1 00:00:00,360 --> 00:00:02,839 Speaker 1: Please have us see now, I'm sure we can do 2 00:00:02,919 --> 00:00:06,000 a bigger run of applause for one of the greatest 3 00:00:06,080 --> 00:00:11,039 two leaders of our history. Let's go ahead. So we're 4 00:00:11,080 --> 00:00:15,679 talking about I lost count. You know, seven to eight 5 00:00:15,720 --> 00:00:19,879 trillion dollars worth of market cap comp I lost count. 6 00:00:20,640 --> 00:00:24,800 But right now we're here to celebrate a historic moment, 7 00:00:25,480 --> 00:00:29,760 a moment that yesterday during the dinner, and thank you 8 00:00:29,800 --> 00:00:34,920 for joining us under the patronage of the Honorable President 9 00:00:34,960 --> 00:00:38,039 and his Royal Highness, the com Prince Musseidi, where we 10 00:00:38,159 --> 00:00:41,880 have the pleasure to hear it firsthand. This is the 11 00:00:42,000 --> 00:00:46,719 greatest alliance between the Kingdom of Sada Arabia and the 12 00:00:46,799 --> 00:00:50,359 United States, where we have joined hands and you have 13 00:00:50,439 --> 00:00:55,520 helped us build our energy based economy, fueling and energizing 14 00:00:55,600 --> 00:00:59,000 the industrial age. And now fast forward going to the 15 00:00:59,039 --> 00:01:05,840 intelligence age where we can fuel AI factories, robotics, eaves 16 00:01:06,359 --> 00:01:09,280 and all of the rest. Speaking of that, let's start 17 00:01:09,280 --> 00:01:11,760 with you Elon. If you don't mind, Jensen, feel free 18 00:01:11,760 --> 00:01:15,760 to chive in. You have a big fascination of something 19 00:01:15,959 --> 00:01:19,680 all of us have admired. First order thinking with gens 20 00:01:19,719 --> 00:01:24,079 is sometimes called first order scaling, which is an opportunity 21 00:01:24,239 --> 00:01:28,120 for how you have dropped the costs of batteries from 22 00:01:28,280 --> 00:01:32,439 one thousand for kilo at hour through sub hundred bucks 23 00:01:32,840 --> 00:01:35,560 And right now you're doing the same thing with robotics, 24 00:01:35,840 --> 00:01:40,680 for actuators, with servo rotors and motors. So I want 25 00:01:40,719 --> 00:01:44,480 to hear from you, how do you manage to always 26 00:01:44,560 --> 00:01:47,120 disrupt every single industry with that. 27 00:01:47,120 --> 00:01:53,480 Speaker 2: Thinking, Well, it's mostly not a disruption, it's creation. So 28 00:01:54,159 --> 00:02:00,200 with say SpaceX with reusable rockets, they really want any 29 00:02:00,280 --> 00:02:07,760 reusable rockets, but the essence of getting of revolutionizing space 30 00:02:07,760 --> 00:02:10,800 travel is reusability. If you throw the rocket away every time, 31 00:02:11,240 --> 00:02:15,919 the cost of access to space is extremely high. With 32 00:02:15,960 --> 00:02:20,280 respect to electric cars, there weren't any electric cars when 33 00:02:20,319 --> 00:02:23,199 we started making them. Really you couldn't buy any, to 34 00:02:23,240 --> 00:02:26,960 the best of my knowledge. So with Tesla, we wanted 35 00:02:26,960 --> 00:02:30,800 to make electric cars compelling at affordable. That was the goal. 36 00:02:32,560 --> 00:02:35,719 The You know, with respect to humanoid robotics, there are 37 00:02:35,759 --> 00:02:40,159 no useful humanoid robotics robots at this point. There are 38 00:02:40,199 --> 00:02:41,960 sort of gimmicks, but there are not There are no 39 00:02:42,080 --> 00:02:46,199 actually useful humanoid robots. And I think Tesla is going 40 00:02:46,240 --> 00:02:49,680 to make the first actually useful humanoid robots, and this 41 00:02:49,759 --> 00:02:52,319 will be quite a revolution. And I think something that 42 00:02:52,439 --> 00:02:56,199 will that everyone will want, because I was thinking, like, 43 00:02:56,520 --> 00:02:59,919 who wouldn't want their own personal C three po are 44 00:03:00,039 --> 00:03:04,560 two D two. Of course everyone we want one, right, 45 00:03:05,400 --> 00:03:08,639 and and then there would be many in industry providing 46 00:03:08,639 --> 00:03:11,639 products and services. This is why I say that humanoid 47 00:03:11,719 --> 00:03:14,520 robots will be the biggest industry or the biggest product ever, 48 00:03:16,240 --> 00:03:19,599 bigger than cell phones or anything else, because everyone's going 49 00:03:19,639 --> 00:03:22,560 to want one and or maybe more more than one, 50 00:03:22,639 --> 00:03:27,199 and there'll be many in industry. On R two D 51 00:03:27,319 --> 00:03:32,759 two in C three pos body, yeah, very well. I 52 00:03:32,759 --> 00:03:34,400 mean a HumanID robot would be better than R two 53 00:03:34,439 --> 00:03:37,759 D two and C through vo combined times ten. So 54 00:03:38,719 --> 00:03:43,479 the and and you know, people often talk about sort 55 00:03:43,520 --> 00:03:46,199 of eliminating poverty and that kind of thing, but really 56 00:03:46,280 --> 00:03:50,599 there's how long they've been talking about that. There's lots 57 00:03:50,639 --> 00:03:54,120 of talk. Uh, there's lots of NGOs sort of trying 58 00:03:54,159 --> 00:03:57,560 to do these things but but really not succeeding. And 59 00:03:58,000 --> 00:04:03,479 and you know the evidence be for ourself, but AI 60 00:04:03,520 --> 00:04:07,840 and humoroid robots will actually eliminate poverty and tell them 61 00:04:07,840 --> 00:04:09,120 won't be the only one that makes them. I think 62 00:04:09,120 --> 00:04:11,400 tells them will pioneer this, but there will be many 63 00:04:11,439 --> 00:04:14,319 other companies that make humanoid robots. But there is only 64 00:04:14,400 --> 00:04:18,199 basically one way to make everyone wealthy, and that is 65 00:04:18,279 --> 00:04:19,399 AI in robotics. 66 00:04:20,399 --> 00:04:24,439 Speaker 1: And we can't talk about robotics without AI factories. And 67 00:04:24,600 --> 00:04:28,120 yesterday was such a historic day for the two nations 68 00:04:28,319 --> 00:04:32,079 but also for all of US where we celebrate the 69 00:04:32,160 --> 00:04:37,079 AI Strategic Partnership with the US signed witnessed by the 70 00:04:37,120 --> 00:04:40,839 Honorable President and His Royal Highness, about how we are 71 00:04:41,079 --> 00:04:48,560 committing our capital, energy land to energize the AI US ecosystem, 72 00:04:48,800 --> 00:04:51,879 to be able to build inference node training nodes, and 73 00:04:51,920 --> 00:04:55,720 to be the most AI enabled nation. With that announcement, 74 00:04:55,839 --> 00:04:59,839 Tell me what's next in AI factories, Jensen, there's. 75 00:04:59,480 --> 00:05:03,720 Speaker 3: A there's a beautiful story about how Saudi Arabia's building 76 00:05:03,759 --> 00:05:07,360 AI refineries and now building AI factor or oil refiners 77 00:05:07,360 --> 00:05:08,160 to AI factors. 78 00:05:08,199 --> 00:05:08,600 Speaker 1: I love that. 79 00:05:10,079 --> 00:05:12,560 Speaker 3: You know, I've said that that AI is an infrastructure 80 00:05:12,560 --> 00:05:14,800 and the reason for that, of course, we understand AI 81 00:05:14,959 --> 00:05:18,480 from the perspective of the technology and how it's revolutionizing 82 00:05:18,560 --> 00:05:24,040 every industry. Digital intelligence, of course has applications into every 83 00:05:24,160 --> 00:05:26,720 every field, and so it's going to be used by 84 00:05:26,759 --> 00:05:31,319 every company, every industry, every country. In that way, it's 85 00:05:31,839 --> 00:05:35,319 foundational and therefore it's part of infrastructure. What is new 86 00:05:35,399 --> 00:05:41,439 about AI from a computer science perspective is that the 87 00:05:41,519 --> 00:05:44,959 way computing was done in the past was largely retrieval 88 00:05:45,079 --> 00:05:49,439 based computing. Somebody typed in a story, or somebody created 89 00:05:49,839 --> 00:05:53,000 a piece of art, or came up with four versions 90 00:05:53,040 --> 00:05:57,399 of a digital ad or it's all pre built by 91 00:05:57,439 --> 00:06:01,360 somebody which is then using a system to retrieve the 92 00:06:01,399 --> 00:06:05,120 appropriate version for you. It's a retrieval based computing model. 93 00:06:06,120 --> 00:06:10,519 Hadoop and many of the frameworks and operating systems that 94 00:06:10,560 --> 00:06:13,959 the past all designed to retrieve the appropriate information for you. 95 00:06:14,360 --> 00:06:18,240 But today software is going to be generated in real time. 96 00:06:18,480 --> 00:06:22,040 It's generative based on the context, based on the circumstance, 97 00:06:22,079 --> 00:06:24,199 based on who you are, based on the problem you 98 00:06:24,199 --> 00:06:27,839 ask that, based on your prompt, it will generate unique 99 00:06:27,839 --> 00:06:31,319 content for you every single time, for everybody. It's unique 100 00:06:31,439 --> 00:06:33,439 when you use GROC every time you use it as 101 00:06:33,480 --> 00:06:37,120 different which based on the right, based on based on 102 00:06:37,160 --> 00:06:39,240 the prompt that you give in and based on the circumstance, 103 00:06:39,279 --> 00:06:44,639 and so therefore it used to be retrieval based. Today 104 00:06:44,759 --> 00:06:49,399 it's generative. And if it's generative, then and every time 105 00:06:49,480 --> 00:06:51,680 is different than you need AI factories all over the 106 00:06:51,720 --> 00:06:54,560 world to generate the content in real time, which is 107 00:06:54,600 --> 00:06:56,839 the reason why you need AI factories. And this is 108 00:06:56,879 --> 00:06:59,519 a unique way of doing computation. But the benefit, of 109 00:06:59,600 --> 00:07:05,360 course is that everything isn't preconceived and predocumented and it's contextually, 110 00:07:05,959 --> 00:07:08,439 contextually sensible and therefore intelligence. 