1 00:00:02,720 --> 00:00:15,640 Speaker 1: Bloomberg Audio Studios, Podcasts, radio News. 2 00:00:17,760 --> 00:00:20,680 Speaker 2: Hello man, Welcome to another episode of the All Thoughts Podcast. 3 00:00:20,800 --> 00:00:22,119 Speaker 2: I'm Tracy Alloway. 4 00:00:21,920 --> 00:00:25,239 Speaker 3: And I'm Joe Wisenthal. Joe, Yes, have you. 5 00:00:25,200 --> 00:00:27,240 Speaker 2: Ever wanted to tour a data center? 6 00:00:28,280 --> 00:00:28,840 Speaker 3: I'd love to. 7 00:00:29,800 --> 00:00:32,120 Speaker 2: Part of me thinks it would be really boring, like 8 00:00:32,200 --> 00:00:34,400 Speaker 2: if you actually went inside, because it would just be 9 00:00:34,760 --> 00:00:37,320 Speaker 2: a bunch of the same cables, like over and over 10 00:00:37,360 --> 00:00:40,280 Speaker 2: and over again. But part of me thinks just seeing 11 00:00:40,320 --> 00:00:42,400 Speaker 2: the scale of it, yeah, would be amazing. 12 00:00:42,680 --> 00:00:44,120 Speaker 4: You know what I would do if I toured a 13 00:00:44,200 --> 00:00:47,559 Speaker 4: data center is I would look at those cables and 14 00:00:47,680 --> 00:00:52,080 Speaker 4: every nut and bolt and vec and heating thing and 15 00:00:52,159 --> 00:00:55,200 Speaker 4: get every brand name I could on any little piece 16 00:00:55,200 --> 00:00:58,360 Speaker 4: of equipment in the entire thing, and then see are 17 00:00:58,400 --> 00:01:00,280 Speaker 4: there any of these that are AI play? Is that 18 00:01:00,280 --> 00:01:03,800 Speaker 4: the market doesn't really you know, any random screwdriver that's 19 00:01:03,840 --> 00:01:05,600 Speaker 4: being used in this operation? 20 00:01:06,319 --> 00:01:08,319 Speaker 3: What is the name? Is it publicly traded? 21 00:01:08,440 --> 00:01:09,720 Speaker 2: So literally picks and shovel? 22 00:01:09,880 --> 00:01:11,000 Speaker 3: Yeah, just exactly. 23 00:01:11,160 --> 00:01:13,360 Speaker 4: Oh here is here is a company that makes the 24 00:01:14,200 --> 00:01:17,280 Speaker 4: you know, the doorstop. Are they an AI play? Because 25 00:01:17,280 --> 00:01:18,880 Speaker 4: they're going to need a lot of doorstops for all 26 00:01:18,880 --> 00:01:20,240 Speaker 4: these doors. That's what I would do. 27 00:01:20,400 --> 00:01:25,040 Speaker 2: Okay, So I guess the Joe hyper Scale data Center 28 00:01:25,120 --> 00:01:29,160 Speaker 2: ETF coming to a market near you eventually, that would 29 00:01:29,160 --> 00:01:29,600 Speaker 2: be a good idea. 30 00:01:29,640 --> 00:01:30,640 Speaker 5: It's a great idea. 31 00:01:30,800 --> 00:01:33,479 Speaker 2: Well, today we're going to talk about some of these 32 00:01:33,520 --> 00:01:37,039 Speaker 2: data centers, yeah, with someone who has actually gone on 33 00:01:37,080 --> 00:01:38,800 Speaker 2: a field trip and looked at them, and then we're 34 00:01:38,800 --> 00:01:42,760 Speaker 2: going to talk about broader AI because of course with 35 00:01:42,920 --> 00:01:45,880 Speaker 2: markets still, you know, basically at records, there's a lot 36 00:01:45,880 --> 00:01:48,280 Speaker 2: of concern that valuations are getting out of hand. We've 37 00:01:48,280 --> 00:01:52,440 Speaker 2: had these financing deals which just do my head in 38 00:01:52,440 --> 00:01:55,360 Speaker 2: in terms of like trying to track who's lending to who, 39 00:01:55,400 --> 00:01:57,760 Speaker 2: and who's buying from who and all of that. So 40 00:01:57,760 --> 00:01:58,680 Speaker 2: we should talk about. 41 00:01:58,520 --> 00:02:00,880 Speaker 4: It totally because I mean two things, which is one, 42 00:02:01,040 --> 00:02:04,680 Speaker 4: we know how important AI is to the stock market. Two, 43 00:02:05,120 --> 00:02:07,600 Speaker 4: you know, there's all these sort of complicated financing deals 44 00:02:07,640 --> 00:02:09,359 Speaker 4: and the number is changed by the day. So just 45 00:02:09,400 --> 00:02:12,200 Speaker 4: today we got news that in the latest open AI 46 00:02:12,320 --> 00:02:14,240 Speaker 4: is allowing some of its employees to sell shares at 47 00:02:14,240 --> 00:02:17,160 Speaker 4: a half a trillion dollar valuation. It just seems like 48 00:02:17,200 --> 00:02:20,520 Speaker 4: it's constant fundraised, constant numbers that go higher and hire 49 00:02:20,520 --> 00:02:23,480 Speaker 4: and hire every day. Various estimates for how much capital 50 00:02:23,480 --> 00:02:26,800 Speaker 4: expenditure is going to happen in this cycle seem to 51 00:02:26,840 --> 00:02:29,000 Speaker 4: go up. I feel like what we really need to 52 00:02:29,040 --> 00:02:31,200 Speaker 4: do is have one of these conversations every month and 53 00:02:31,280 --> 00:02:33,960 Speaker 4: just get for real, like we need to keep like 54 00:02:34,200 --> 00:02:36,440 Speaker 4: it changes so fast and the stake seems so high. 55 00:02:36,560 --> 00:02:40,079 Speaker 2: Yeah, you definitely need frequent updates. Yes. The other thing, 56 00:02:40,160 --> 00:02:43,120 Speaker 2: of course, hovering in the background is the importance of 57 00:02:43,320 --> 00:02:46,320 Speaker 2: the AI build out to the overall economy, right and 58 00:02:46,400 --> 00:02:48,560 Speaker 2: this is the question, like, is the I think Deutsche 59 00:02:48,600 --> 00:02:50,760 Speaker 2: Bank actually put out a note on this is the 60 00:02:50,919 --> 00:02:54,120 Speaker 2: entire US economy being propped up by AI right now? 61 00:02:54,639 --> 00:02:57,160 Speaker 2: Maybe that's not a problem if you know, if the 62 00:02:57,200 --> 00:03:00,640 Speaker 2: AI stuff actually gets built and materializes in real money. 