- 7 weeks ago
Elon إيلون Musk ماسك Alswaha السواحه Jensen جينسين Huang هوانغ AI ذكاء ✨
Watch شاهد this هذا exclusive حصري panel جلسة featuring تضم Elon إيلون Musk ماسك, Jensen جينسين Huang هوانغ, and و Abdullah عبدالله Alswaha السواحه 🎬🚀. They هم discuss يناقشون future مستقبل AI ذكاء and و technology تقنية trends اتجاهات at في Saudi سعودي Investment استثمار Forum منتدى ⏳. Don't لا forget تنسَ to لـ follow اتبع the الـ channel قناة for لـ more مزيد AI ذكاء updates تحديثات ✨🌟.
Tags الكلمات المفتاحية
Elon إيلون Musk ماسك Jensen جينسين Huang هوانغ Alswaha السواحه AI ذكاء Tech تقنية Future مستقبل Saudi السعودية Investment استثمار Forum منتدى Vision رؤية 2030 Global عالمي Innovation ابتكار ✨
Watch شاهد this هذا exclusive حصري panel جلسة featuring تضم Elon إيلون Musk ماسك, Jensen جينسين Huang هوانغ, and و Abdullah عبدالله Alswaha السواحه 🎬🚀. They هم discuss يناقشون future مستقبل AI ذكاء and و technology تقنية trends اتجاهات at في Saudi سعودي Investment استثمار Forum منتدى ⏳. Don't لا forget تنسَ to لـ follow اتبع the الـ channel قناة for لـ more مزيد AI ذكاء updates تحديثات ✨🌟.
Tags الكلمات المفتاحية
Elon إيلون Musk ماسك Jensen جينسين Huang هوانغ Alswaha السواحه AI ذكاء Tech تقنية Future مستقبل Saudi السعودية Investment استثمار Forum منتدى Vision رؤية 2030 Global عالمي Innovation ابتكار ✨
Category
🤖
TechTranscript
00:00Please have a seat.
00:02Now I'm sure we can do a bigger round of applause
00:04for one of the greatest two leaders of our history.
00:08Let's go ahead.
00:10So we're talking about, I lost count.
00:14You know, $7 to $8 trillion worth of market cap, comp, I lost count.
00:20But right now, we're here to celebrate a historic moment.
00:25A moment that yesterday, during the dinner,
00:29and thank you for joining us under the patronage of the Honorable President
00:35and His Royal Highness the Crown Prince, Moussaidi,
00:38where we had the pleasure to hear it firsthand.
00:41This is the greatest alliance between the Kingdom of Saudi Arabia
00:46and the United States, where we have joined hands
00:50and you have helped us build our energy-based economy,
00:54fueling and energizing the industrial age,
00:57and now fast forward, going to the intelligence age,
01:00where we can fuel AI, factories, robotics, EVs, and all of the rest.
01:08Speaking of that, let's start with you, Ilan, if you don't mind, Jensen.
01:12Feel free to chime in.
01:13You have a big fascination of something all of us have admired,
01:18first-order thinking, which Jensen sometimes calls first-order scaling,
01:23which is an opportunity for how you have dropped the cost of batteries
01:28from 1,000 for kilowatt hour to sub-hundred bucks.
01:33And right now, you're doing the same thing with robotics
01:36for actuators with servo rotors and motors.
01:40So I want to hear from you.
01:42How do you manage to always disrupt every single industry with that thinking?
01:49Well, it's mostly not disruption.
01:51It's creation.
01:54So with, say, SpaceX, with reusable rockets,
01:59there really weren't any reusable rockets.
02:03But the essence of getting, of revolutionizing space travel is reusability.
02:09If you throw the rocket away every time,
02:11the cost of access to space is extremely high.
02:16With respect to electric cars, there weren't any electric cars
02:20when we started making them, really.
02:22You couldn't buy any, to the best of my knowledge.
02:25So with Tesla, we wanted to make electric cars compelling and affordable.
02:30That was the goal.
02:33With respect to humanoid robotics,
02:36there are no useful humanoid robots at this point.
02:40There are sort of gimmicks,
02:41but there are no actually useful humanoid robots.