111 00:07:08,639 --> 00:07:12,639 Speaker 1: So AI factories and robotics, and we heard it yesterday 112 00:07:12,639 --> 00:07:16,160 from his Royal Highness, his vision how to augment our 113 00:07:16,240 --> 00:07:20,399 workforce with roughly tens of millions of robotics to be 114 00:07:20,439 --> 00:07:23,959 able to infuse the next wave of productivity and progress. 115 00:07:24,560 --> 00:07:27,199 But this cares a lot of folks here when it 116 00:07:27,199 --> 00:07:30,279 comes to the future of jobs. So let's hear about 117 00:07:30,319 --> 00:07:32,319 your thoughts, Elon and Denson on that. 118 00:07:34,399 --> 00:07:39,279 Speaker 2: Sure, well, you say, like in the long term, where 119 00:07:39,279 --> 00:07:41,319 will things end up long term? I don't know what 120 00:07:41,360 --> 00:07:44,800 long term is. Maybe it's ten twenty years or something 121 00:07:44,839 --> 00:07:48,959 like that. For me, that's long term. My prediction is 122 00:07:48,959 --> 00:07:58,079 that work will be optional optional optional, So we'll take that. Yeah, 123 00:07:58,319 --> 00:08:00,600 I mean it'll be like a playing sports or a 124 00:08:00,680 --> 00:08:03,680 video game or something like that. If you want to 125 00:08:03,720 --> 00:08:06,480 work you know, in the same way like you can 126 00:08:07,000 --> 00:08:10,720 you can go to the store and just buy some vegetables, 127 00:08:10,839 --> 00:08:13,920 or you could grow vegetables in your backyard. It's much 128 00:08:13,920 --> 00:08:16,079 harder to grow veg vegetables in your backyard, but some 129 00:08:16,079 --> 00:08:19,240 people still do it because they like growing growing vegetables. 130 00:08:19,920 --> 00:08:24,399 That will be what work is like optional, and between 131 00:08:24,399 --> 00:08:26,000 now and then, there's actually a lot of work to 132 00:08:26,079 --> 00:08:30,759 get to that point. I always recommend people really read 133 00:08:30,879 --> 00:08:34,840 in banks culture books to get a sense for what 134 00:08:34,919 --> 00:08:41,559 a probable positive AI future is like. And interestingly, in 135 00:08:41,679 --> 00:08:46,759 those books, money is no longer doesn't exist. It's kind 136 00:08:46,759 --> 00:08:50,600 of interesting. And my guess is if you go out 137 00:08:50,639 --> 00:08:54,679 long enough, assuming there's a continued improvement in AI and robotics, 138 00:08:54,720 --> 00:08:59,679 which seems likely, the money will will stop being relevant. 139 00:08:59,679 --> 00:09:03,360 At some point in the future. There will still be 140 00:09:03,919 --> 00:09:08,240 constraints on power like in it, like electricity and mass. 141 00:09:09,240 --> 00:09:13,159 The fundamental physics elements will still be still be constraints, 142 00:09:14,159 --> 00:09:20,320 but I think at some point a currency becomes irrelevant. 143 00:09:22,639 --> 00:09:23,639 Speaker 1: Jensen any thoughts. 144 00:09:30,279 --> 00:09:34,519 Speaker 2: By the way, that video earning schools later today, and 145 00:09:34,519 --> 00:09:34,799 by the. 146 00:09:34,799 --> 00:09:44,200 Speaker 3: Way, since currency is a relevant, Elon just wants to 147 00:09:44,360 --> 00:09:52,600 share with you the two of us would like to 148 00:09:52,600 --> 00:09:56,679 share some breaking news. Uh, let's see. I would say, 149 00:09:56,759 --> 00:09:59,960 I would say, there there's different horizons you can look at. 150 00:10:00,480 --> 00:10:03,519 Everybody's jobs will be different that I think that that's 151 00:10:03,519 --> 00:10:08,360 for sure. How how students learn will be different. How 152 00:10:08,399 --> 00:10:11,039 people do their work will be different, obviously, because a 153 00:10:11,039 --> 00:10:14,679 lot of the things that we do mundanely or arduously 154 00:10:15,120 --> 00:10:17,799 or very difficultly are going to be done very simply, 155 00:10:18,559 --> 00:10:21,399 and so we're going to be more productive from that sense. 