63 00:03:00,680 --> 00:03:03,160 Speaker 2: But you can imagine if it is a bubble, then 64 00:03:03,160 --> 00:03:03,840 Speaker 2: that would be bad. 65 00:03:04,000 --> 00:03:06,200 Speaker 4: The way I see it is, there's essentially two ways 66 00:03:06,400 --> 00:03:09,440 Speaker 4: this is going to end up. One is it turns 67 00:03:09,440 --> 00:03:11,400 Speaker 4: out it was a bubble, we have a massive recession 68 00:03:11,440 --> 00:03:14,000 Speaker 4: and we all lose our jobs. And the other possibility 69 00:03:14,040 --> 00:03:16,560 Speaker 4: is that AI is not a bubble, we have AGI 70 00:03:16,840 --> 00:03:18,880 Speaker 4: and we all lose our jobs. And so I'm curious 71 00:03:18,919 --> 00:03:21,040 Speaker 4: which of the two paths is going to happen. 72 00:03:21,280 --> 00:03:24,839 Speaker 2: Well, I gotta say there are some alternate possibilities. Did 73 00:03:24,840 --> 00:03:26,560 Speaker 2: you see the Dallas fed chart. 74 00:03:27,240 --> 00:03:28,120 Speaker 3: I haven't seen that one. 75 00:03:28,160 --> 00:03:31,560 Speaker 2: So it showed GDP per capita and one line was 76 00:03:31,639 --> 00:03:35,920 Speaker 2: like for agi and it just goes straightly. Yeah. But 77 00:03:36,000 --> 00:03:39,280 Speaker 2: then the other line is what if the robots take 78 00:03:39,320 --> 00:03:41,960 Speaker 2: over the world and kill everyone? And that line went 79 00:03:42,040 --> 00:03:42,600 Speaker 2: straight down. 80 00:03:42,960 --> 00:03:47,840 Speaker 3: So yeah, okay, okay, lots of good outcomes here. 81 00:03:47,960 --> 00:03:50,280 Speaker 2: Yes, all right, So we have the perfect guest, someone 82 00:03:50,320 --> 00:03:55,160 Speaker 2: we've spoken to before, who writes absolutely phenomenal research about 83 00:03:55,240 --> 00:03:58,680 Speaker 2: AI and actually digs into literally the nuts and bolts 84 00:03:58,720 --> 00:04:01,200 Speaker 2: of everything we're going to be even with. James Van Geln. 85 00:04:01,440 --> 00:04:05,680 Speaker 2: He is the author of the Satrini newsletter. James, thanks 86 00:04:05,720 --> 00:04:06,440 Speaker 2: for coming back on. 87 00:04:06,680 --> 00:04:10,040 Speaker 5: Thanks for having me. I feel like the last time 88 00:04:10,040 --> 00:04:12,440 Speaker 5: we were here we were also discussing whether a is 89 00:04:12,440 --> 00:04:13,920 Speaker 5: a bubble or not. And yeah, we get to do 90 00:04:13,960 --> 00:04:14,280 Speaker 5: it again. 91 00:04:15,240 --> 00:04:17,279 Speaker 3: Well, it'll be right right. 92 00:04:17,600 --> 00:04:21,440 Speaker 2: Analyst going on field trips is one of my favorite 93 00:04:21,520 --> 00:04:25,960 Speaker 2: genres of research. Explain to us how you managed to 94 00:04:26,040 --> 00:04:26,440 Speaker 2: do this. 95 00:04:27,279 --> 00:04:33,640 Speaker 5: So drones have become very commoditized, the likelihood of anything 96 00:04:33,680 --> 00:04:35,920 Speaker 5: in the world being more than a ten mile radius 97 00:04:35,920 --> 00:04:41,160 Speaker 5: from someone with a drone is extremely low and we 98 00:04:41,279 --> 00:04:43,320 Speaker 5: kind of use that to our advantage because we've been 99 00:04:43,320 --> 00:04:45,479 Speaker 5: sitting in front of a screen watching green numbers go 100 00:04:45,600 --> 00:04:47,679 Speaker 5: up and up and up. And it's not my fault 101 00:04:47,720 --> 00:04:49,880 Speaker 5: that they named it stargate. It kind of lends itself 102 00:04:49,920 --> 00:04:52,840 Speaker 5: to not taking it as serious as one otherwise might. 103 00:04:53,480 --> 00:04:57,000 Speaker 5: And someone I've read a tweet that said, well, where's 104 00:04:57,040 --> 00:05:00,200 Speaker 5: the proof. There probably should be proof, right, So we 105 00:05:00,240 --> 00:05:02,880 Speaker 5: looked at some satellite images and we went, Okay, a 106 00:05:02,920 --> 00:05:05,600 Speaker 5: year ago this was dirt, and now there's this thing, 107 00:05:06,200 --> 00:05:07,920 Speaker 5: and it's the size of Lower Manhattan. 108 00:05:08,400 --> 00:05:10,840 Speaker 2: Wow. So in Texas it's. 109 00:05:10,760 --> 00:05:14,120 Speaker 5: An apple in Texas, and the I mean this is 110 00:05:14,160 --> 00:05:16,120 Speaker 5: actually one of the smaller ones. I mean, if you 111 00:05:16,200 --> 00:05:21,679 Speaker 5: look at Louisiana metas Hyperion, it would start the footprint 112 00:05:21,720 --> 00:05:23,640 Speaker 5: would start at the top of Central Park and it 113 00:05:23,640 --> 00:05:25,520 Speaker 5: would go down to Soho wow. 114 00:05:25,760 --> 00:05:29,039 Speaker 2: So so that's why they call it hyper Scare. 