02:45And I think Tesla's going to make the first actually useful humanoid robots.
02:49And this will be quite a revolution.
02:51And I think something that everyone will want.
02:55Because I always think of, like,
02:57who wouldn't want their own personal C-3PO R2-D2?
03:01Oh, yeah?
03:02Of course.
03:03Everyone will want one, right?
03:05And then there would be many in industry providing products and services.
03:10This is why I say that humanoid robots will be the biggest industry,
03:14or the biggest product ever.
03:16Bigger than cell phones or anything else,
03:18because everyone's going to want one,
03:21or maybe more than one,
03:23and there will be many in industry.
03:26I just want R2-D2 and C-3PO's body.
03:30Yeah.
03:31There you go.
03:32Well, I mean, a humanoid robot would be better than R2-D2 and C-3PO combined times 10.
03:37Yeah.
03:38So, the...
03:40And, you know, people often talk about sort of eliminating poverty and that kind of thing.
03:46But, really, the...
03:48How long have they been talking about that?
03:50There's lots of talk.
03:52You know, there's lots of NGOs sort of trying to do these things,
03:55but really not succeeding.
03:58And, you know, the evidence speaks for itself.
04:03But AI and humanoid robots will actually eliminate poverty.
04:08And Tesla won't be the only one that makes them.
04:09I think Tesla will pioneer this,
04:11but there will be many other companies that make humanoid robots.
04:13But there is only basically one way to make everyone wealthy,
04:18and that is AI and robotics.
04:20And we can't talk about robotics without AI factories.
04:24And yesterday was such a historic day for the two nations,
04:28but also for all of us,
04:30where we celebrate the AI strategic partnership with the U.S.,
04:35signed witness by the Honorable President and His Royal Highness,
04:40about how we are committing our capital energy land
04:45to energize the AI U.S. ecosystem
04:49to be able to build inference nodes, training nodes,
04:52and to be the most AI-enabled nation.
04:55With that announcement,
04:56tell me what's next in AI factories, Jensen.
04:59There's a beautiful story about how Saudi Arabia is building AI refineries.
05:05Now building AI factories, or oil refiners to AI factories.
05:08I love that.
05:10You know, I've said that AI is an infrastructure,
05:13and the reason for that, of course,
05:14we understand AI from the perspective of the technology
05:17and how it's revolutionizing every industry.
05:20Digital intelligence, of course,
05:22has applications into every field.
05:25And so it's going to be used by every company,
05:28every industry, every country.
05:29In that way, it's foundational,
05:33and therefore it's part of infrastructure.
05:35What is new about AI from a computer science perspective
05:40is that the way computing was done in the past
05:44was largely retrieval-based computing.
05:46Somebody typed in a story,
05:48or somebody created a piece of art,
05:51or came up with four versions of a digital ad,
05:55or it's all pre-built by somebody,
05:58which is then using a system
06:00to retrieve the appropriate version for you.
06:03It's a retrieval-based computing model.
06:06Hadoop and many of the frameworks
06:09and operating systems of the past
06:11are all designed to retrieve the appropriate information for you.
06:14But today, software is going to be generated in real time.
06:19It's generative.
06:20Based on the context, based on the circumstance,
06:22based on who you are, based on the problem you asked,
06:25based on your prompt,
06:26it will generate unique content for you.
06:29Every single time, for everybody, it's unique.
06:31When you use Grok, every time you use it, it's different.
06:34We all love Grok.
06:35Right?
06:36Based on the prompt that you give it
06:38and based on the circumstance.
06:40And so therefore,
06:42it used to be retrieval-based.
06:44Today, it's generative.
06:46And if it's generative,
06:48and every time it's different,
06:50then you need AI factories all over the world
06:52to generate the content in real time,
06:55which is the reason why you need AI factories.
06:56And this is a unique way of doing computation,
06:59but the benefit, of course,
07:00is that everything isn't preconceived and pre-documented,
07:04and it's contextually sensible and therefore intelligent.
07:09So AI factories and robotics,
07:11and we heard it yesterday from His Royal Highness,
07:14his vision,
07:15how to augment our workforce
07:17with roughly tens of millions of robotics
07:20to be able to infuse
07:22the next wave of productivity and progress.
