 2025, at 5:00 p.m. ET, CSIS will host NVIDIA founder and CEO Jensen Huang for a fireside chat with CSIS president and CEO Dr. John J. Hamre on securing U.S. leadership in AI. As AI infrastructure becomes central to national security and economic competitiveness, Huang will offer insights from leading one of the most influential companies in AI. The discussion will cover policies to accelerate U.S. AI innovation, the role of public-private partnerships in building AI infrastructure, and how allied collaboration can strengthen national security. This event is made possible through general support to CSIS. --------------------------------------------- A nonpartisan institution, CSIS is the top national security think tank in the world. Visit www.csis.org to find more of our work as we bring bipartisan solutions to the world's greatest challenges. Want to see more videos and virtual events? Subscribe to this channel and turn on notifications: https://cs.is/2dCfTve Follow CSIS on: • Twitter: https://twitter.com/csis • Facebook: https://facebook.com/CSIS.org • Instagram: https://instagram.com/csis/

字幕文本（200 句）

Well, welcome to all of you out in
0:02
cyberspace. Uh, and I have a wonderful
0:06
collection of colleagues here in
0:07
physical space and we're going to have
0:10
an interesting conversation with Jensen
0:12
Hang. Um,
0:15
I would waste your time by introducing
0:17
him. You know, everybody knows Jensen,
0:20
but um, what you may not know is that he
0:25
he started from fairly humble roots.
0:28
Your mom was a school teacher. Your dad
0:31
was a petroleum engineer. You and I
0:34
share one thing in common is that's we
0:36
both started off our first job was
0:39
running a dishwashing machine in a res
0:41
in a restaurant.
0:42
>> Denny's you
0:44
>> what was what was your restaurant?
0:46
>> It was at Mount Rushmore when all the
0:48
you but you did better than I did. He
0:50
became he became a uh bus boy and later
0:54
a waiter and then I guess led him to
0:58
Nvidia the head of Nidia. You know it's
1:01
a remarkable story. It's I think it's
1:04
the quitessential American story. Yes.
1:07
>> That uh you know we welcome people who
1:10
come with just energy and imagination
1:13
and creativity and they make an
1:16
astounding success and
1:18
>> thank you. Congratulations and thank you
1:20
thank you for joining us. We're going to
1:23
uh we're going to have a very
1:24
interesting conversation today
1:27
colleagues and u you know it's u nvidi
1:30
is not only a huge economic success but
1:35
it's a national security platform and I
1:38
think we want to talk about that today.
1:41
I've been looking at your um at your
1:44
website and you do talk about uh
1:47
Nvidia's being a platform. What does
1:50
that mean? A platform? A platform is
1:55
something that you build other things
1:56
upon. Nvidia is the largest pure play
2:01
technology company the United States has
2:04
ever seen. In fact, we're the largest
2:06
pure play technology company the world's
2:07
ever known.
2:09
We create technology out of nothing. Our
2:13
final product is pure technology and in
2:18
order to use it, you have to create
2:20
software and applications for various
2:22
industries above it. You know, if you
2:24
look at most of the technology companies
2:25
today, some of it could be in social
2:27
media, some of it could be in
2:28
e-commerce, some of it it could be in
2:30
information search and and these are all
2:32
amazing technology companies whose
2:34
business models are something else. our
2:36
business model is purely technology. Now
2:38
the way that that AI works and and uh
2:42
our technology works is that in the
2:44
final analysis the technology platform
2:47
is built in layers and that's one of the
2:49
reasons why we think of it as a
2:50
platform. You're standing on top of it.
2:53
An application or an industry stands on
2:56
top of that platform. That platform
2:58
starts with energy on the bottom. One of
3:00
the reasons why why uh this
3:02
administration uh has made such a huge
3:05
difference right away if it's is this
3:07
pro- energy growth initiative its
3:10
attitude about energy is that if we
3:12
don't have energy we can't enable this
3:14
new industry to thrive it is absolutely
3:17
true so layer one is energy layer two
3:20
are essentially the chips and systems
3:22
but the chips that's where Nvidia comes
3:25
in layer three is a whole bunch of
3:28
software and we build a whole bunch of
3:29
software on top of our chips. And we're
3:31
well known for this piece of software
3:32
called CUDA, but there's hundreds of
3:34
different pieces of software that we
3:36