115 00:05:29,160 --> 00:05:31,920 Speaker 4: So just to be clear with Stargate, this is like 116 00:05:31,960 --> 00:05:36,040 Speaker 4: the thing that, like Trump and Oracle and probably Soft Bank, 117 00:05:36,160 --> 00:05:39,279 Speaker 4: is in the middle, like, what is this complex? Who's 118 00:05:39,320 --> 00:05:43,240 Speaker 4: behind it? This thing that you went and toured and 119 00:05:43,279 --> 00:05:44,360 Speaker 4: got some drone footage. 120 00:05:44,480 --> 00:05:48,880 Speaker 5: So it's essentially a mix of you know, Oracle Open AI. 121 00:05:49,200 --> 00:05:53,400 Speaker 5: There's some involvement from Nvideo, there's Saudi interest with mg X, 122 00:05:53,640 --> 00:05:58,119 Speaker 5: there is some financing involved, and it is just one 123 00:05:58,400 --> 00:06:02,520 Speaker 5: of currently data centers that are planned with a half 124 00:06:02,560 --> 00:06:05,000 Speaker 5: a trillion dollars of investment. That's why we went and 125 00:06:05,000 --> 00:06:07,440 Speaker 5: looked at it because when you really think about it, 126 00:06:07,480 --> 00:06:13,320 Speaker 5: this represents probably the largest infrastructure build out and effort 127 00:06:13,400 --> 00:06:14,280 Speaker 5: since World War Two. 128 00:06:15,160 --> 00:06:16,760 Speaker 2: Where are they getting the energy from? 129 00:06:17,080 --> 00:06:19,920 Speaker 5: So that's that's the best part. That was the first 130 00:06:19,920 --> 00:06:22,040 Speaker 5: thing we saw on the drone is you fly over 131 00:06:22,080 --> 00:06:26,719 Speaker 5: it and they just built their own natural gas plant 132 00:06:27,360 --> 00:06:31,360 Speaker 5: and going like Joe, great idea, just going and looking 133 00:06:31,400 --> 00:06:33,000 Speaker 5: at all the but you got to look at the 134 00:06:33,040 --> 00:06:37,679 Speaker 5: big stuff too, right and without i mean without power. 135 00:06:37,720 --> 00:06:40,760 Speaker 5: These are just kind of you know, hunks of metal. 136 00:06:40,960 --> 00:06:47,159 Speaker 5: So you have outside of Stargate aveline ten natural gas turbines. 137 00:06:47,680 --> 00:06:51,039 Speaker 5: And the interesting thing is these aren't like the really 138 00:06:51,040 --> 00:06:54,599 Speaker 5: good natural gas turbines because if you wanted so, natural 139 00:06:54,600 --> 00:06:58,120 Speaker 5: gas turbines fall along simple cycle combined cycle. These are 140 00:06:58,240 --> 00:07:01,200 Speaker 5: simple cycle. They're each thirty five megawatts, which is very 141 00:07:01,240 --> 00:07:03,480 Speaker 5: much on the lower end. Half of them are from 142 00:07:03,560 --> 00:07:05,520 Speaker 5: geev right now, about half of them are from Caterpillar, 143 00:07:05,520 --> 00:07:08,000 Speaker 5: a company the Caterpillar on it's called solar turbines. And 144 00:07:08,080 --> 00:07:10,080 Speaker 5: the reason why they're not you would think, oh, you're 145 00:07:10,080 --> 00:07:12,240 Speaker 5: spending half a trillion dollars on these things. You could 146 00:07:12,240 --> 00:07:15,360 Speaker 5: probably get the best thing ever, but that'll take you 147 00:07:15,400 --> 00:07:15,960 Speaker 5: seven years. 148 00:07:16,160 --> 00:07:18,680 Speaker 4: Oh right, because there's a there's a natural guest turbine. 149 00:07:18,800 --> 00:07:19,760 Speaker 3: Yeah, shortage. 150 00:07:19,920 --> 00:07:22,640 Speaker 4: By the way, there's a headline. 151 00:07:22,880 --> 00:07:24,040 Speaker 3: So there's a headline. 152 00:07:24,200 --> 00:07:27,360 Speaker 4: An article on the Bloomberg Today by Matthew Griffin, the 153 00:07:27,480 --> 00:07:30,080 Speaker 4: hunt for winners in the artificial intelligence gold rush has 154 00:07:30,160 --> 00:07:33,800 Speaker 4: landed on an unlikely target, old line industrial equipment maker 155 00:07:33,840 --> 00:07:36,400 Speaker 4: Caterpillar in So you got to look at every one 156 00:07:36,440 --> 00:07:40,960 Speaker 4: of these brands that's supplying it, and Caterpillar is up 157 00:07:41,000 --> 00:07:44,440 Speaker 4: a lot. It's up today and the stock is done 158 00:07:44,800 --> 00:07:48,040 Speaker 4: very well. So okay, Caterpillar, who else? What are some 159 00:07:48,080 --> 00:07:50,400 Speaker 4: of the what else do you see besides these big 160 00:07:50,440 --> 00:07:53,280 Speaker 4: turbines when you go out and tour this facility. So 161 00:07:53,960 --> 00:07:54,960 Speaker 4: the stock is a rocket. 162 00:07:54,960 --> 00:07:58,040 Speaker 2: So that's any people like securities are. 163 00:07:57,920 --> 00:08:02,760 Speaker 5: So many people and I mean it's there. I don't 164 00:08:02,760 --> 00:08:05,360 Speaker 5: know you've seen like Silicon Valley and there's that scene 165 00:08:05,360 --> 00:08:07,440 Speaker 5: where there's like one guy in the data center and 166 00:08:07,480 --> 00:08:09,760 Speaker 5: he's like, yeah, you know, night and day don't really 167 00:08:09,760 --> 00:08:12,000 Speaker 5: matter much down here, and it's just one guy running. 