156 00:10:21,440 --> 00:10:23,480 From that sense, one of the things that I will 157 00:10:23,519 --> 00:10:28,039 say is that for most people or company, if some 158 00:10:28,279 --> 00:10:31,960 if your life becomes more productive, and if the things 159 00:10:31,960 --> 00:10:36,320 that you're doing with great difficulty becomes simpler, it is 160 00:10:36,440 --> 00:10:39,720 very likely, because you have so many ideas, you'll have 161 00:10:39,759 --> 00:10:41,879 more time to go pursue things. It is my guess 162 00:10:41,919 --> 00:10:44,240 that Elon will be busier as a result of AI. 163 00:10:44,679 --> 00:10:46,480 I'm going to be busier as a result of AI. 164 00:10:46,799 --> 00:10:48,519 And the reason for that is because we have so 165 00:10:48,559 --> 00:10:51,080 many ideas we want to pursue, so many things that 166 00:10:51,600 --> 00:10:54,200 we still have in our backlog inside our company that 167 00:10:54,240 --> 00:10:57,600 we can go pursue. If we're more productive, we can 168 00:10:57,600 --> 00:11:00,679 get to those things faster, and so in the near term, 169 00:11:00,720 --> 00:11:04,440 I would say that there's every evidence that we will 170 00:11:04,480 --> 00:11:07,080 be more productive and yet still be busier because we 171 00:11:07,120 --> 00:11:09,240 have so many ideas. One thing that I will say, 172 00:11:09,840 --> 00:11:13,120 give you some evidence is that and I was just 173 00:11:13,159 --> 00:11:17,639 telling Elan about this earlier. Radiology, for example, has largely 174 00:11:17,720 --> 00:11:21,559 been converted to AI driven radiology, and there's some really 175 00:11:21,559 --> 00:11:26,320 great companies doing that. And the surprising thing is the 176 00:11:26,360 --> 00:11:29,919 prediction that all radiologists would be the first jobs to 177 00:11:30,000 --> 00:11:34,600 go with exactly the opposite. The trend shows that there 178 00:11:34,600 --> 00:11:37,879 are more radiologists being hired now as a result of AI. 179 00:11:38,360 --> 00:11:40,200 And the reason for that, if you take a step back, 180 00:11:40,480 --> 00:11:43,440 it's because the goal of a radiologist is not to 181 00:11:43,440 --> 00:11:47,120 study the images. The goal of a radiologist is to 182 00:11:47,159 --> 00:11:50,919 diagnose a disease. Now the studying of the images became 183 00:11:51,039 --> 00:11:55,279 so productive, they could study more images, study more modalities, 184 00:11:55,440 --> 00:11:58,320 spend more time with the patients, and as a result, 185 00:11:58,559 --> 00:12:01,919 they're actually accepting more patient We're doing more radiology all 186 00:12:01,919 --> 00:12:04,559 around the world, we're doing a better job with diagnosing disease. 187 00:12:05,080 --> 00:12:08,080 And so that's that's kind of the the near term, 188 00:12:08,360 --> 00:12:13,720 near term outcome of AI and productivity, and and we'll see, 189 00:12:13,879 --> 00:12:18,320 we'll see what happens long term, you know, I when 190 00:12:18,399 --> 00:12:21,240 when currency doesn't matter anymore, just you know, let me 191 00:12:21,279 --> 00:12:22,360 know right before. 192 00:12:24,399 --> 00:12:29,080 Speaker 2: You'll see it coming. You'll see it coming, like we. 193 00:12:29,000 --> 00:12:31,960 Speaker 3: Text often, so just yeah, just text it off. 194 00:12:32,480 --> 00:12:34,279 Speaker 1: Let me I kind of I kind of agree with 195 00:12:34,279 --> 00:12:36,720 with both of you because if you look at every 196 00:12:36,799 --> 00:12:41,440 technological trend, every general purpose technology has been that new 197 00:12:41,639 --> 00:12:45,559 positive for for the globe, for humanity and so forth. 198 00:12:45,600 --> 00:12:48,879 Speaker 3: And let me share with you too, and your excellency, 199 00:12:48,919 --> 00:12:51,159 I think it's precisely the reason. The reason for that 200 00:12:51,279 --> 00:12:54,440 is because all the great ideas from from innovators like 201 00:12:54,519 --> 00:12:59,159 elon that you have so many ideas that Jackson is wrong. Yeah, 202 00:12:59,240 --> 00:13:00,519 well yeah, thank you. 203 00:13:01,080 --> 00:13:03,159 Speaker 1: So let me share share with you two stories from 204 00:13:03,279 --> 00:13:06,399 two Saudi innovators in collaboration with a lot of great 205 00:13:06,440 --> 00:13:10,200 the great work that Nvidia does, that gro does. One 206 00:13:10,279 --> 00:13:17,360 is Professor Mariagi. All right, m got a game. It's 207 00:13:17,440 --> 00:13:20,000 Professor Mariagi. 208 00:13:23,440 --> 00:13:32,639 Speaker 2: That might need to might need to move the Micay. 209 00:13:29,080 --> 00:13:33,200 Speaker 3: Come on, what's share this one. 210 00:13:33,879 --> 00:13:36,600 Speaker 1: Let's let's let's write this one more time. So one 211 00:13:36,639 --> 00:13:40,679 of them is Professor Aumriagi, who's the first American Saudi 212 00:13:40,919 --> 00:13:44,679 to win a Nobel Prize in creating new chemistry, and 213 00:13:44,799 --> 00:13:46,879 the way he has done that he has leveraged you 214 00:13:46,960 --> 00:13:51,440 AI accelerators and models like Grock to be able to 215 00:13:51,519 --> 00:13:54,600 create new chemistry. When it comes to metal organic frameworks, 216 00:13:54,960 --> 00:13:59,080 those are metal allions that are positively charged with organic 217 00:13:59,159 --> 00:14:02,960 clinkers to be able to effectively create a sponge withero 218 00:14:02,960 --> 00:14:07,879 point