WEBVTT

00:00:00.480 --> 00:00:02.717
Minister of Communications and

00:00:02.800 --> 00:00:07.757
Information Technology of the Kingdom of
Saudi Arabia. Your Excellency, thank you

00:00:07.840 --> 00:00:09.677
so much. Good [music] to see you. Also

00:00:09.760 --> 00:00:13.017
with us this morning

00:00:13.100 --> 00:00:16.237
[music] are two incredible visionaries.

00:00:16.320 --> 00:00:18.157
Please join me in welcoming to the stage

00:00:18.240 --> 00:00:20.877
Elon Musk, the CEO of Tesla, founder of

00:00:20.960 --> 00:00:23.757
SpaceX, founder of XAI, and Jensen

00:00:23.840 --> 00:00:26.777
Huang, founder and CEO of Nvidia.

00:00:26.860 --> 00:00:28.880
[music]

00:00:31.440 --> 00:00:33.460
>> [music]

00:00:35.120 --> 00:00:37.917
>> Please, please have a seat. Now, I'm

00:00:38.000 --> 00:00:39.597
sure we can do a bigger round of

00:00:39.680 --> 00:00:42.077
applause for one of the greatest two

00:00:42.160 --> 00:00:46.397
leaders of our history. Let's go ahead.

00:00:46.480 --> 00:00:50.237
So, we're talking about I lost count,

00:00:50.320 --> 00:00:52.477
you know, 7 to8 trillion dollars worth

00:00:52.560 --> 00:00:56.717
of market cap comp. I lost count. But

00:00:56.800 --> 00:00:59.677
right now we're here to celebrate a

00:00:59.760 --> 00:01:03.837
historic moment. A moment that yesterday

00:01:03.920 --> 00:01:06.077
during the dinner and thank you for for

00:01:06.160 --> 00:01:09.757
joining us under the patronage of the

00:01:09.840 --> 00:01:13.757
honorable president and his royal
highness the crown prince Musidi where

00:01:13.840 --> 00:01:15.437
we had the pleasure to hear it

00:01:15.520 --> 00:01:19.437
firsthand. This is the greatest alliance

00:01:19.520 --> 00:01:22.397
between the Kingdom of Saudi Arabia and

00:01:22.480 --> 00:01:25.117
the United States where we have joined

00:01:25.200 --> 00:01:27.757
hands and you have helped us build our

00:01:27.840 --> 00:01:30.477
energy based economy fueling and

00:01:30.560 --> 00:01:33.277
energizing the industrial age and now

00:01:33.360 --> 00:01:35.437
fast forward going to the intelligence

00:01:35.520 --> 00:01:39.597
age where we can fuel AI factories

00:01:39.680 --> 00:01:44.157
robotics EVs and all
of the rest. Speaking of

00:01:44.240 --> 00:01:45.997
that, let's start with you, Elon, if you

00:01:46.080 --> 00:01:47.677
don't mind, Jensen, and feel free to

00:01:47.760 --> 00:01:51.197
chive in. You have a big fascination of

00:01:51.280 --> 00:01:53.917
something all of us have admired, first

00:01:54.000 --> 00:01:56.077
order thinking, which Jensen sometimes

00:01:56.160 --> 00:01:59.277
calls first order scaling, which is an

00:01:59.360 --> 00:02:02.557
opportunity for how you have dropped the

00:02:02.640 --> 00:02:05.037
cost of batteries from a,000 for

00:02:05.120 --> 00:02:09.037
kilowatt hour to subund bucks. And right

00:02:09.120 --> 00:02:10.637
now you're doing the same thing with

00:02:10.720 --> 00:02:15.037
robotics for actuators with servo rotors

00:02:15.120 --> 00:02:18.397
and motors. So I want to hear from you.

00:02:18.480 --> 00:02:22.957
How do you manage to always disrupt
every single industry with that

00:02:23.040 --> 00:02:24.797
thinking?

00:02:24.880 --> 00:02:27.677
>> Well, it's mostly not disruption, it's

00:02:27.760 --> 00:02:32.397
uh creation. So with uh say SpaceX with

00:02:32.480 --> 00:02:35.517
the reusable rockets, uh there really

00:02:35.600 --> 00:02:39.197
weren't any reusable rockets. Um,

00:02:39.280 --> 00:02:44.077
but the essence of getting
of revolutionizing space travel is

00:02:44.160 --> 00:02:45.917
reusability. If you throw the rocket

00:02:46.000 --> 00:02:48.397
away every time, the cost of access to

00:02:48.480 --> 00:02:51.757
space is extremely high. Um, with

00:02:51.840 --> 00:02:56.237
respect to electric cars, there there
weren't any electric cars when we

00:02:56.320 --> 00:03:01.117
started making them really. You couldn't
buy any to the best of my knowledge. Um,

00:03:01.200 --> 00:03:03.037
so with Tesla, we wanted to make

00:03:03.120 --> 00:03:05.197
electric cars compelling um, and

00:03:05.280 --> 00:03:08.957
affordable. That was the goal. Um the uh

00:03:09.040 --> 00:03:13.197
you know with respect to humanoid
robotics, there are no useful humanoid

00:03:13.280 --> 00:03:17.277
robotics robots at this point. Um there
are sort of gimmicks, but there are

00:03:17.360 --> 00:03:22.077
there are no actually useful humanoid
robots. Um and I think Tesla's going to

00:03:22.160 --> 00:03:26.157
make the first actually useful humanoid
robots. Um and this will be quite a

00:03:26.240 --> 00:03:31.197
revolution and I think something that will
that everyone will want. Uh because

00:03:31.280 --> 00:03:36.477
I always think of like who who wouldn't
want their own personal C3PO R2-D2.