168 00:08:12,200 --> 00:08:15,280 Speaker 5: But in the buildout, I mean build up, there are 169 00:08:16,520 --> 00:08:20,160 Speaker 5: every single HVAC technician or guy that knows how to 170 00:08:20,240 --> 00:08:26,400 Speaker 5: lay fiber or guy that knows electrical or plumbing within 171 00:08:26,440 --> 00:08:28,720 Speaker 5: one hundred mile radius of valley and those are they're 172 00:08:28,760 --> 00:08:32,480 Speaker 5: all there, right, There's seven thousand people working on building 173 00:08:32,520 --> 00:08:33,960 Speaker 5: this and this is just one site. 174 00:08:34,040 --> 00:08:37,679 Speaker 2: And yeah, so people have been talking about the picks 175 00:08:37,679 --> 00:08:42,320 Speaker 2: and shovels phase of AI investment for a while now, 176 00:08:42,360 --> 00:08:44,920 Speaker 2: but it seems like stuff is still getting built and 177 00:08:44,960 --> 00:08:46,079 Speaker 2: it has further to run. 178 00:08:46,520 --> 00:08:49,280 Speaker 5: My kind of favorite we can you know, and I'm 179 00:08:49,320 --> 00:08:51,360 Speaker 5: sure we will discuss whether this is a bubble or not, 180 00:08:51,520 --> 00:08:53,880 Speaker 5: but at the end of the day, the money is 181 00:08:53,920 --> 00:08:58,880 Speaker 5: getting spent. And my favorite kind of setup in the 182 00:08:58,920 --> 00:09:02,360 Speaker 5: market is you have something that has a multiple that 183 00:09:02,400 --> 00:09:05,080 Speaker 5: reflects kind of a wrong reality or doesn't really fully 184 00:09:05,400 --> 00:09:10,840 Speaker 5: integrate secular theme and then you have a cycle that's 185 00:09:10,880 --> 00:09:15,880 Speaker 5: really bad. So whether it's kind of the robotics supply 186 00:09:15,960 --> 00:09:18,160 Speaker 5: chain goes throughout the automotive cycle and if you guys 187 00:09:18,160 --> 00:09:21,360 Speaker 5: are aware, that's been a nightmare. And then you have 188 00:09:21,679 --> 00:09:24,080 Speaker 5: some of these A lot of these companies took really 189 00:09:24,120 --> 00:09:29,960 Speaker 5: significant drawdowns when Trump got elected. Because this isn't really 190 00:09:30,080 --> 00:09:33,760 Speaker 5: tech Capex, it's much more similar to like LNG terminal cabets. Ye, 191 00:09:34,120 --> 00:09:37,840 Speaker 5: that kind of entails the same companies that were benefiting 192 00:09:37,880 --> 00:09:41,240 Speaker 5: from the IRA oh, and so everyone sold them. And 193 00:09:41,280 --> 00:09:44,080 Speaker 5: now it turns out it's kind of like the horseshoe theory, 194 00:09:44,160 --> 00:09:44,520 Speaker 5: right you. 195 00:09:44,440 --> 00:09:48,040 Speaker 4: Can now that's super interesting. Like it's just these sort 196 00:09:48,080 --> 00:09:51,280 Speaker 4: of like oldline industrials, and they did very well under 197 00:09:51,320 --> 00:09:54,120 Speaker 4: the Obama administration because of all the factories and all 198 00:09:54,160 --> 00:09:56,319 Speaker 4: that stuff that we talked about. And then it's like, okay, 199 00:09:56,320 --> 00:09:58,560 Speaker 4: we're pulling the plug on this, but now there's just 200 00:09:58,640 --> 00:10:02,520 Speaker 4: this title wave money. By the way, Caterpillar was a 201 00:10:02,600 --> 00:10:04,920 Speaker 4: three hundred and thirty four dollars stock on April second, 202 00:10:05,080 --> 00:10:06,760 Speaker 4: right before libration day, which is sort of where we 203 00:10:06,760 --> 00:10:09,800 Speaker 4: should benchmark things now, four hundred and eighty six at 204 00:10:09,800 --> 00:10:10,480 Speaker 4: an auld time high. 205 00:10:10,559 --> 00:10:13,560 Speaker 3: What are some of the other companies see for real? 206 00:10:13,640 --> 00:10:15,720 Speaker 5: Like preferably the ones that haven't gone up? Yeah? 207 00:10:15,800 --> 00:10:19,280 Speaker 4: Yeah, seriously, Like who else is who else is getting 208 00:10:19,280 --> 00:10:20,280 Speaker 4: a piece of this action. 209 00:10:20,920 --> 00:10:23,920 Speaker 5: So something that I want to make sure I mentioned 210 00:10:23,960 --> 00:10:26,320 Speaker 5: just to frame this whole thing. If you look at 211 00:10:26,320 --> 00:10:29,040 Speaker 5: the building side plans, they have the names of all 212 00:10:29,080 --> 00:10:33,040 Speaker 5: the data halls and they are literally each called ludicrous building. 213 00:10:33,160 --> 00:10:35,240 Speaker 5: So it's ludicrous building, one ludicrous building. 214 00:10:35,320 --> 00:10:35,720 Speaker 4: Amazing. 215 00:10:35,840 --> 00:10:37,880 Speaker 2: Wait is it Ludicris with a K? 216 00:10:38,720 --> 00:10:39,160 Speaker 5: No? 217 00:10:39,679 --> 00:10:43,960 Speaker 2: That would be great. I mean Ludacris spelled correctly is 218 00:10:44,000 --> 00:10:44,920 Speaker 2: also fantastic. 219 00:10:45,000 --> 00:10:50,880 Speaker 5: But the company, I mean really I would encourage anyone 220 00:10:50,920 --> 00:10:53,680 Speaker 5: to go and look at this video because the it's 221 00:10:54,120 --> 00:10:58,800 Speaker 5: it's just a there is such a disconnect between looking 222 00:10:58,840 --> 00:11:03,079 Speaker 5: at green numbers going up like like caterpillar going okay, cool, Yeah, 223 00:11:03,120 --> 00:11:06,160 Speaker 5: and then you see it and it is mind blowing. 