three three nanometers pores to capture water from air 219 00:14:08,480 --> 00:14:13,080 and also to capture carbon dioxide. The second story has 220 00:14:13,120 --> 00:14:17,039 also to do with AI accelerated by Nvidia and with 221 00:14:17,200 --> 00:14:21,120 models like Rock, which is nanopalm, which is effectively creating 222 00:14:21,120 --> 00:14:25,519 a nano robot five hundred nanometers by one thousand nanometers 223 00:14:25,559 --> 00:14:29,080 to be able to do gene editing, leveraging the crisper 224 00:14:29,159 --> 00:14:33,039 technology to take out sickill cell disease. Now, in both 225 00:14:33,080 --> 00:14:37,559 these instances they originated twenty years ago in research, but 226 00:14:37,720 --> 00:14:42,320 AI was able to really accelerate the outcomes and the 227 00:14:42,440 --> 00:14:45,720 outputs such that we can move into new value pools. 228 00:14:45,879 --> 00:14:49,399 So I think with every technological trend, humanity is going 229 00:14:49,480 --> 00:14:52,679 to always manage to shift to new value pools when 230 00:14:52,720 --> 00:14:55,679 it comes to workforce on productivity. But we have some 231 00:14:55,720 --> 00:14:58,879 great announcements to talk about here today. Let's begin with 232 00:14:58,919 --> 00:15:01,320 you Elon. The thing that we're doing with x AI. 233 00:15:02,159 --> 00:15:05,039 Speaker 2: Oh yeah, we're excited to announce that we're doing a 234 00:15:05,639 --> 00:15:10,600 a five hundred megawatt I mean yeah, five hundred sorry, 235 00:15:12,360 --> 00:15:13,639 five mega mega. 236 00:15:13,879 --> 00:15:17,399 Speaker 1: Yeah, we're doing. 237 00:15:19,559 --> 00:15:24,279 Speaker 2: Sorry yeah, yeah, it's five hundred gigwak one laftway so 238 00:15:25,960 --> 00:15:37,960 that that that'll be eight bazilion trillion dollars that uh so, yeah, 239 00:15:37,960 --> 00:15:43,000 we're doing a x A I and the Kingdom Saudi 240 00:15:43,000 --> 00:15:45,039 Arabia are doing a. 241 00:15:45,720 --> 00:15:48,639 Speaker 1: Mega watts starting with fifty megawatts phase one, and we're 242 00:15:48,639 --> 00:15:52,120 doing it with Nvidio. Congratulations to the Humane team to 243 00:15:52,279 --> 00:15:56,840 Target team, such a fantastic job, Jensen. I think we're 244 00:15:56,879 --> 00:15:59,120 also doing some great an ausments this week. 245 00:15:59,759 --> 00:16:03,919 Speaker 3: We are, We're we're announcing, we're announcing all kinds of things. 246 00:16:04,720 --> 00:16:07,840 Our partnership with Humane is going incredibly well. First of all, 247 00:16:08,679 --> 00:16:12,360 we work together to get this company started and off 248 00:16:12,399 --> 00:16:16,639 the ground and just got an incredible customer with Elon. 249 00:16:17,720 --> 00:16:21,919 Could you imagine a startup company approximately zero billion dollars 250 00:16:21,919 --> 00:16:25,679 in revenues now going to build a data center for Elon. 251 00:16:26,240 --> 00:16:30,200 Five hundred megawatts is gigantic. This company is off the charts, 252 00:16:30,480 --> 00:16:36,519 right away. In addition to that, we're working working AWS, 253 00:16:36,559 --> 00:16:37,320 as you know, is. 254 00:16:37,279 --> 00:16:42,200 Speaker 1: Also caraculations to the HUMANE team, with the AWS starting 255 00:16:42,200 --> 00:16:46,399 with one hundred megawatts with a gigawatts ambition and counting. 256 00:16:47,000 --> 00:16:50,120 Speaker 3: So AWS is also coming to Humane. We're working with 257 00:16:50,759 --> 00:16:55,759 Humane on Omniverse Digital twins. As you know that AI 258 00:16:55,879 --> 00:17:00,240 is not just well just agentic AI and chatbot and 259 00:17:00,840 --> 00:17:04,400 cognitive AI is incredibly important to the world, but AI 260 00:17:04,519 --> 00:17:08,480 applies to everything chemicals and proteins and genes, and physics 261 00:17:08,519 --> 00:17:12,480 and fluid dynamnamics and particles and of course robotics and activation. 262 00:17:12,640 --> 00:17:16,720 And we created this world called Omniverse where robots can 263 00:17:16,799 --> 00:17:18,119 learn how to be good robots, and. 264 00:17:19,680 --> 00:17:20,640 Speaker 2: It's physically based. 265 00:17:20,680 --> 00:17:23,440 Speaker 3: It obeys the laws of physics, and so robots can 266 00:17:23,519 --> 00:17:26,559 learn in these environments. And we're working with Humane to 267 00:17:26,799 --> 00:17:31,440 apply Omniverse to all kinds of digital factories and robotics 268 00:17:31,440 --> 00:17:34,400 and warehouses and things like that, and so that's another 269 00:17:35,000 --> 00:17:39,000 We're also working in Saudi Arabia to build supercomputers to 270 00:17:39,039 --> 00:17:44,599 simulate quantum computers and using our computers to be the 271 00:17:44,680 --> 00:17:49,160 controller and the error correction. Quantum error correction requires an 272 00:17:49,279 --> 00:17:51,920 enormous amount of computation and so we're doing a lot 273 00:17:51,960 --> 00:17:54,799 of great work there too, So a big partnership with Humane. 