00:03:36.560 --> 00:03:40.077
>> Oh yeah, >> of course.
Everyone would want one,

00:03:40.160 --> 00:03:44.957
right? And and then there would be many
in industry uh providing products and

00:03:45.040 --> 00:03:46.957
services. This is why I say that

00:03:47.040 --> 00:03:51.997
humanoid robots will be the biggest
industry or the biggest product ever. Um

00:03:52.080 --> 00:03:56.877
bigger than cell phones or anything else
because everyone's going to want one and

00:03:56.960 --> 00:03:58.557
uh or maybe more more than one. And

00:03:58.640 --> 00:04:01.997
there will be many in industry. Um,

00:04:02.080 --> 00:04:05.677
>> I just want R2-D2 in C3PO's body.

00:04:05.760 --> 00:04:09.597
>> Yeah, >> there you go. >> Well,
I mean, a humanoid robot would be

00:04:09.680 --> 00:04:11.757
better than R2-D2 and C3vo combined.

00:04:11.840 --> 00:04:13.437
>> Yeah. >> Times 10. >> Yeah.

00:04:13.520 --> 00:04:17.037
>> So, the it and and you know, people

00:04:17.120 --> 00:04:20.157
often talk about uh sort of eliminating

00:04:20.240 --> 00:04:24.157
poverty and that kind of thing,
but really the there's how long have they

00:04:24.240 --> 00:04:28.477
been talking about that? Um there's
lots of talk uh you know there's lots of

00:04:28.560 --> 00:04:30.717
NOS's sort of trying to do these things

00:04:30.800 --> 00:04:33.757
but but really not succeeding um and and

00:04:33.840 --> 00:04:35.997
and you know the evidence speaks for

00:04:36.080 --> 00:04:39.837
itself uh but but but AI and humanoid

00:04:39.920 --> 00:04:44.397
robots will actually eliminate poverty
and Tesla won't be the only one that

00:04:44.480 --> 00:04:47.437
makes them. I think Tesla will pioneer
this, but there will be many other

00:04:47.520 --> 00:04:52.317
companies that make humanoid robots. But
there there is only basically one way to

00:04:52.400 --> 00:04:56.157
uh make everyone wealthy,
and that is AI and robotics.

00:04:56.240 --> 00:04:58.397
>> And we can't talk
about robotics without

00:04:58.480 --> 00:05:01.357
AI factories. And yesterday was such a

00:05:01.440 --> 00:05:04.317
historic day for the two nations but

00:05:04.400 --> 00:05:07.677
also for all of us where we celebrate

00:05:07.760 --> 00:05:10.797
the AI strategic partnership with the US

00:05:10.880 --> 00:05:13.437
s signed witnessed by the honorable

00:05:13.520 --> 00:05:15.917
president and his royal highness about

00:05:16.000 --> 00:05:19.837
how we are committing our capital energy

00:05:19.920 --> 00:05:24.797
land to energize the AI US ecosystem to

00:05:24.880 --> 00:05:26.797
be able to build inference node training

00:05:26.880 --> 00:05:29.837
nodes and to be the most AI enabled

00:05:29.920 --> 00:05:34.237
nation with that announcement. Tell
me what's what's next in AI factories.

00:05:34.320 --> 00:05:36.957
Jensen, >> there there's
a there's a beautiful

00:05:37.040 --> 00:05:39.517
story about how Saudi Arabia's building

00:05:39.600 --> 00:05:43.997
AI refineries now building AI
facil refineries to AI factories.

00:05:44.080 --> 00:05:47.837
>> I love that. >> I you know
I've said that that AI is an

00:05:47.920 --> 00:05:51.117
infrastructure and the reason for that
of course we understand AI from the

00:05:51.200 --> 00:05:55.597
perspective of the technology and how
it's revolutionizing every industry.

00:05:55.680 --> 00:05:57.597
Digital intelligence of course has

00:05:57.680 --> 00:06:01.197
applications into every every field and

00:06:01.280 --> 00:06:02.797
so it's going to be used by every

00:06:02.880 --> 00:06:05.357
company, every industry, every country.

00:06:05.440 --> 00:06:08.637
In that way it's foundational and

00:06:08.720 --> 00:06:10.397
therefore it's part of infrastructure.

00:06:10.480 --> 00:06:14.477
What is new about AI from a computer

00:06:14.560 --> 00:06:17.757
science perspective is that the way

00:06:17.840 --> 00:06:22.157
computing was done in the past
was largely retrievalbased computing.

00:06:22.240 --> 00:06:24.717
Somebody typed in a story or somebody

00:06:24.800 --> 00:06:27.757
created a a piece of art or came came up

00:06:27.840 --> 00:06:30.637
with four versions of a digital ad or

00:06:30.720 --> 00:06:34.477
it's all pre-built by somebody which is

00:06:34.560 --> 00:06:37.197
then using a system to retrieve the

00:06:37.280 --> 00:06:38.797
appropriate version for you. It's a

00:06:38.880 --> 00:06:42.397
retrievalbased computing model. Hadoop

00:06:42.480 --> 00:06:47.277
and many of the the the frameworks
and operating systems of the past all

00:06:47.360 --> 00:06:51.997
designed to retrieve the appropriate
information for you. But today software

00:06:52.080 --> 00:06:54.237
is going to be generated in real time.

00:06:54.320 --> 00:06:58.397
It's generative based on the context,
based on the circumstance, based on who

00:06:58.480 --> 00:07:02.237
you are, based on the problem you
ask that based on your prompt. It will

00:07:02.320 --> 00:07:07.197
generate unique content for you every
single time for everybody. It's unique.