07:24But this scares a lot of folks
07:27when it comes to the future of jobs.
07:29So let's hear about your thoughts,
07:31Elon and Jensen, on that.
07:34Sure, well,
07:38say like in the long term,
07:39where will things end up?
07:40Long term, I don't know what long term is.
07:42Maybe it's 10, 20 years,
07:45something like that.
07:46For me, that's long term.
07:48My prediction is that work will be optional.
07:51Optional.
07:52Optional.
07:56We'll take that.
07:57Yeah.
07:58I mean, it'll be like playing sports
08:01or a video game or something like that.
08:03If you want to work,
08:05you know, in the same way,
08:06like you can go to the store
08:08and just buy some vegetables,
08:11or you could grow vegetables in your backyard.
08:14It's much harder to grow vegetables in your backyard,
08:16but some people still do it
08:17because they like growing vegetables.
08:20That will be what work is like, optional.
08:24And between now and then,
08:25there's actually a lot of work to get to that point.
08:28And I always recommend people
08:30read Yen Banks' culture books
08:33to get a sense for
08:34what a probable positive AI future is like.
08:40And interestingly, in those books,
08:43money is no longer, doesn't exist.
08:46It's kind of interesting.
08:47And my guess is,
08:50if you go out long enough,
08:51assuming there's a continued improvement
08:53in AI and robotics,
08:55which this seems likely,
08:56the money will stop being relevant
09:00at some point in the future.
09:02There will still be constraints on power,
09:06like electricity and mass.
09:09The fundamental physics elements
09:11will still be constraints.
09:14But I think at some point,
09:19currency becomes irrelevant.
09:23Jensen, any thoughts?
09:30By the way, the NVIDIA earnings call is later today.
09:34And by the way, since currency is relevant.
09:38Cheers.
09:43Elon just wants to share with you
09:45some breaking news.
09:51The two of us would like to share
09:53some breaking news.
09:56Let's see.
09:56I would say there's
09:58different horizons you could look at.
10:01Everybody's jobs will be different.
10:03I think that's for sure.
10:04How the students learn will be different.
10:08How people do their work
10:09will be different, obviously,
10:10because a lot of the things
10:12that we do mundanely
10:14or arduously
10:15or very difficultly
10:16are going to be done very simply.
10:19And so we're going to be more productive
10:21from that sense.
10:22One of the things that I will say
10:24is that for most people
10:27or company,
10:28if your life becomes more productive
10:31and if the things that you're doing
10:34with great difficulty
10:35become simpler,
10:36it is very likely
10:37because you have so many ideas,
10:40you'll have more time
10:40to go pursue things.
10:41It is my guess
10:42that Elon will be busier
10:43as a result of AI.
10:45I'm going to be busier
10:46as a result of AI.
10:47And the reason for that
10:48is because we have so many ideas
10:49we want to pursue,
10:50so many things
10:51that we still have
10:52in our backlog
10:53inside our company
10:54that we can go pursue.
10:55If we're more productive,
10:57we can get to those things faster.
10:59And so in the near term,
11:01I would say
11:01that there's every evidence
11:03that we will be more productive
11:05and yet still be busier
11:07because we have so many ideas.
11:08One thing that I will say,
11:10give you some evidence,
11:12is that,
11:13and I was just telling Elon
11:14about this earlier,
11:15radiology, for example,
11:16has largely been converted
11:19to AI-driven radiology
11:21and there's some really great companies
11:22doing that.
11:23And the surprising thing
11:25is the prediction
11:27that all radiologists
11:29would be the first jobs to go
11:30was exactly the opposite.
11:33The trend shows
11:34that there are more radiologists
11:36being hired now
11:37as a result of AI.
11:38And the reason for that,
11:39if you take a step back,
11:41it's because the goal
11:42of a radiologist
11:43is not to study the images.
11:45The goal of a radiologist
11:47is to diagnose a disease.
11:49Now the studying of the images
11:51became so productive,
11:52they could study more images,
11:54study more modalities,
11:56spend more time
11:56with the patients
11:57and as a result,
11:59they were actually
11:59accepting more patients.