224 00:11:06,240 --> 00:11:11,000 Speaker 5: I mean they have rows of dry and atabiatic coolers 225 00:11:11,120 --> 00:11:13,960 Speaker 5: that are you know, hundreds and hundreds for the liquid 226 00:11:13,960 --> 00:11:18,320 Speaker 5: cooling loops, and then you have the kind of all 227 00:11:18,400 --> 00:11:22,560 Speaker 5: the HVAC, the power generation and transmission. We compile the 228 00:11:22,600 --> 00:11:26,480 Speaker 5: list of basically two hundred companies, fifty of which were 229 00:11:26,520 --> 00:11:29,280 Speaker 5: high confidence in kind of seeing as part of this 230 00:11:29,360 --> 00:11:32,480 Speaker 5: build out. And the interesting thing is you have these 231 00:11:32,559 --> 00:11:37,840 Speaker 5: multinational OEMs. This scale of this project is so large 232 00:11:38,240 --> 00:11:40,680 Speaker 5: that you can call one of them and say, okay, 233 00:11:40,720 --> 00:11:43,920 Speaker 5: we want to go with you know, Mitsubishi heav or 234 00:11:43,960 --> 00:11:47,679 Speaker 5: we want to use Eton for this thing, and they 235 00:11:47,679 --> 00:11:49,720 Speaker 5: will say, yeah, okay, well we can do sixty percent 236 00:11:49,760 --> 00:11:53,959 Speaker 5: of it. I mean, the biggest companies in the world 237 00:11:54,320 --> 00:11:56,600 Speaker 5: are Yeah, they're just fighting. They don't have to fight 238 00:11:56,600 --> 00:11:57,840 Speaker 5: with each other because the pie is so big. 239 00:11:58,200 --> 00:12:01,880 Speaker 3: How scarce you know, there's all this around the country. 240 00:12:02,120 --> 00:12:04,800 Speaker 4: There's been all this tensions like is there gonna be 241 00:12:04,880 --> 00:12:06,720 Speaker 4: enough energy for data centers? 242 00:12:07,120 --> 00:12:10,000 Speaker 3: What about the water? What about the politics? People? 243 00:12:10,120 --> 00:12:13,319 Speaker 4: We see these town halls where the locals do want 244 00:12:13,320 --> 00:12:17,400 Speaker 4: a data center. How scarce are locations like Abilene? I 245 00:12:17,400 --> 00:12:19,760 Speaker 4: mean there's a lot of empty space in Texas. Could 246 00:12:19,840 --> 00:12:22,960 Speaker 4: you see Are there more stargates or more Abilenees out 247 00:12:23,000 --> 00:12:25,600 Speaker 4: there waiting where the companies can just say, you know what, 248 00:12:25,880 --> 00:12:27,680 Speaker 4: we're just going to build our own infrastructure. We're going 249 00:12:27,720 --> 00:12:29,600 Speaker 4: to build our own natural gas plant, and we're just 250 00:12:29,640 --> 00:12:31,160 Speaker 4: going to do it in the desert in Texas where 251 00:12:31,200 --> 00:12:33,400 Speaker 4: we don't have to worry about nimbi's or local politics 252 00:12:33,440 --> 00:12:33,880 Speaker 4: or anything. 253 00:12:34,040 --> 00:12:35,680 Speaker 3: Is that the future for a lot of this stuff? 254 00:12:35,800 --> 00:12:40,120 Speaker 5: Yeah, there's already another facility that's planned in Shackleford, Texas, 255 00:12:40,120 --> 00:12:42,680 Speaker 5: which is not too far from Abilene. Abilene is already 256 00:12:42,679 --> 00:12:45,160 Speaker 5: planning on expanding by another six hundred megawats. I mean, 257 00:12:45,200 --> 00:12:47,760 Speaker 5: that's how we're measuring data centers now, is just in 258 00:12:47,840 --> 00:12:53,360 Speaker 5: megawatts and gigawatts. In Louisiana, that's where Meta is building Hyperion. 259 00:12:53,640 --> 00:12:59,360 Speaker 5: You have some unidentified or unannounced Midwest facility that'll probably 260 00:12:59,400 --> 00:13:03,160 Speaker 5: be in Ohi, Io. Basically, yeah, anywhere that you can 261 00:13:03,200 --> 00:13:06,040 Speaker 5: get the gas, anywhere you can get the power, anywhere 262 00:13:06,040 --> 00:13:10,920 Speaker 5: that there's the land. There was one planned in Indianapolis. 263 00:13:11,280 --> 00:13:13,240 Speaker 5: And it's interesting you bring up water because we wrote 264 00:13:13,240 --> 00:13:15,880 Speaker 5: about water a little while ago as well, and we 265 00:13:15,880 --> 00:13:18,400 Speaker 5: were basing our projections off of what the water consumption 266 00:13:18,520 --> 00:13:21,600 Speaker 5: of like existing data centers were. The advancements in liquid 267 00:13:21,600 --> 00:13:23,800 Speaker 5: cooling have changed that a lot, so there was a 268 00:13:23,880 --> 00:13:29,439 Speaker 5: huge kind of uproar about, well, this data center is 269 00:13:29,440 --> 00:13:32,079 Speaker 5: going to use so much water. And if you look 270 00:13:32,080 --> 00:13:33,959 Speaker 5: at it, this was an Indianapolis and if you look 271 00:13:33,960 --> 00:13:36,160 Speaker 5: at it, it's about the same amount of water as 272 00:13:36,160 --> 00:13:39,480 Speaker 5: a thousand households. Because it's all closed loop liquid cooling. Now, 273 00:13:39,679 --> 00:13:44,720 Speaker 5: the water just gets recycled and filtered back, so it's 274 00:13:44,800 --> 00:13:49,920 Speaker 5: really just something where the constraint is still power and 275 00:13:49,960 --> 00:13:53,280 Speaker 5: getting these kind of turbines, and it's going to increasingly 276 00:13:53,280 --> 00:13:55,360 Speaker 5: become a thing where it's behind the meter power generation. 