00:07:07.280 --> 00:07:11.117
When you use Grock, every time you
use it is different just based on right

00:07:11.200 --> 00:07:14.397
based on based on the based on the prompt
that you give it and based on the

00:07:14.480 --> 00:07:17.597
circumstance and and so therefore it

00:07:17.680 --> 00:07:19.997
used to be retrievalbased

00:07:20.080 --> 00:07:22.557
today it's generative and if it's

00:07:22.640 --> 00:07:27.197
generative then and every time is
different then you need AI factories all

00:07:27.280 --> 00:07:30.877
over the world to generate the content
in real time which is the reason why you

00:07:30.960 --> 00:07:34.877
need AI factories and and this is a
unique way of doing computation but the

00:07:34.960 --> 00:07:39.517
benefit of course is that Everything
isn't preconceived and pre-documented

00:07:39.600 --> 00:07:43.677
and it's it's contextually s
contextually sensible and and therefore

00:07:43.760 --> 00:07:47.277
intelligent. >> So AI
factories and robotics and we

00:07:47.360 --> 00:07:49.037
heard it yesterday from his royal

00:07:49.120 --> 00:07:51.997
highness his vision how to augment our

00:07:52.080 --> 00:07:54.717
workforce with roughly tens of millions

00:07:54.800 --> 00:07:57.677
of robotics to be able to infuse the

00:07:57.760 --> 00:08:00.317
next wave of productivity and progress.

00:08:00.400 --> 00:08:02.797
But this scares a lot of folks here when

00:08:02.880 --> 00:08:07.357
it comes to the future of jobs. So let's
hear about your thoughts uh Elon and

00:08:07.440 --> 00:08:09.757
Jensen on that.

00:08:09.840 --> 00:08:15.037
>> Uh sure. Well um you say
like in the long term where

00:08:15.120 --> 00:08:19.437
will things end up longterm?
I don't know what long term is. Maybe it's

00:08:19.520 --> 00:08:21.437
10 20 years something like that. For me

00:08:21.520 --> 00:08:26.557
that's long term. Um my prediction
is that work will be optional.

00:08:26.640 --> 00:08:31.757
>> Optional. >> Optional. Um so

00:08:31.840 --> 00:08:35.837
>> we'll take that. >> Yeah.
I mean it it'll be like uh playing

00:08:35.920 --> 00:08:37.517
sports or a video game or something like

00:08:37.600 --> 00:08:40.797
that. Um if you want to work uh you know

00:08:40.880 --> 00:08:43.117
in the same way like you can you can go

00:08:43.200 --> 00:08:47.997
to the store and just buy some
vegetables or you could grow vegetables

00:08:48.080 --> 00:08:51.677
in your backyard. It's much harder
to grow vegetables in your backyard but

00:08:51.760 --> 00:08:53.277
some people still do it because they

00:08:53.360 --> 00:08:55.837
like growing growing vegetables. Um that

00:08:55.920 --> 00:09:00.717
will be what work is like optional.
Um and between now and then there's

00:09:00.800 --> 00:09:02.957
actually a lot of work
to get to that point.

00:09:03.040 --> 00:09:05.757
>> Mh. Um, and I always recommend people

00:09:05.840 --> 00:09:08.797
read read Yin Banks uh culture books to

00:09:08.880 --> 00:09:12.957
get a sense for what a probable

00:09:13.040 --> 00:09:16.477
positive AI future is like. Um, and

00:09:16.560 --> 00:09:18.797
interestingly in those books, money is

00:09:18.880 --> 00:09:22.557
no longer doesn't exist. It's kind of

00:09:22.640 --> 00:09:25.597
interesting. And I I my guess is in in

00:09:25.680 --> 00:09:29.677
if you go out long enough, assuming
there's a continued improvement in AI

00:09:29.760 --> 00:09:32.317
and robotics, which there seems likely,

00:09:32.400 --> 00:09:35.357
the money will will stop being relevant

00:09:35.440 --> 00:09:38.397
at some point in the future. Um now

00:09:38.480 --> 00:09:41.357
there will still be constraints on power

00:09:41.440 --> 00:09:45.117
like electricity and mass. Uh the

00:09:45.200 --> 00:09:47.117
fundamental physics elements will still

00:09:47.200 --> 00:09:51.117
be still be constraints. Um but um I

00:09:51.200 --> 00:09:55.437
think at some point uh currency becomes

00:09:55.520 --> 00:09:58.237
irrelevant.

00:09:58.320 --> 00:10:00.877
>> Jensen, any thoughts?

00:10:00.960 --> 00:10:05.917
>> Um [laughter]

00:10:06.000 --> 00:10:09.997
by the way, the Nvidia earnings
call is later today. [laughter]

00:10:10.080 --> 00:10:13.997
>> And by the way,
since currency is relevant.

00:10:14.080 --> 00:10:18.237
>> Cheers. >> [laughter]

00:10:18.320 --> 00:10:24.280
>> Elon just wants to share with you
some breaking news. [laughter]

00:10:26.840 --> 00:10:31.597
[laughter] The two of us who like
to share some breaking news. Uh let's see.

00:10:31.680 --> 00:10:36.237
I I would say I would say there there's
um different horizons you could look at.

00:10:36.320 --> 00:10:38.397
Everybody's jobs will be different. That

00:10:38.480 --> 00:10:43.437
I think that that's for sure. Uh how
how the students learn will be different.

00:10:43.520 --> 00:10:47.197
um uh how people do their work will be
different obviously because a lot of the

00:10:47.280 --> 00:10:49.677
things that that we do mundanely or

00:10:49.760 --> 00:10:52.477
arduously or very difficultly are going

00:10:52.560 --> 00:10:57.117
to be done very simply and and so we're
going to be more productive from that s

00:10:57.200 --> 00:10:59.037
from that sense one of the things that I

00:10:59.120 --> 00:11:02.717
will say is that for most people or a

00:11:02.800 --> 00:11:06.877
company if some if your life
becomes more productive

00:11:06.960 --> 00:11:09.357
and if the things that you're doing uh

00:11:09.440 --> 00:11:14.077
with great difficulty becomes simpler it
is very likely because you have so many

00:11:14.160 --> 00:11:18.157
ideas, you'll have more time to go
pursue things. It is my guess that Elon

00:11:18.240 --> 00:11:22.477
will be busier as a result of AI,
I'm going to be busier as a result of AI.