12:01We're doing more radiology
12:02all around the world.
12:03We're doing a better job
12:04with diagnosing disease.
12:05And so that's kind of
12:06the near-term outcome
12:10of AI and productivity.
12:13And we'll see what happens
12:15long-term.
12:16You know,
12:18when currency doesn't matter anymore,
12:20just let me know right before.
12:24You'll see it coming.
12:27You'll see it coming.
12:29We text often,
12:30so just...
12:31Yeah, we do.
12:31Yeah, just text it out.
12:32Let me know before.
12:33I kind of agree
12:34with both of you
12:35because if you look
12:36at every technological trend,
12:39every general purpose technology
12:40has been net new positive
12:43for the globe,
12:44for humanity,
12:45and so forth.
12:46And let me share
12:46with you two case studies.
12:48Your Excellency,
12:49I think it's precisely
12:50the reason for that
12:51is because all the great ideas
12:53from innovators like Ilan,
12:55you have so many good ideas.
12:58And Jensen as well.
12:59Yeah, well,
13:00thank you.
13:01So let me share with you
13:02two stories
13:03from two Saudi innovators
13:05in collaboration
13:05with a lot of the great work
13:07that NVIDIA does,
13:09that Grok does.
13:10So one of them
13:11is Professor Omar Yagi,
13:13who's the first American Saudi
13:15to win a Nobel Prize
13:16in creating new chemistry.
13:18And the way he has done that,
13:20he has leveraged
13:21new AI accelerators
13:22and models like Grok
13:24to be able
13:25to create new chemistry
13:26when it comes
13:27to metal organic frameworks.
13:29Those are metal ions
13:30that are positively charged
13:32with organic linkers
13:34to be able
13:34to effectively create
13:35a sponge
13:36with 0.33 nanometers pores
13:40to capture water from air
13:42and also to capture carbon dioxide.
13:45The second story
13:47has also to do
13:48with AI
13:48accelerated by NVIDIA
13:50and with models like Grok,
13:52which is Nanopalm,
13:54which is effectively
13:55creating a nanorobot
13:56500 nanometers
13:58by 1,000 nanometers
13:59to be able
14:00to do gene editing
14:01leveraging the CRISPR technology
14:04to take out
14:05sickle cell disease.
14:06Now,
14:07in both these instances,
14:08they originated
14:0920 years ago
14:10in research,
14:11but AI was able
14:13to really accelerate
14:15the outcomes
14:16and the outputs
14:17such that we can move
14:18into new value pools.
14:20So I think
14:20with every technological trend,
14:23humanity is going
14:23to always manage
14:24to shift
14:25to new value pools
14:26when it comes
14:27to workforce
14:28and productivity.
14:29But we have
14:30some great announcements
14:30to talk about here today.
14:32Let's begin with you,
14:33Ilan,
14:34the things that we're doing
14:35with XAI.
14:36Yeah,
14:37we're excited
14:37to announce
14:37that we're doing
14:39a 500 megawatt,
14:43I mean,
14:43yeah,
14:44500,
14:44sorry.
14:46500 megawatt.
14:47500 megawatt,
14:48yeah.
14:50We're doing 500 megawatts.
14:53Sorry,
14:55yeah,
14:55500 gigawatt
14:56won't have to wait.
14:58So,
15:00that'll be
15:01eight bazillion
15:01trillion dollars.
15:05Stop that.
15:11So,
15:12yeah,
15:12we're doing
15:15XAI
15:16and
15:16the Kingdom of Saudi Arabia
15:17are doing
15:18a...
15:19Humane,
15:19500 megawatts,
15:20starting with 50 megawatts
15:21phase one,
15:22and we're doing it
15:23with NVIDIA.
15:24Congratulations
15:25to the Humane team,
15:26to Target team,
15:27such a fantastic
15:28job.
15:30Jensen,
15:30I think we're also doing
15:31some great announcements
15:32this week.
15:34We are?
15:34Yeah?
15:36We're announcing
15:37all kinds of things.
15:38Our partnership
15:39with Humane
15:40is going incredibly well.