274 00:17:54,839 --> 00:17:57,440 They're off the charts, off the ground and off the 275 00:17:57,519 --> 00:17:58,319 charts at the same time. 276 00:17:58,519 --> 00:18:01,319 Speaker 1: This is how we walk the top in the Kingdom 277 00:18:01,359 --> 00:18:05,440 of sud Arabia and partnership with the US. Yesterday the 278 00:18:05,480 --> 00:18:09,119 President and His Royal Highness announced the AI Strategic Framework 279 00:18:09,160 --> 00:18:13,839 and partnership. Today we're going big with Elon and Jensen. 280 00:18:14,039 --> 00:18:21,759 So thank you for those opportunities. Now, they told me 281 00:18:21,799 --> 00:18:25,319 I have time for two last questions. So last night 282 00:18:25,440 --> 00:18:28,640 at the dinner I got a number of questions because 283 00:18:28,640 --> 00:18:32,640 it seems that the schedule leaked and everybody was getting 284 00:18:32,680 --> 00:18:34,880 me hints about the last two questions I'm going to do. 285 00:18:35,480 --> 00:18:38,920 So the first one was for you, Elon, and there's 286 00:18:38,960 --> 00:18:42,400 a big one for you, Jensen, So prepare for that one. 287 00:18:43,680 --> 00:18:46,359 AI in space Is that possible? 288 00:18:48,279 --> 00:18:53,039 Speaker 2: Yes, If civilization continues, which it probably will, then AI 289 00:18:53,039 --> 00:19:00,480 in space is inevitable. You know, I have to like 290 00:19:00,640 --> 00:19:04,720 preface that, you know, we shouldn't take civilization for granted. 291 00:19:04,839 --> 00:19:08,119 We need to make sure to take care to ensure 292 00:19:08,119 --> 00:19:11,720 that civilization hasn't an upward arc. I mean any student 293 00:19:11,720 --> 00:19:14,680 of history knows that civilization does not always have an 294 00:19:14,720 --> 00:19:18,640 upward arc, and in fact, civilizations have life cycles. So 295 00:19:18,759 --> 00:19:21,200 hopefully we are in a strong upward arc. I think 296 00:19:21,240 --> 00:19:23,319 we are for now, but we don't want to take 297 00:19:23,359 --> 00:19:27,759 that for granted or be complacent. But in order to 298 00:19:28,279 --> 00:19:30,880 the way to think of AI in space is that 299 00:19:30,960 --> 00:19:33,960 in order to achieve any meaningful percentage of a Kardashev 300 00:19:34,000 --> 00:19:37,759 two scale civilization where you're using even a millionth of 301 00:19:37,960 --> 00:19:42,559 a millionth of the Sun's energy, you must have solar 302 00:19:42,640 --> 00:19:49,160 powered AI satellites in deep space. So that once you realize, 303 00:19:49,160 --> 00:19:51,440 like once you think in terms of a Cartoshev two 304 00:19:51,480 --> 00:19:55,519 scale civilization, which is what percentage of the Sun's energy 305 00:19:55,799 --> 00:19:59,960 are you turning into useful work, then then it becomes 306 00:20:00,079 --> 00:20:06,559 obvious that space is overwhelmingly what matters. Overwhelmingly, the Sun 307 00:20:06,640 --> 00:20:10,440 only receives one roughly one two billionth of the Earth 308 00:20:10,440 --> 00:20:14,559 only receives roughly uh one two billionth of the Sun's energy. 309 00:20:15,680 --> 00:20:19,039 So if you want to have something that is, say 310 00:20:19,400 --> 00:20:22,440 a million times more energy than Earth could possibly produce, 311 00:20:23,400 --> 00:20:30,400 you must go into space. It's and so this is 312 00:20:30,400 --> 00:20:32,359 where it's kind of handy to have a space company, 313 00:20:32,400 --> 00:20:33,359 I guess. 314 00:20:35,000 --> 00:20:38,119 Speaker 3: Sell the books to cool chips in space too. Yes, 315 00:20:38,279 --> 00:20:39,759 easier to cool chips in space. 316 00:20:40,160 --> 00:20:42,119 Speaker 2: Yes, there's definitely no water in space, so you're gonna 317 00:20:42,119 --> 00:20:46,440 have to do something that doesn't involve water. Well, it's good, 318 00:20:46,519 --> 00:20:51,960 just got to radiate, that's right. So my estimate is 319 00:20:52,000 --> 00:20:58,359 that actually that that that the cost of electricity, like 320 00:20:58,359 --> 00:21:02,680 like the cost effectiveness of AI in space will be 321 00:21:03,000 --> 00:21:06,079 overwhelmingly better than AI on the ground. So for long 322 00:21:06,119 --> 00:21:11,519 before you exhaust potential energy sources on Earth, long long before, 323 00:21:11,680 --> 00:21:15,160 meaning like I think even peraps in the four or 324 00:21:15,200 --> 00:21:18,880 five year time frame, the lowest cost way to do 325 00:21:19,400 --> 00:21:24,839 AI compute will be with solar powered AI satellites. So 326 00:21:26,160 --> 00:21:28,160 I'd say not more than five years from now. 327 00:21:28,519 --> 00:21:32,720 Speaker 3: Wow, And just look at the supercomputers we're building together. 