00:11:22.560 --> 00:11:26.237
And the reason for that is because we
have so many ideas we want to pursue, so

00:11:26.320 --> 00:11:30.077
many things that that we still have
in our backlog inside our company that we

00:11:30.160 --> 00:11:34.077
can go pursue. If we were more
productive, we can get to those things

00:11:34.160 --> 00:11:36.717
faster. And so in the near term I would

00:11:36.800 --> 00:11:41.197
say that that there is every evidence
that that we will be more productive and

00:11:41.280 --> 00:11:45.517
yet still be busier because we have so
many ideas. One thing that I will say

00:11:45.600 --> 00:11:48.157
give you give you some evidence is that

00:11:48.240 --> 00:11:49.597
uh and I was just telling Elon about

00:11:49.680 --> 00:11:52.717
this earlier radiology for example has

00:11:52.800 --> 00:11:55.917
largely been uh converted to AIdriven

00:11:56.000 --> 00:11:57.517
radiology and there's some really great

00:11:57.600 --> 00:12:00.557
companies doing that and the surprising

00:12:00.640 --> 00:12:03.517
thing is the prediction that all

00:12:03.600 --> 00:12:05.677
radiologists would be the first jobs to

00:12:05.760 --> 00:12:08.717
go was exactly the opposite. The trend

00:12:08.800 --> 00:12:14.157
shows that there are more radiologists
being hired now as a result of AI. And

00:12:14.240 --> 00:12:17.917
the reason for that, if you take
a step back, it's because the goal of a

00:12:18.000 --> 00:12:20.797
radiologist is not to study the images.

00:12:20.880 --> 00:12:23.277
The goal of a radiologist is to diagnose

00:12:23.360 --> 00:12:28.157
a disease. Now, the studying of the
images became so productive. They could

00:12:28.240 --> 00:12:32.237
study more images, study more
modalities, spend more time with the

00:12:32.320 --> 00:12:36.397
patients. And as a result, they were
actually accepting more patients. We're

00:12:36.480 --> 00:12:39.277
doing more radiology all around the world.
We're doing a better job with

00:12:39.360 --> 00:12:41.837
diagnosing disease. And so that's that's

00:12:41.920 --> 00:12:45.837
kind of the the nearterm near uh outcome

00:12:45.920 --> 00:12:48.957
of of uh AI and productivity. And and

00:12:49.040 --> 00:12:53.757
we'll see we'll see what
happens long term. You know, I I

00:12:53.840 --> 00:12:57.437
when when currency doesn't matter
anymore, just, you know, let me know

00:12:57.520 --> 00:13:00.077
right before.

00:13:00.160 --> 00:13:04.477
>> You'll see it coming.
[laughter] You'll see it coming. like like

00:13:04.560 --> 00:13:06.717
>> we text often so just >> Yeah, we do.

00:13:06.800 --> 00:13:08.637
>> Yeah, just text it off.
Yeah, let me know.

00:13:08.720 --> 00:13:12.557
>> I kind of I kind of agree with with
both of you because if you look at every

00:13:12.640 --> 00:13:15.277
technological trend, every general

00:13:15.360 --> 00:13:20.557
purpose technology has been net new
positive for for the globe for humanity

00:13:20.640 --> 00:13:22.877
and so forth. And let me
share with you two nice

00:13:22.960 --> 00:13:26.397
>> statements and and your excellency
I think it's precisely the reason the

00:13:26.480 --> 00:13:30.717
reason for that is because all the great
ideas from from innovators like Elon

00:13:30.800 --> 00:13:34.637
that you have so many good
ideas >> that and Jensen as well.

00:13:34.720 --> 00:13:38.557
>> Yeah. Well, thank you.
>> So let me share with you two stories

00:13:38.640 --> 00:13:42.317
from two Saudi innovators
in collaboration with a lot of great the

00:13:42.400 --> 00:13:44.877
great work that Nvidia does that Gro

00:13:44.960 --> 00:13:47.917
does. One is professor

00:13:48.000 --> 00:13:49.597
Maryagi.

00:13:49.680 --> 00:13:53.197
All right, let me say that again. It's

00:13:53.280 --> 00:13:59.080
Professor Omaragi. Who is

00:13:59.200 --> 00:14:01.517
>> I might need to >>
might need to move the mic.

00:14:01.600 --> 00:14:04.477
>> All right,

00:14:04.560 --> 00:14:08.760
>> here. Come on. We'll share this one.

00:14:08.880 --> 00:14:11.277
>> Get closer. >> Let's
let's let's try this one more

00:14:11.360 --> 00:14:13.357
time. So one of them is professor

00:14:13.440 --> 00:14:16.557
Omaragi who's the first American Saudi

00:14:16.640 --> 00:14:19.277
to win a Nobel prize in creating new

00:14:19.360 --> 00:14:21.677
chemistry and the way he has done that

00:14:21.760 --> 00:14:24.157
he has leveraged your AI accelerators

00:14:24.240 --> 00:14:28.877
and models like Grock to be able
to create new chemistry when it comes to

00:14:28.960 --> 00:14:31.037
metal organic frameworks. Those are

00:14:31.120 --> 00:14:33.517
metal ions that are positively charged

00:14:33.600 --> 00:14:36.237
with organic linkers to be able to

00:14:36.320 --> 00:14:39.837
effectively create a sponge with 33

00:14:39.920 --> 00:14:43.277
nanometers pores to capture water from

00:14:43.360 --> 00:14:47.357
air and also to capture carbon dioxide.