15:41First of all,
15:42we worked together
15:44to get this company
15:45started and off the ground
15:47and just got
15:49an incredible customer
15:50with Elon.
15:51Could you imagine
15:52a startup company,
15:54approximately $0 billion
15:56in revenues,
15:57now going to build
15:59a data center
15:59for Elon?
16:00500 megawatts
16:01is gigantic.
16:03This company
16:03is off the charts
16:04right away.
16:05In addition to that,
16:07we're working...
16:10AWS,
16:10as you know,
16:11is also coming...
16:12to the Humane team
16:14with AWS,
16:16starting with 100 megawatts
16:17with a gigawatt ambition
16:19and counting.
16:21So AWS
16:22is also coming
16:23to Humane.
16:23We're working
16:24with Humane
16:25on Omniverse
16:26digital twins.
16:28As you know,
16:29that AI
16:30is not just...
16:31Well,
16:32just agentic AI
16:33and chatbots
16:34and cognitive AI
16:36is incredibly important
16:37to the world,
16:37but AI applies
16:39to everything,
16:40chemicals
16:40and proteins
16:41and genes
16:41and physics
16:42and fluid dynamics
16:44and particles
16:45and, of course,
16:45robotics and activation.
16:47And we created
16:48this world
16:49called Omniverse
16:49where robots
16:51can learn
16:51how to be good robots.
16:52And it's physically based.
16:55It obeys
16:55the laws of physics.
16:56And so robots
16:57can learn
16:58in these environments.
16:59And we're working
16:59with Humane
17:00to apply Omniverse
17:02to all kinds
17:03of digital factories
17:05and robotics
17:05and warehouses
17:06and things like that.
17:07And so that's another.
17:09We're also
17:10working in Saudi Arabia
17:11to build supercomputers
17:13to simulate
17:13quantum computers
17:15and using our computers
17:18to be the controller
17:19and the error correction,
17:21quantum error correction,
17:22requires an enormous amount
17:24of computation.
17:25And so we're doing
17:26a lot of great work there, too.
17:27So a big partnership
17:28with Humane.
17:29They're off the charts,
17:30off the ground
17:31and off the charts
17:32at the same time.
17:32This is how
17:33we walk the talk
17:34in the kingdom
17:35of Saudi Arabia
17:36in partnership
17:37with the U.S.
17:38Yesterday,
17:39the president
17:40and his royal highness
17:41announced the AI
17:42strategic framework
17:43and partnership.
17:44Today,
17:45we're going big
17:46with Elon and Jensen.
17:48So thank you
17:49for those opportunities.
17:54Now,
17:55they told me
17:56I have time
17:56for two last questions.
17:58So last night
17:59at the dinner,
18:00I got a number
18:02of questions
18:02because it seems
18:03that the schedule leaked
18:04and everybody
18:06was giving me hints
18:07about the last
18:08two questions
18:08I'm going to do.
18:09So the first one
18:10was for you, Elon.
18:12And there's a big one
18:14for you, Jensen.
18:15So prepare for that one.
18:18AI in space,
18:19is that possible?
18:22Yes.
18:23If civilization continues,
18:25which it probably will,
18:26then AI in space
18:27is inevitable.
18:33You know,
18:34I always have to
18:34like preface that,
18:35you know.
18:37We shouldn't take
18:38civilization for granted.
18:39We need to make sure
18:40to take care
18:41to ensure
18:42that civilization
18:43has an upward arc.
18:45I mean,
18:45any student of history
18:46knows that civilization
18:47does not always
18:48have an upward arc
18:49and in fact,
18:50civilizations have
18:51life cycles.
18:52So hopefully,
18:53we are in a strong
18:54upward arc.
18:55I think we are for now.
18:56But we don't want
18:57to take that for granted
18:58or be complacent.
19:00But in order to,
19:02the way to think
19:03of AI in space
19:04is that in order
19:05to achieve
19:06any meaningful percentage
19:07of a Kardashev
19:08two-scale civilization
19:09where you're using
19:10even a millionth
19:13of the sun's energy,
19:15you must have
19:16solar-powered AI satellites
19:18in deep space.