create that enables people to do AI for
3:39
different fields of science or language
3:41
or images or whatever it happens to be,
3:43
robotics or manufacturing and such.
3:44
>> But that third layer is called
3:46
infrastructure. Basically software.
3:48
>> Now people historically have thought of
3:51
infrastructure really as cloud. But
3:53
increasingly it's really important to
3:55
realize that infrastructure includes
3:57
land, power, shell because these are and
4:01
I'll speak about this in a second that
4:04
this industry
4:06
spawn another industry altogether and
4:08
I'll come back to that. But the third
4:10
layer is basically infrastructure and
4:12
that infrastructure includes financial
4:14
services because it takes an enormous
4:15
amount of capital to do what we do and
4:18
historically all of that software. the
4:20
layer above that and this is where
4:22
people largely focus on when they talk
4:25
about AI which is the AI models. This is
4:27
the revolutionary of course Chad GPT uh
4:31
incredible work that anthropic does with
4:33
cloud and uh Google does with Gemini and
4:37
uh what XAI does with Grock. But the
4:40
important thing to realize those are
4:42
four
4:44
of the one and a half million AI models
4:47
in the world.
4:49
AI
4:51
is not just intelligence that
4:53
understands English or language, but
4:56
it's AI that understands
4:59
genes, proteins, chemicals, the laws of
5:03
physics, AI that understands quantum. AI
5:07
that understands physical articulation,
5:11
otherwise known as robotics. AI that
5:14
understands patterns across long
5:16
sequence of time, financial services,
5:21
>> AI that understand longitudinally across
5:24
multiple modalities, healthcare. And so
5:27
AI has all of these different reaches
5:30
and domains. We talk about this one
5:32
area. We just have to be very careful to
5:35
understand that AI AC spans basically
5:38
every form of information across every
5:41
field of science across every single
5:43
industry.
5:45
One and a half million AIs around the
5:47
world. On top of that are all the
5:50
applications and we never should forget
5:52
that in the final analysis these AI
5:54
models are technologies
5:57
but technologies are about enabling
6:00
application and use whether you're in
6:03
health care
6:05
you know you could be in entertainment
6:07
manufacturing self-driving cars
6:09
transportation each one of these
6:11
industries have AIs that deeply affect
6:14
them and these are the five layer stack
6:18
Nvidia is at the lower level, the
6:21
platform level. The reason why we say
6:24
that, you know, we're AI company that
6:27
works with every single AI company in
6:28
the world is because we're the platform
6:30
by which we are able to work with all of
6:33
these technology companies and all these
6:35
application companies across all these
6:36
industries. And so that is a platform
6:39
that we created. the mode that people
6:42
describe, not so much a mode, but
6:44
basically the the language by which all
6:47
of these different applications and
6:50
these different technologies speak to us
6:52
is an architecture we invented 25 years
6:55
ago called CUDA. And on top of that CUDA
6:58
libraries, a whole bunch of algorithms
7:00
that we invented over the years and that
7:02
is basically Nvidia's platform. In the
7:06
end, we don't build self-driving cars,
7:08
but we work with every single
7:09
self-driving car company in the world.
7:11
In the end, we don't discover drugs, but
7:14
we just have we work with every single
7:16
drug discovery company in the world.
7:18
We're the platform by which they build
7:20
their things. We're a platform company.
7:22
>> I I think half of I had a half a dozen
7:25
young kids come up to me today say
7:27
Nvidia gave them a better gaming
7:29
experience. I mean you're all here.
7:32
>> Before [laughter] I before we invented
7:34
the AI industry,
7:36
>> the first industry we created is the
7:38
modern gaming industry. Absolutely.
7:40
Yeah. You you don't know this. I'm very
7:42
proud of this. Nvidia is the world's
7:44
largest gaming platform.
7:46
>> Yeah. And
7:47
>> probably probably didn't know that there
7:49
>> and that's how most of the kids were
7:51
talking to me about that today. I want
7:53
to see a
7:54
>> 300 million active users. Yeah. a 100
7:57
million of them Nintendo Switch. Yeah,
8:01
Nvidia's in that.
8:02
>> Okay. So, let me ask you. You said
8:05
something um recently that was quite
8:08
provocative. You you said that China was
8:12
winning the AI
8:14
race, the AI competition. Um
8:19
um I know that you've got a powerful,

在 Linglish 上深度学习这个视频

AI 助手 · 单词高亮 · 影子跟读 · 间隔重复 · 完全免费

立即免费学习