00:14:47.440 --> 00:14:50.477
The second story has also to do with AI

00:14:50.560 --> 00:14:53.277
accelerated by Nvidia and with models

00:14:53.360 --> 00:14:55.837
like Brock which is nanop which is

00:14:55.920 --> 00:14:59.037
effectively creating a nano robot 500

00:14:59.120 --> 00:15:01.677
nanometers by a,000 nanometers to be

00:15:01.760 --> 00:15:04.397
able to do gene editing leveraging the

00:15:04.480 --> 00:15:06.957
crisper technology to take out sickle

00:15:07.040 --> 00:15:09.117
cell disease. Now in both these

00:15:09.200 --> 00:15:11.917
instances they originated 20 years ago

00:15:12.000 --> 00:15:15.677
in research but AI was able to really

00:15:15.760 --> 00:15:20.877
accelerate the outcomes and the outputs
such that we can move into new value

00:15:20.960 --> 00:15:25.357
pools. So I think with every
technological trend humanity is going to

00:15:25.440 --> 00:15:29.997
always manage to shift to new value
pools when it comes to workforce and

00:15:30.080 --> 00:15:33.917
productivity. But we have some great
announcements to talk about here today.

00:15:34.000 --> 00:15:38.237
Let's begin with you Elon. the things
that we're doing with XAI. Uh yeah,

00:15:38.320 --> 00:15:40.397
we're excited to announce that um we're

00:15:40.480 --> 00:15:44.957
doing a a a uh a 500 megawatt I mean

00:15:45.040 --> 00:15:48.877
yeah 500 sorry >> 500 megawatt

00:15:48.960 --> 00:15:52.397
>> 500 megawatt yeah >>
we're doing 500 [laughter]

00:15:52.480 --> 00:15:56.397
>> we're doing 500 megawatt >> sorry

00:15:56.480 --> 00:15:58.397
>> yeah yeah the 500 g one will have to

00:15:58.480 --> 00:16:02.717
wait um so um that that'll be eight

00:16:02.800 --> 00:16:06.057
bazillion trillion dollars Um,

00:16:06.140 --> 00:16:09.837
[laughter] stop that.

00:16:09.920 --> 00:16:13.037
>> Uh, [laughter]

00:16:13.120 --> 00:16:17.597
so yeah, we we're we're doing a um XCAI

00:16:17.680 --> 00:16:20.397
and the Kingdom of Saudi
Arabia are doing a a

00:16:20.480 --> 00:16:24.557
>> Humane 500 megawatt starting with 50
megawatts phase one and we're doing it

00:16:24.640 --> 00:16:26.957
with Nvidia. Congratulations to the

00:16:27.040 --> 00:16:29.437
Humane team to target team. Such a

00:16:29.520 --> 00:16:32.637
fantastic job. Jensen, I think we're

00:16:32.720 --> 00:16:37.757
also doing some great announcements
uh this week. We are [laughter]

00:16:37.840 --> 00:16:41.037
we're announcing we're announcing all
kinds of things. Um uh our partnership

00:16:41.120 --> 00:16:43.037
with Humane is is going incredibly well.

00:16:43.120 --> 00:16:45.917
First of all, uh we we uh work together

00:16:46.000 --> 00:16:48.237
to get this company started and off the

00:16:48.320 --> 00:16:51.197
ground and just got an incredible

00:16:51.280 --> 00:16:54.877
customer with Elon. Could you imagine a

00:16:54.960 --> 00:16:57.037
startup company approximately zero

00:16:57.120 --> 00:17:00.157
billion dollars in revenues now going to

00:17:00.240 --> 00:17:02.397
build a data center for Elon 500

00:17:02.480 --> 00:17:04.957
megawatts is gigantic. Uh this company

00:17:05.040 --> 00:17:07.437
is off the charts uh right away. Uh in

00:17:07.520 --> 00:17:10.157
addition to that we're um uh working

00:17:10.240 --> 00:17:13.357
working uh AWS as you know is also

00:17:13.440 --> 00:17:17.517
>> congratulations
to the humane team with AWS

00:17:17.600 --> 00:17:19.357
>> starting with 100 megawws with a

00:17:19.440 --> 00:17:22.957
gigawatt ambition and uh counting. So

00:17:23.040 --> 00:17:25.517
AWS is also coming to Humane. Uh we're

00:17:25.600 --> 00:17:28.237
working with uh Humane on Omniverse uh

00:17:28.320 --> 00:17:31.757
digital twins. Uh as you know that AI is

00:17:31.840 --> 00:17:35.357
not just well just agentic AI and chat

00:17:35.440 --> 00:17:40.237
bots and uh cognitive AI is incredibly
important to the world. Uh but AI

00:17:40.320 --> 00:17:45.037
applies to everything, chemicals and
proteins and genes and physics and fluid

00:17:45.120 --> 00:17:49.517
dynamics and particles and of course
robotics and activation and um and we

00:17:49.600 --> 00:17:51.517
created this world called Omniverse

00:17:51.600 --> 00:17:53.197
where robots can learn how to be good

00:17:53.280 --> 00:17:56.557
robots and and it's physically based. It

00:17:56.640 --> 00:18:00.797
obeys the laws of physics and so robots
can learn in these environments and

00:18:00.880 --> 00:18:05.917
we're working with Humane to apply
Omniverse to all kinds of uh digital

00:18:06.000 --> 00:18:09.437
factories and robotics and warehouses
and things like that. And so that's

00:18:09.520 --> 00:18:13.677
that's another uh we're we're also
working uh in Saudi Arabia to build

00:18:13.760 --> 00:18:15.837
supercomputers to simulate quantum

00:18:15.920 --> 00:18:19.037
computers and and uh uh using our

00:18:19.120 --> 00:18:23.437
computers to be the controller
and the error correction quantum error

00:18:23.520 --> 00:18:27.517
correction requires an enormous amount
of computation and so so we're doing a

00:18:27.600 --> 00:18:31.117
lot of great work there too. So a big
partnership with Humane they're off the

00:18:31.200 --> 00:18:35.037
charts um off the ground and off the
charts at the same time. This is how we

00:18:35.120 --> 00:18:37.357
walk the talk in the Kingdom of Saudi

00:18:37.440 --> 00:18:39.917
Arabia in partnership with the US.