19:22So once you realize,
19:23once you think
19:24in terms of a
19:25Kardashev two-scale
19:26civilization
19:26which is
19:27what percentage
19:28of the sun's energy
19:29are you turning
19:31into useful work,
19:32then it becomes obvious
19:34that space
19:36is overwhelmingly
19:37what matters.
19:39Overwhelmingly.
19:40The sun only receives
19:41roughly one,
19:42two billionth of,
19:44the Earth only receives
19:45roughly one,
19:47two billionth
19:47of the sun's energy.
19:49So if you want
19:50to have something
19:51that is,
19:53say,
19:53a million times
19:54more energy
19:55than Earth
19:55could possibly produce,
19:57you must go
19:58into space.
20:01And so,
20:04this is where
20:05it's kind of handy
20:05to have a space company,
20:06I guess.
20:09Sell the work,
20:10as they say.
20:10cool chips in space too.
20:12Yes.
20:12Easier to cool chips
20:13in space.
20:14Yes,
20:15there's definitely
20:15no water in space
20:16so you're going
20:16to have to do something
20:17that doesn't involve water.
20:19Just hang out.
20:20Well,
20:20you've just got to radiate.
20:21That's right.
20:25my estimate is that actually
20:26that the cost of electricity,
20:32like the cost effectiveness
20:34of AI in space
20:35will be overwhelmingly
20:38better than AI on the ground.
20:39So long before you exhaust
20:42potential energy sources
20:43on Earth,
20:44long before,
20:46meaning like I think
20:46even perhaps in the
20:48four or five year time frame,
20:51the lowest cost way
20:52to do AI compute
20:54will be
20:55with solar powered
20:56AI satellites.
20:59So,
21:00I'd say
21:01not more than
21:01five years from now.
21:03Wow.
21:03And just look at
21:05the supercomputers
21:06we're building together.
21:07Let's say each one
21:08of the racks
21:08is two tons.
21:09Out of that two tons,
21:111.95 of it
21:12is probably for cooling.
21:14Right.
21:14Oh yeah.
21:15Just imagine
21:16how tiny
21:17that little supercomputer is,
21:18right?
21:18Each one of these
21:19GB300 racks.
21:20It'll just be
21:20a little tiny thing.
21:21And just electricity
21:22generation is already
21:24becoming a challenge.
21:26So if you start
21:27doing any kind of scaling
21:29for both electricity
21:30generation and cooling,
21:32you realize,
21:33okay, space
21:34is incredibly compelling.
21:36So,
21:38like let's say
21:39you wanted to do
21:39I don't know,
21:41two or three hundred
21:42gigawatts per year
21:45of AI compute.
21:48It's very difficult
21:49to do that
21:49on earth.
21:51So the U.S.
21:52average electricity
21:53usage,
21:54last time I checked,
21:55was around
21:56460 gigawatts
21:57per year
21:58average usage.
22:00So something like,
22:02say,
22:04you know,
22:04if you're doing
22:05300 gigawatts
22:06a year,
22:07that would be like
22:07two-thirds of U.S.
22:09electricity production
22:09per year.
22:11There's no way
22:12you're building
22:12power plants
22:13at that level.
22:14And if you take it up
22:15to, say,
22:16a terawatt per year,
22:17impossible.
22:18Like you have
22:19to do that in space.
22:20There just is no way
22:23to do a terawatt
22:24per year on earth.
22:28And in space,
22:30you've got
22:30continuous solar.
22:32You've got,
22:34you actually don't
22:35need batteries
22:36because it's always
22:36sunny in space.
22:37Right, exactly.
22:40And the solar panels
22:41actually become cheaper
22:42because you don't need
22:43glass or framing.
22:45And the cooling
22:46is just radiative.
22:47So that's why I think
22:49That's the dream.
22:50Yes.
22:51That's the dream.
22:52So, Jensen,
22:53everybody last night
22:54was asking me,
22:55and I'm mindful
22:55it's earnings call
22:58for you today.
22:59So I'm going to say
22:59this delicately.
23:01Everybody has been
23:01asking me to ask you,
23:03are we going to have
23:04an AI bubble?