277 00:13:55,840 --> 00:13:58,120 Speaker 5: They're just going to build their own power plants. They're 278 00:13:58,160 --> 00:13:59,880 Speaker 5: not going to rely on the grid. You guys remember 279 00:14:00,000 --> 00:14:02,840 Speaker 5: what happened in Texas and twenty one with a Yeah, 280 00:14:03,160 --> 00:14:06,640 Speaker 5: they you cannot underwrite that as a data center, like 281 00:14:08,160 --> 00:14:12,560 Speaker 5: you just can't be offline. So that's kind of where 282 00:14:12,600 --> 00:14:15,040 Speaker 5: we're headed, is that this power and energy kind of 283 00:14:15,800 --> 00:14:19,760 Speaker 5: bottleneck is the bottleneck. One of the more interesting things 284 00:14:19,760 --> 00:14:23,040 Speaker 5: we ran across is you look at Xai's data center, 285 00:14:23,080 --> 00:14:24,600 Speaker 5: and we went through a lot of the filings of 286 00:14:24,680 --> 00:14:28,440 Speaker 5: all these different data centers. They're permitted for fifteen thirty 287 00:14:28,480 --> 00:14:32,960 Speaker 5: five megawatt turbines, and there's thermal imaging of the site 288 00:14:33,000 --> 00:14:36,560 Speaker 5: that shows they're using thirty five ge Vanova. 289 00:14:36,560 --> 00:14:37,640 Speaker 3: Another one that was a three. 290 00:14:39,120 --> 00:14:40,320 Speaker 2: You're just typing tickers. 291 00:14:40,360 --> 00:14:42,360 Speaker 4: Yeah, there was a three hundred and thirty dollars stock 292 00:14:42,440 --> 00:14:44,920 Speaker 4: on April second. That's a six hundred and nine dollars stocks. 293 00:14:44,920 --> 00:14:46,360 Speaker 5: Okay, check out Stemens Energy. 294 00:14:46,440 --> 00:14:48,680 Speaker 2: Thank you, Joe. Oh he's doing it. 295 00:14:49,080 --> 00:14:50,320 Speaker 3: You could keep going. 296 00:14:50,400 --> 00:14:54,040 Speaker 2: All right, I have kind of this isn't an essential question, 297 00:14:54,120 --> 00:14:56,200 Speaker 2: but okay, wait, I got it. 298 00:14:56,480 --> 00:14:57,720 Speaker 5: I knew this is going to happen. 299 00:14:58,280 --> 00:15:01,520 Speaker 4: Semens Energy was a fifty six dollars stock on April second. 300 00:15:01,600 --> 00:15:02,360 Speaker 5: That's one hundred and eight. 301 00:15:02,360 --> 00:15:03,640 Speaker 4: That's a string lineup. 302 00:15:03,720 --> 00:15:23,240 Speaker 2: Okay, okay, clearly someone's making money. Like how big are 303 00:15:23,240 --> 00:15:26,120 Speaker 2: the walls of these things? And what does security actually 304 00:15:26,120 --> 00:15:28,920 Speaker 2: look like? Because you know, we talk about people being 305 00:15:29,040 --> 00:15:32,400 Speaker 2: upset about higher energy prices. I understand that maybe some 306 00:15:32,480 --> 00:15:34,720 Speaker 2: of the some of them are building out their own 307 00:15:34,840 --> 00:15:38,440 Speaker 2: energy capacity, but you know, in a populist mindset, people 308 00:15:38,480 --> 00:15:41,800 Speaker 2: might still be upset at AI, still upset that AI 309 00:15:42,000 --> 00:15:44,520 Speaker 2: is going to take jobs and things like that. They 310 00:15:44,520 --> 00:15:46,040 Speaker 2: could become targets. 311 00:15:45,640 --> 00:15:49,680 Speaker 5: Right, yeah, absolutely, I mean apparently security isn't tight enough 312 00:15:49,720 --> 00:15:52,880 Speaker 5: to shoot down at drone yet, oh of course. Yeah. 313 00:15:53,000 --> 00:15:55,280 Speaker 5: But when you talk about the walls, like, the interesting 314 00:15:55,320 --> 00:15:58,000 Speaker 5: thing is everything in this data center has You can't 315 00:15:58,080 --> 00:16:00,720 Speaker 5: just build it like a house. Everything needs to have 316 00:16:00,760 --> 00:16:05,480 Speaker 5: the right anti corrosion insulation. You can't just be losing heat. 317 00:16:05,480 --> 00:16:08,360 Speaker 5: You're already spending so much money on cooling. So everything 318 00:16:08,360 --> 00:16:10,920 Speaker 5: in the building envelope is also kind of the most 319 00:16:10,960 --> 00:16:12,080 Speaker 5: expensive thing it could be. 320 00:16:12,480 --> 00:16:14,560 Speaker 2: And I assume it's like a sealed envelope. 321 00:16:14,600 --> 00:16:22,840 Speaker 5: I guess, yeah, exactly, And the opportunity of just how 322 00:16:22,920 --> 00:16:25,000 Speaker 5: big is this going to be if you actually believe that. 323 00:16:25,240 --> 00:16:28,240 Speaker 5: The people that are building this, and this is kind 324 00:16:28,280 --> 00:16:31,800 Speaker 5: of an interesting dynamic of the hyperscalars are in somewhat 325 00:16:31,800 --> 00:16:34,200 Speaker 5: of a prisoner's dilemma. We can't speak to whether they 326 00:16:34,240 --> 00:16:38,640 Speaker 5: actually believe or not, but the party line is we're 327 00:16:38,640 --> 00:16:42,160 Speaker 5: building machine God. And if we're building machine God, then 328 00:16:42,320 --> 00:16:45,720 Speaker 5: that's kind of a religious zelotry. So three years ago, 329 00:16:45,800 --> 00:16:50,080 Speaker 5: most of these hyperscalers were making net zero. By twenty thirty, 330 00:16:50,280 --> 00:16:53,040 Speaker 5: that is not happening. Yeah, there's no way that's happening. 