00:18:40.000 --> 00:18:42.317
Yesterday, the president and his Royal

00:18:42.400 --> 00:18:44.317
Highness announced the AI strategic

00:18:44.400 --> 00:18:47.037
framework and partnership. Today, we're

00:18:47.120 --> 00:18:53.880
going big with Elon and Jensen.
So, thank you for those opportunities.

00:18:54.470 --> 00:18:55.997
[applause]

00:18:56.080 --> 00:18:58.637
Now, they they told me I have time for

00:18:58.720 --> 00:19:01.437
two last questions. So last night at the

00:19:01.520 --> 00:19:03.917
dinner I got a number of questions

00:19:04.000 --> 00:19:08.477
because it seems that the schedule
leaked and uh everybody was giving me

00:19:08.560 --> 00:19:10.317
hints about the last two questions I'm

00:19:10.400 --> 00:19:12.397
going to do. So the first one was for

00:19:12.480 --> 00:19:16.477
you Elon and there's a big one for you

00:19:16.560 --> 00:19:20.077
Jensen. So prepare for that one. AI in

00:19:20.160 --> 00:19:23.197
space is that possible?

00:19:23.280 --> 00:19:26.797
Uh yes, if if civilization continues,

00:19:26.880 --> 00:19:33.757
which it probably will,
uh then AI in space is inevitable. Um

00:19:33.840 --> 00:19:38.717
um you know, I always have
to like preface that, you know,

00:19:38.800 --> 00:19:42.717
we shouldn't take civilization for
granted. We we need to make sure to take

00:19:42.800 --> 00:19:45.357
care to ensure that civilization has an

00:19:45.440 --> 00:19:49.837
upward arc. I mean, any student of history
knows that civilization does not

00:19:49.920 --> 00:19:54.477
always have an upward arc. And in fact,
civilizations have life life cycles. So

00:19:54.560 --> 00:19:56.557
hopefully we are in a strong upward arc.

00:19:56.640 --> 00:19:58.717
I think we are for now. Um but we don't

00:19:58.800 --> 00:19:59.917
want to take that for granted or be

00:20:00.000 --> 00:20:04.157
complacent. Um but the in order to the

00:20:04.240 --> 00:20:08.397
way to think of AI in space is that
in order to achieve any meaningful

00:20:08.480 --> 00:20:10.237
percentage of a cautev 2 scale

00:20:10.320 --> 00:20:13.037
civilization where you're using even a

00:20:13.120 --> 00:20:15.677
millionth a millionth of the sun's

00:20:15.760 --> 00:20:19.197
energy you must have solar powered AI

00:20:19.280 --> 00:20:23.597
satellites in in deep space. Um so so

00:20:23.680 --> 00:20:27.517
that once you realize like once you
think in terms of a kadeshv two scale

00:20:27.600 --> 00:20:29.757
civilization which is what what

00:20:29.840 --> 00:20:32.077
percentage of the sun's energy are you

00:20:32.160 --> 00:20:35.037
turning into useful work um then you

00:20:35.120 --> 00:20:38.237
then it becomes obvious that space is

00:20:38.320 --> 00:20:42.637
overwhelmingly what matters
overwhelmingly. the the sun only

00:20:42.720 --> 00:20:44.877
receives one roughly one two billionth

00:20:44.960 --> 00:20:48.557
of the earth only receives roughly u one

00:20:48.640 --> 00:20:51.757
two billionth of the sun's energy. So if

00:20:51.840 --> 00:20:55.357
you want to have something that is say a

00:20:55.440 --> 00:20:57.037
million times more energy than earth

00:20:57.120 --> 00:20:59.197
could possibly produce,

00:20:59.280 --> 00:21:04.877
>> you must go into space. It's and and so

00:21:04.960 --> 00:21:10.637
um you know this is where it's kind of
handy to have a space company I guess um

00:21:10.720 --> 00:21:13.677
sell the book. >> Easier
to cool chips in space too.

00:21:13.760 --> 00:21:15.997
>> Yes. >> Easier to cool
chips in space. Yeah.

00:21:16.080 --> 00:21:18.317
>> Yes. There's definitely no water
in space. So you're going to have to do

00:21:18.400 --> 00:21:20.637
something. Yeah.
>> Uh that doesn't involve water.

00:21:20.720 --> 00:21:23.117
>> Just hang out. >> Well,
it's you just got to radiate.

00:21:23.200 --> 00:21:26.877
>> That's right. Um so

00:21:26.960 --> 00:21:28.397
my estimate is that actually that that

00:21:28.480 --> 00:21:33.997
that the cost of of electricity like

00:21:34.080 --> 00:21:36.957
like the cost effectiveness of AI in

00:21:37.040 --> 00:21:40.157
space will be overwhelmingly better than

00:21:40.240 --> 00:21:43.037
AI on the ground. So far long before you

00:21:43.120 --> 00:21:45.197
uh exhaust potential energy sources on

00:21:45.280 --> 00:21:47.917
on Earth. Long long before meaning like

00:21:48.000 --> 00:21:50.557
I think even perhaps in the

00:21:50.640 --> 00:21:53.517
four or five year time frame the lowest

00:21:53.600 --> 00:21:57.197
cost way to do AI compute will be with

00:21:57.280 --> 00:22:00.557
solar powered AI satellites.

00:22:00.640 --> 00:22:04.317
So I'd say not more
than five years from now.

00:22:04.400 --> 00:22:07.597
>> Wow. >> And just look
at the supercomputers

00:22:07.680 --> 00:22:11.437
we're building together. Let's say each
one of the racks is two tons. Out of

00:22:11.520 --> 00:22:15.757
that two tons, 1.95 of it
is probably for cooling, >> right?

00:22:15.840 --> 00:22:18.717
>> Oh, yeah.
>> Just imagine how tiny that little

00:22:18.800 --> 00:22:22.477
supercomputer is, right? Each one
of these GB300 racks will just be a little

00:22:22.560 --> 00:22:24.797
tiny thing. >> And and just
electricity generation is

00:22:24.880 --> 00:22:26.717
is already becoming a challenge.