328 00:21:33,039 --> 00:21:35,039 Let's say each one of the racks is two tons. 329 00:21:35,440 --> 00:21:37,960 Out of that two tons, one point nine to five 330 00:21:38,000 --> 00:21:42,240 of it is probably for cooling, right, Just imagine how 331 00:21:42,279 --> 00:21:44,599 tiny that little supercomputer is, right, each one of these 332 00:21:44,640 --> 00:21:47,160 GB three hundred racks and just be a little tiny thing. 333 00:21:47,359 --> 00:21:51,880 Speaker 2: And just electricity generation is already becoming a challenge. So 334 00:21:52,119 --> 00:21:55,599 if you start doing any kind of scaling for both 335 00:21:55,640 --> 00:22:01,079 electricity generation and cooling. You realize, Okay, space incredibly compelling. 336 00:22:02,559 --> 00:22:07,079 So like let's say you wanted to do you know, 337 00:22:07,119 --> 00:22:13,119 two or three hundred gigawatts per year of AI compute. 338 00:22:14,079 --> 00:22:17,240 It's very difficult to do that on Earth. So the 339 00:22:18,000 --> 00:22:21,240 US average electricity usage last time I checked it was 340 00:22:21,279 --> 00:22:25,119 around four hundred and sixty gigawatts per year average usage. 341 00:22:26,240 --> 00:22:31,119 So so something like say, you know, three three hundred, 342 00:22:31,119 --> 00:22:32,920 If you're doing three hundred giga, what's a year, that 343 00:22:32,960 --> 00:22:35,839 would be like two thirds of US electricity production per year. 344 00:22:37,000 --> 00:22:39,319 There's no way you're building power plants at that level. 345 00:22:40,599 --> 00:22:41,799 And then if you take it up to say a 346 00:22:41,880 --> 00:22:45,079 taro wat per year, impossible, Like you have to do 347 00:22:45,119 --> 00:22:48,440 that in space. There there just is. There just is 348 00:22:48,480 --> 00:22:51,480 no way to do a tarowa per year on Earth. 349 00:22:53,720 --> 00:22:59,079 And in space you've you've got continuous solar you've got 350 00:23:00,160 --> 00:23:02,440 you don't you actually don't need batteries because it's always 351 00:23:02,440 --> 00:23:07,079 sunny in space, right exactly, And and the solar panels 352 00:23:07,119 --> 00:23:09,440 actually become cheaper because you don't need glass or framing, 353 00:23:11,079 --> 00:23:14,880 and the cooling is just RADIATI So that's that's why 354 00:23:14,920 --> 00:23:15,359 I think. 355 00:23:15,200 --> 00:23:16,279 Speaker 3: That's the dream. 356 00:23:16,599 --> 00:23:18,000 Speaker 2: Yes, that's the dream. 357 00:23:18,240 --> 00:23:21,079 Speaker 1: So Jensen, everybody last night was asking me, and I'm 358 00:23:21,119 --> 00:23:25,400 mindful it's Earning's call for you today. So I'm gonna 359 00:23:25,400 --> 00:23:28,759 say this delicately. Everybody has been asking me to ask you. 360 00:23:29,160 --> 00:23:30,519 Are we going to have an AI bubble? 361 00:23:32,039 --> 00:23:33,000 Speaker 3: That's the last question. 362 00:23:33,640 --> 00:23:34,400 Speaker 1: All right, let's not. 363 00:23:36,400 --> 00:23:37,799 Speaker 3: All right, let me see well, let me just tell 364 00:23:37,799 --> 00:23:40,599 you what we see. Okay. So I think it's really 365 00:23:40,640 --> 00:23:43,279 important when you look at what's happening around the world 366 00:23:43,319 --> 00:23:45,559 and go back to the first principles of what's happening in 367 00:23:45,559 --> 00:23:49,200 computer science and computing. There are three things that's happening. 368 00:23:49,519 --> 00:23:52,279 The first thing is that we all know that Moore's 369 00:23:52,359 --> 00:23:54,920 laws run its course, and the ability the the amount 370 00:23:54,920 --> 00:23:57,640 of demand for computing versus the amount of computation we 371 00:23:57,680 --> 00:24:00,359 can get out of general purpose computing has really chanting, lenging, 372 00:24:00,759 --> 00:24:03,359 and so the world's been moving to accelerated computing for 373 00:24:03,400 --> 00:24:05,279 some time. We've been pushing this now for some over 374 00:24:05,279 --> 00:24:08,319 twenty years. Let me give you one statistic. I was 375 00:24:08,400 --> 00:24:15,920 just as supercomputing six years ago. CPUs were ninety percent 376 00:24:16,559 --> 00:24:19,759 of the world supercomputers top five hundred supercomputers. 