23:06That's the last question.
23:07All right.
23:08Let's, no.
23:10All right.
23:11Let me just tell you
23:12what we see.
23:13Okay?
23:13So I think it's really
23:15important when you look
23:16at what's happening
23:17around the world
23:17and go back to first
23:18principles of what's
23:19happening in computer
23:20science and computing.
23:21There are three things
23:22that's happening.
23:23The first thing
23:24is that we all know
23:26that Moore's Law
23:26has run its course
23:27and the ability,
23:28the amount of demand
23:29for computing
23:30versus the amount
23:31of computation
23:31we can get out
23:32of general purpose
23:33computing is really
23:34challenging.
23:34And so the world
23:35has been moving
23:36to accelerated computing
23:37for some time.
23:38We've been pushing
23:38this now for some
23:39over 20 years.
23:40Let me give you
23:41one statistic.
23:42I was just at
23:43supercomputing.
23:44Six years ago,
23:48CPUs were 90%
23:50of the world's
23:51supercomputers,
23:52top 500 supercomputers.
23:53Six years ago.
23:54This year,
23:56less than 15%.
23:57Went from 90%
23:59to 10%.
23:59And meanwhile,
24:01accelerated computing
24:02went from the other
24:02way, 10%
24:03to now 90%.
24:05Okay, so you're
24:06seeing that inflection
24:07point, the transition
24:08in high-performance
24:09computing from
24:10general purpose
24:11computing to
24:11accelerated computing.
24:12Well, one of the
24:14most data intensive,
24:15one of the most
24:16intensive computation
24:17things that the world
24:17does in cloud
24:18is data processing.
24:19Several hundred
24:20billion dollars
24:21of computation
24:22is done
24:23on just raw data
24:24processing.
24:25That Rexis
24:26is the engine
24:27of the internet today.
24:28that's going
24:29generative AI.
24:30It used to be
24:31running on CPUs,
24:32now it runs on GPUs.
24:33Which then says
24:33the third thing,
24:35if you just look
24:36at those two
24:37applications,
24:39many of the
24:39internet companies
24:40can build
24:41enormous number
24:42of GPU supercomputers
24:44just doing that.
24:45Of course,
24:46then it creates
24:47the third opportunity
24:48on top of it,
24:49which is agentic AI.
24:50This is Grok,
24:51and this is OpenAI,
24:52this is Anthropic,
24:53this is Gemini.
24:55Agentic AI sits
24:56on top of that.
24:56But don't forget
24:58to think about
24:59what is happening
25:01above,
25:02underneath,
25:03what everybody
25:04sees as AI today.
25:05There's a whole
25:06movement of computing
25:07from general purpose
25:08computing
25:09to accelerated computing.
25:10And that,
25:10if you take that
25:12into consideration,
25:13you'll come
25:14to the conclusion
25:14that in fact,
25:16what is left over
25:17to fuel
25:18that revolutionary
25:19agentic AI
25:20is not only
25:21substantially less
25:23than you thought,
25:24and all of it
25:25justified.
25:26Well,
25:26I was just
25:27informed by the team
25:28that my boss
25:30and your bosses
25:31is going to talk next,
25:32the Honorable President
25:33and His Royal Highness
25:34the Crown Prince,
25:36and hence we ran
25:37out of time.
25:38But in essence,
25:41this is our,
25:43such,
25:43so much love
25:44for you,
25:45Ilan,
25:46and Jensen.
25:47But this,
25:48in essence,
25:48is a 92 alliance
25:51that shifted
25:52from energy
25:53to digital
25:54to the intelligence age
25:55powered by pioneers
25:57such as Ilan
25:58and Jensen
25:59to serve humanity
26:00and create
26:01on a net new basis
26:02new economies,
26:04new jobs,
26:05and a better future
26:06for humanity
26:06powered by
26:08the Kingdom
26:08of Saudi Arabia
26:09and the United States.
26:10Thank you for
26:11our lifetime
26:12partnership
26:12and friendship.
26:13Thank you,
26:14Ilan.
26:14Thank you,
26:15Jensen.
26:18All right.
Comments