00:22:26.800 --> 00:22:30.157
>> Um so if you if you
start doing any kind

00:22:30.240 --> 00:22:32.077
of scaling for both electricity

00:22:32.160 --> 00:22:35.037
generation and cooling, um you realize,

00:22:35.120 --> 00:22:38.317
okay, space is incredibly compelling. Um

00:22:38.400 --> 00:22:39.917
so

00:22:40.000 --> 00:22:42.637
like let's say you wanted to do um I

00:22:42.720 --> 00:22:49.037
don't know two or 300 gawatt
per year um of of AI compute.

00:22:49.120 --> 00:22:51.437
>> Yeah. >> Um it's very
difficult to do that on

00:22:51.520 --> 00:22:54.637
Earth. Uh the so the the uh US average

00:22:54.720 --> 00:22:56.877
electricity usage uh last time I checked

00:22:56.960 --> 00:23:00.397
was around 460 gawatts per year average

00:23:00.480 --> 00:23:05.437
usage. Um so so something like say uh

00:23:05.520 --> 00:23:07.597
you know three 300 if you're doing 300

00:23:07.680 --> 00:23:12.717
gawatts a year that would be like 2/3
of US electricity production per year.

00:23:12.800 --> 00:23:16.797
>> There's no way you're building power
plants at that level. Um and then if you

00:23:16.880 --> 00:23:20.077
take it up to say a terowatt
per year impossible. Yeah.

00:23:20.160 --> 00:23:22.317
>> Like you have to do that in space.

00:23:22.400 --> 00:23:24.557
>> There there just
is there just is no way

00:23:24.640 --> 00:23:27.917
to do a terowatt uh per year on Earth.

00:23:28.000 --> 00:23:29.437
Um

00:23:29.520 --> 00:23:32.157
and uh and and in space you you've got

00:23:32.240 --> 00:23:37.597
continuous solar. Um you've got uh you
don't you actually don't need batteries

00:23:37.680 --> 00:23:39.037
because it's always sunny in space.

00:23:39.120 --> 00:23:42.877
>> Right. Exactly.
>> Um and um and and the solar panels

00:23:42.960 --> 00:23:44.077
actually become cheaper because you

00:23:44.160 --> 00:23:47.117
don't need glass or framing. Um and the

00:23:47.200 --> 00:23:52.317
cooling is just radiative. So that's
that's why I think >> that's the dream.

00:23:52.400 --> 00:23:55.277
>> Yes, >> that's the dream.
So Jensen, everybody

00:23:55.360 --> 00:23:57.277
last night was asking me and I'm mindful

00:23:57.360 --> 00:24:00.637
it's a earnings uh call for you today.

00:24:00.720 --> 00:24:02.397
So I'm going to say this delicately.

00:24:02.480 --> 00:24:04.877
Everybody has been asking me to ask you

00:24:04.960 --> 00:24:07.487
are we going to have an AI bubble?

00:24:07.570 --> 00:24:12.157
[laughter] That's the last
question. All right, let's

00:24:12.240 --> 00:24:15.997
all right let me well let me just hear
you what we see. Okay, so so I I think

00:24:16.080 --> 00:24:19.357
it's really important when you look at
what's happening around the world and go

00:24:19.440 --> 00:24:22.317
back to first principles of what's
happening in computer science and

00:24:22.400 --> 00:24:26.557
computing. There are three things
that that's happening. The first thing is

00:24:26.640 --> 00:24:30.477
that we all know that Moore's law has
run its course and the ability that the

00:24:30.560 --> 00:24:33.837
amount of demand for computing versus
the amount of computation we can get out

00:24:33.920 --> 00:24:37.437
of general purpose computing is really
challenging and so the world's been

00:24:37.520 --> 00:24:40.557
moving to accelerated computing for some
time. We've been pushing this now for

00:24:40.640 --> 00:24:43.277
some over 20 years. Let me give you one

00:24:43.360 --> 00:24:45.917
statistic. I was just at supercomputing

00:24:46.000 --> 00:24:49.437
six years ago. Uh

00:24:49.520 --> 00:24:52.877
CPUs were 90% of the world's

00:24:52.960 --> 00:24:54.877
supercomputers, top 500 supercomputers

00:24:54.960 --> 00:24:59.517
six years ago. This year less than 15%.

00:24:59.600 --> 00:25:02.637
Went from 90% to 10%. And meanwhile,

00:25:02.720 --> 00:25:07.517
accelerated computing went from the
other way, 10% to now 90%. Okay, so

00:25:07.600 --> 00:25:11.917
you're seeing that inflection point, the
transition in high performance computing

00:25:12.000 --> 00:25:15.117
from general purpose computing
to accelerated computing. Well, one of the

00:25:15.200 --> 00:25:18.717
one of the most data intensive one
of the most intensive computation things

00:25:18.800 --> 00:25:22.717
that the world does in cloud is data
processing. Several hundred billion

00:25:22.800 --> 00:25:25.357
dollars of computation is done on just

00:25:25.440 --> 00:25:30.157
raw data processing. Has nothing to do
with AI, just SQL processing, data

00:25:30.240 --> 00:25:33.917
frames, you know, everybody's names,
address, their their sex, their their

00:25:34.000 --> 00:25:37.757
age, where they live, you know, how much
money they make. All of that sits into a

00:25:37.840 --> 00:25:41.277
data frame. And that data frame drives
the world today. Whether it's in banking

00:25:41.360 --> 00:25:43.517
or you know whether it's in credit cards

00:25:43.600 --> 00:25:48.317
or of course e-commerce
and everything from ad recommendation

00:25:48.400 --> 00:25:51.517
everything is driven off of that data
frame that data frame costs hundreds

00:25:51.600 --> 00:25:55.117
hundreds of dollars to go compute and so
that's the number one thing end of mors