377 00:24:19,119 --> 00:24:19,920 Speaker 1: Six years ago. 378 00:24:20,400 --> 00:24:24,880 Speaker 3: This year, less than fifteen percent went from ninety percent 379 00:24:24,960 --> 00:24:28,200 to ten percent, and meanwhile accelerated computing went from the 380 00:24:28,200 --> 00:24:31,680 other way ten percent to now ninety percent. Okay, so 381 00:24:31,759 --> 00:24:35,440 you're seeing that inflection point, the transition in high performance 382 00:24:35,440 --> 00:24:38,680 computing from general purpose of computing to accelerated computer. Well, 383 00:24:38,920 --> 00:24:41,519 of the one of the most data intensive, one of 384 00:24:41,559 --> 00:24:43,759 the most intensive computation things that the world does in 385 00:24:43,880 --> 00:24:48,519 cloud is data processing. Several hundred billion dollars of computation 386 00:24:48,720 --> 00:24:51,000 is done on just raw data. Process has nothing to 387 00:24:51,000 --> 00:24:54,960 do with AI, just CQL processing data frames. You know 388 00:24:54,960 --> 00:24:59,160 everybody's names, address, their sex, their their age, where they live, 389 00:24:59,240 --> 00:25:01,279 you know how much money they make. All of that 390 00:25:01,319 --> 00:25:03,759 sits into a data frame, and that data frame drives 391 00:25:03,799 --> 00:25:06,839 the world today, whether it's in banking, or whether it's 392 00:25:06,880 --> 00:25:10,440 in credit cards or of course e commerce and everything 393 00:25:10,480 --> 00:25:13,799 from ad recommendation and everything is driven off of that 394 00:25:13,920 --> 00:25:16,039 data frame. That data frame costs hundreds of billions of 395 00:25:16,039 --> 00:25:18,480 dollars to go compute. And so that's the number one 396 00:25:18,519 --> 00:25:21,720 thing end of More's laws. The second thing is generative AI. 397 00:25:23,720 --> 00:25:28,720 The most important application of the last fifteen years is 398 00:25:28,759 --> 00:25:32,519 called rexis recommended systems. How do we know what information 399 00:25:32,599 --> 00:25:35,519 to recommend to us in a social feed. How do 400 00:25:35,559 --> 00:25:38,400 you know what ad to recommend to somebody, what book 401 00:25:38,480 --> 00:25:41,039 to recommend, what movie to recommend? The world is the 402 00:25:41,160 --> 00:25:44,319 Internet is so gigantic without a recommender system that a 403 00:25:44,359 --> 00:25:46,799 little tiny flo of us will have no chance. I've 404 00:25:46,839 --> 00:25:50,759 ever seen the right information that rexis is the engine 405 00:25:50,759 --> 00:25:54,200 of the Internet today. That's going generative AI. It used 406 00:25:54,240 --> 00:25:56,319 to be running on CPUs, now it runs on GPS, 407 00:25:56,440 --> 00:25:59,519 which then says the third thing when if you just 408 00:25:59,599 --> 00:26:04,079 look at those two applications, many of the Internet companies 409 00:26:04,480 --> 00:26:08,839 can build an enormous number of GPUs supercomputers. Just doing that, 410 00:26:09,319 --> 00:26:12,480 of course, then it creates this the third opportunity on 411 00:26:12,519 --> 00:26:14,880 top of it, which is agentic AI. This is GROCK 412 00:26:14,920 --> 00:26:17,319 and this is open AI, this is anthropic you know, 413 00:26:17,400 --> 00:26:20,119 this is Gemini. Agentic AI sits on top of that. 414 00:26:20,359 --> 00:26:23,160 But don't you know, don't forget to think about what 415 00:26:23,319 --> 00:26:29,039 is happening above underneath what everybody sees as AI today, 416 00:26:29,319 --> 00:26:32,599 there's a whole movement of computing from Jenner purpose computing 417 00:26:32,640 --> 00:26:35,440 to accelerated computing, and that if you just if you 418 00:26:35,480 --> 00:26:39,039 take that into consideration, you'll come to the conclusion that 419 00:26:39,200 --> 00:26:43,359 in fact, what is left over to fuel that revolutionary 420 00:26:43,400 --> 00:26:47,400 agentic AI is not only substantially less than your thought 421 00:26:47,839 --> 00:26:49,359 and all of it justified. 422 00:26:50,000 --> 00:26:52,799 Speaker 1: Well, I was just informed by the team that my 423 00:26:53,160 --> 00:26:56,160 boss and your bosses is going to talk next the 424 00:26:56,200 --> 00:26:59,960 Honorable President and his Royal Highness the Common Prince and him. 425 00:27:00,119 --> 00:27:05,720 We ran out of time. But in essence, this is 426 00:27:05,880 --> 00:27:10,359 our such so much love for you Elon and Jensen. 427 00:27:10,880 --> 00:27:15,119 But this, in essence is a ninety two alliance that 428 00:27:15,279 --> 00:27:20,039 shifted from energy to digital to the intelligence age, powered 429 00:27:20,079 --> 00:27:24,039 by pioneers such as Elon and Jensen to serve humanity 430 00:27:24,160 --> 00:27:28,359 and create on a net new basis, new economies, new jobs, 431 00:27:28,400 --> 00:27:31,960 and a better future for humanity, powered by the Kingdom 432 00:27:31,960 --> 00:27:34,920 of Arabia and the United States. Thank you for our 433 00:27:34,960 --> 00:27:38,680 lifetime partnership and friendship. Thank you Elon, Thank you Jensen.