00:25:55.200 --> 00:26:02.040
law the second thing
is generative what the

00:26:02.720 --> 00:26:05.357
the last 15 years is called rexes

00:26:05.440 --> 00:26:10.317
recommener systems how do we know what
information to recommend to us uh in a

00:26:10.400 --> 00:26:14.397
social feed. How do you know what ad
to recommend to somebody? Uh what book to

00:26:14.480 --> 00:26:18.397
recommend, what movie to recommend?
The world is the internet is so gigantic

00:26:18.480 --> 00:26:22.237
without a recommener system that a
little tiny phone of us will have no

00:26:22.320 --> 00:26:26.797
chance of ever seeing the right
information. That Rexus is the engine of

00:26:26.880 --> 00:26:30.717
the internet today. That's going
generative AI. It used to be running on

00:26:30.800 --> 00:26:33.117
CPUs, now runs on GPUs. Which then says

00:26:33.200 --> 00:26:36.237
the third thing when if you just look at

00:26:36.320 --> 00:26:38.477
those two applications

00:26:38.560 --> 00:26:41.037
many of the internet companies can build

00:26:41.120 --> 00:26:43.677
enormous number of GPU supercomputers

00:26:43.760 --> 00:26:48.397
just doing that of course then it
creates this the third opportunity on

00:26:48.480 --> 00:26:52.157
top of it which is agentic AI this
is grock and this is open AI this is

00:26:52.240 --> 00:26:56.957
anthropic you know this is Gemini
agentic AI sits on top of that but don't

00:26:57.040 --> 00:26:58.957
you know don't forget to think about

00:26:59.040 --> 00:27:02.797
what is happening above underneath ath

00:27:02.880 --> 00:27:07.597
what everybody sees as AI today.
There's a whole movement of computing from

00:27:07.680 --> 00:27:11.357
general purpose computing to accelerated
computing and that if you just if you

00:27:11.440 --> 00:27:13.517
take that into consideration you'll come

00:27:13.600 --> 00:27:16.237
to the conclusion that in fact what is

00:27:16.320 --> 00:27:19.197
left over to fuel that revolutionary

00:27:19.280 --> 00:27:22.397
agentic AI is not only substantially

00:27:22.480 --> 00:27:24.717
less than you thought and all of it

00:27:24.800 --> 00:27:27.197
justified. Well, I was just informed by

00:27:27.280 --> 00:27:30.637
the team that my boss and your bosses is

00:27:30.720 --> 00:27:34.397
going to talk next, the honorable
president and his royal highness, the

00:27:34.480 --> 00:27:40.920
crown prince. And hence,
we ran out of time. But in essence,

00:27:41.200 --> 00:27:45.277
this is a such so much love for you Elon

00:27:45.360 --> 00:27:48.477
and and Jensen. But this in essence is a

00:27:48.560 --> 00:27:50.877
92 alliance

00:27:50.960 --> 00:27:53.917
that shifted from energy to digital to

00:27:54.000 --> 00:27:56.797
the intelligence age powered by pioneers

00:27:56.880 --> 00:27:59.277
such as Elon and Jensen to serve

00:27:59.360 --> 00:28:02.317
humanity and create on a net new basis

00:28:02.400 --> 00:28:04.877
new economies, new jobs and a better

00:28:04.960 --> 00:28:09.517
future for humanity powered by the
Kingdom of Saudi Arabia and the United

00:28:09.600 --> 00:28:13.277
States. Thank you for our lifetime
partnership and friendship. Thank you

00:28:13.360 --> 00:28:16.387
Elon. Thank you Jensen. >> Thank you.

00:28:16.470 --> 00:28:21.110
>> [applause] >> All right. [music]

00:28:27.600 --> 00:28:31.757
[music] Ladies and gentlemen, what we
just heard was a conversation about the

00:28:31.840 --> 00:28:34.397
future of technology spoken by the

00:28:34.480 --> 00:28:36.637
people building it. To his excellency

00:28:36.720 --> 00:28:41.437
Abdullah Hailon Mosque and Jensen Huang,
thank you for a conversation that did

00:28:41.520 --> 00:28:43.437
more than describe the future of

00:28:43.520 --> 00:28:46.557
technology. It actually showed us who's

00:28:46.640 --> 00:28:48.877
building it and how. And now ladies and

00:28:48.960 --> 00:28:51.037
gentlemen, we ask you to remain in the

00:28:51.120 --> 00:28:54.477
plenary. We will soon proceed with the

00:28:54.560 --> 00:28:56.717
special address by the leaders. As his

00:28:56.800 --> 00:28:59.677
excellency Swaha mentioned, our leaders

00:28:59.760 --> 00:29:03.357
are about to arrive. So the sponsorship

00:29:03.440 --> 00:29:06.317
recognition segment will take place

00:29:06.400 --> 00:29:10.397
after their departure. And I just want
to mention that uh his royal highness

00:29:10.480 --> 00:29:12.557
the crown prince Muhammad bin Salman Aiz

00:29:12.640 --> 00:29:15.517
aloud will deliver his speech in Arabic.

00:29:15.600 --> 00:29:18.237
For English speaking guests, we have

00:29:18.320 --> 00:29:20.557
live translation devices available for

00:29:20.640 --> 00:29:27.240
you. Our ground team will be more
than happy to assist you. Thank you.

00:29:32.720 --> 00:29:34.317
>> Waiting [music]

00:29:34.400 --> 00:29:37.037
for

00:29:37.120 --> 00:29:41.757
sing [music and singing] his song.

00:29:41.840 --> 00:29:43.997
So many

00:29:44.080 --> 00:29:45.837
dreams [music]

00:29:45.920 --> 00:29:50.637
I kept deep inside [singing] me.

00:29:50.720 --> 00:29:54.797
alone [music] in the dark.

00:29:54.880 --> 00:29:57.357
But now you've come [singing]

00:29:57.440 --> 00:30:02.797
along and you

00:30:02.880 --> 00:30:06.480
light [singing] up