331 00:16:53,520 --> 00:16:56,400 Speaker 5: Maybe they buy carbon off, but that's not happening because 332 00:16:56,760 --> 00:16:58,840 Speaker 5: if you need to do it now, you're using natural 333 00:16:58,880 --> 00:17:03,840 Speaker 5: gas and that then you say, well, once we create 334 00:17:04,119 --> 00:17:07,679 Speaker 5: machine God, machine God will fix it, right, and you 335 00:17:07,680 --> 00:17:09,240 Speaker 5: can justify a lot of stuff. 336 00:17:11,600 --> 00:17:14,280 Speaker 2: Wait, okay, so let's talk about is AI in a bubble, 337 00:17:14,280 --> 00:17:18,520 Speaker 2: because when we're conversing about buildings that are literally called 338 00:17:18,640 --> 00:17:22,080 Speaker 2: ludacris and we're talking about machine God will solve all 339 00:17:22,119 --> 00:17:23,439 Speaker 2: our problems. 340 00:17:23,119 --> 00:17:24,080 Speaker 5: That sounds a little bubbly. 341 00:17:24,200 --> 00:17:26,080 Speaker 2: Yeah, that's the kind of stuff you look back on 342 00:17:26,240 --> 00:17:29,240 Speaker 2: and say, oh, yeah, that was the top I know. 343 00:17:29,359 --> 00:17:31,359 Speaker 5: That we we have to talk about this and that 344 00:17:31,440 --> 00:17:35,240 Speaker 5: requires some numbers, so I have say notes. But the 345 00:17:35,440 --> 00:17:39,320 Speaker 5: interesting I mean, you can make a case for either side, 346 00:17:39,560 --> 00:17:43,280 Speaker 5: and it kind of hinges on not so much what 347 00:17:43,320 --> 00:17:45,760 Speaker 5: AI looks like in the future, but more where we 348 00:17:45,800 --> 00:17:48,520 Speaker 5: are in the cycle. If the question is is AI bubble? Yeah, probably, 349 00:17:48,880 --> 00:17:51,640 Speaker 5: it's a new technology. It's exciting every single time we've 350 00:17:51,840 --> 00:17:56,200 Speaker 5: the railroads, is the you know, dot com? Probably if 351 00:17:56,200 --> 00:17:58,240 Speaker 5: it's not a bubble, it will be, and it probably 352 00:17:58,320 --> 00:18:00,720 Speaker 5: is already. But where are are we in the bubble? 353 00:18:00,760 --> 00:18:05,360 Speaker 5: That's the interesting thing for investors. So if you look 354 00:18:05,400 --> 00:18:11,160 Speaker 5: at kind of where are we in the financing side. 355 00:18:11,520 --> 00:18:13,640 Speaker 5: When we had the dot com bubble, the big thing. 356 00:18:14,000 --> 00:18:17,920 Speaker 5: Are you guys familiar with dark fiber? No? So essentially 357 00:18:18,320 --> 00:18:21,520 Speaker 5: they were building out this kind of infrastructure for the Internet, 358 00:18:21,920 --> 00:18:25,159 Speaker 5: and the conviction of the day was the Internet's going 359 00:18:25,200 --> 00:18:27,400 Speaker 5: to be the biggest thing in the world. Correct, And 360 00:18:28,560 --> 00:18:30,720 Speaker 5: ninety seven percent of the fiber that was laid was 361 00:18:30,760 --> 00:18:33,760 Speaker 5: called dark fiber because it wasn't lit up. It was 362 00:18:33,920 --> 00:18:36,439 Speaker 5: just laying there because eventually we're going to need it. 363 00:18:37,400 --> 00:18:40,000 Speaker 5: That's not the case right now. Every single time that 364 00:18:40,040 --> 00:18:42,520 Speaker 5: one of these data centers comes online, you're at one 365 00:18:42,600 --> 00:18:47,040 Speaker 5: hundred percent utilization. There's no dark fiber, and we're continuing 366 00:18:47,040 --> 00:18:48,280 Speaker 5: to need more. I mean, I don't know if you 367 00:18:48,280 --> 00:18:50,760 Speaker 5: guys saw Sora too. More than seventy percent of the 368 00:18:50,760 --> 00:18:54,600 Speaker 5: content that we consume is video, so a video model 369 00:18:54,800 --> 00:18:57,400 Speaker 5: makes a lot of sense. It's also way more compute intensive, 370 00:18:58,040 --> 00:19:02,280 Speaker 5: and the the final form of chat of AI is 371 00:19:02,280 --> 00:19:06,280 Speaker 5: not chatbots. It's gonna be video. It's gonna be robotics, 372 00:19:06,600 --> 00:19:10,359 Speaker 5: which use video models. So that's kind of like the 373 00:19:10,920 --> 00:19:15,520 Speaker 5: bullish side. But I totally see. I have some numbers here, 374 00:19:15,600 --> 00:19:22,560 Speaker 5: So we've had some of these circular deals and so far, 375 00:19:22,680 --> 00:19:26,120 Speaker 5: I mean, there's hundreds of billions of dollars for this. 376 00:19:26,240 --> 00:19:29,080 Speaker 5: Most of the financing is structured much more like toll 377 00:19:29,160 --> 00:19:32,720 Speaker 5: roads or LNG terminals. The Metas hyperion was twenty nine 378 00:19:32,720 --> 00:19:35,959 Speaker 5: billion dollars. Twenty six billion of that was Pimco on 379 00:19:36,040 --> 00:19:38,280 Speaker 5: debt and then blue outl with three billion of equity, 380 00:19:39,040 --> 00:19:41,959 Speaker 5: an off balance sheet arrangement, which also was pretty popular 381 00:19:42,040 --> 00:19:44,600 Speaker 5: during the dot com bubble. Then you have stuff like 382 00:19:44,720 --> 00:19:49,000 Speaker 5: GPU collateralized SPVs core. We've had a seven and a 383 00:19:49,000 --> 00:19:52,560 Speaker 5: half billion dollar Blackstone SPV. There's like an eleven billion 384 00:19:52,600 --> 00:19:57,000 Speaker 5: dollar GPU backed loan market now, which compared to the 385 00:19:57,480 --> 00:20:01,000 Speaker 5: infrastructure thing, I can get that. It's like those GPUs 386 00:20:01,000 --> 00:20:05,480 Speaker 5: are gonna be worth like they will depreciate. So yeah, that's. 387 00:20:05,600 --> 00:20:08,840 Speaker 2: That's abs right, that's like asset backsxety Okay. 388 00:20:08,760 --> 00:20:12,320 Speaker 5: Yeah, but when you look at the where we are 389 00:20:12,480 --> 00:20:15,840 Speaker 5: just saying where are we in the cycle, Well, there's 390 00:20:15,840 --> 00:20:18,120 Speaker 5: been there's been a lot of these balance sheet arrangements. 391 00:20:18,160 --> 00:20:22,200 Speaker 5: But in terms of that prisoner's dilemma that we spoke about, 392 00:20:22,200 --> 00:20:25,280 Speaker 5: where nobody's going to back off because then maybe they 393 00:20:25,359 --> 00:20:28,200 Speaker 5: risk not creating machine GUD. The debt to equity ratio 394 00:20:28,440 --> 00:20:31,840 Speaker 5: on most of the hyperscalars is not crazy. Sure Oracle 395 00:20:31,880 --> 00:20:35,080 Speaker 5: has the highest, and but they still have a lot 396 00:20:35,080 --> 00:20:38,160 Speaker 5: of room to lever up in pursuit of this goal. 397 00:20:38,760 --> 00:20:41,720 Speaker 5: The interesting kind of dynamic of in Vidia finances its 398 00:20:41,720 --> 00:20:45,080 Speaker 5: customers and then its customers buy Nvidia GPUs, and then 399 00:20:45,280 --> 00:20:48,520 Speaker 5: you know that, yeah, that's a bubble dynamic. That's that's 400 00:20:48,600 --> 00:20:52,359 Speaker 5: visible circular financing, and but the demand is real. 401 00:20:52,560 --> 00:20:54,480 Speaker 4: Well, this is this is I guess the question. If 402 00:20:54,520 --> 00:20:58,280 Speaker 4: the demand is real, why are the videos of the 403 00:20:58,320 --> 00:21:02,560 Speaker 4: world financing your customers? Right, intuitively you think, okay, in 404 00:21:02,680 --> 00:21:06,560 Speaker 4: videos the demand for chips is maxed out, right, Then 405 00:21:06,640 --> 00:21:09,040 Speaker 4: intuitively you think, well, then there's. 406 00:21:08,800 --> 00:21:10,040 Speaker 3: Plenty of money elsewhere. 407 00:21:10,119 --> 00:21:13,120 Speaker 4: So do you have a guess for why in VideA 408 00:21:13,280 --> 00:21:16,359 Speaker 4: is compelled to you know, invest in open Ai, invest 409 00:21:16,400 --> 00:21:18,280 Speaker 4: in core weave. It's been a long time investor in 410 00:21:18,320 --> 00:21:19,359 Speaker 4: core weave and so forth. 411 00:21:19,560 --> 00:21:22,440 Speaker 5: Yeah, I could. I could take the very skeptical view 412 00:21:22,520 --> 00:21:25,080 Speaker 5: or the kind of non skeptical view. The non skeptical 413 00:21:25,160 --> 00:21:27,560 Speaker 5: view is the best thing for in Vidia is that 414 00:21:27,600 --> 00:21:30,119 Speaker 5: we accomplish a GI. So anything that it can do 415 00:21:30,280 --> 00:21:33,680 Speaker 5: to get us closer to that massive, massive infrastructure that's 416 00:21:33,720 --> 00:21:36,000 Speaker 5: required for that is great for them as quickly able, 417 00:21:36,000 --> 00:21:39,359 Speaker 5: as quickly as possible, because, uh, every year that you 418 00:21:39,560 --> 00:21:44,479 Speaker 5: don't achieve AGI becomes less likely. So uh, that is 419 00:21:44,920 --> 00:21:48,919 Speaker 5: maybe the core factor. And then there's it's good for 420 00:21:49,280 --> 00:21:55,199 Speaker 5: them and the and playing into that the because this 421 00:21:55,200 --> 00:21:58,080 Speaker 5: this isn't necessarily like the dot com bubble, because the 422 00:21:58,119 --> 00:22:00,960 Speaker 5: dot com bubble had fiber and then it had you know, pets, 423 00:22:01,000 --> 00:22:03,679 Speaker 5: dot Com and Amazon and all that stuff. This is 424 00:22:03,720 --> 00:22:07,040 Speaker 5: all pretty much capex right to tech is capital intensive again, 425 00:22:07,560 --> 00:22:10,879 Speaker 5: which means that it's the bust won't necessarily be like 426 00:22:10,920 --> 00:22:13,240 Speaker 5: the dot com bubble, where it gives the real players 427 00:22:13,280 --> 00:22:16,240 Speaker 5: time to shine. If the CAPEX spending grinds to a 428 00:22:16,280 --> 00:22:20,000 Speaker 5: halt because the market goes down, that's the worst thing 429 00:22:20,000 --> 00:22:23,000 Speaker 5: in the world. So in a way, it's also it's 430 00:22:23,040 --> 00:22:25,720 Speaker 5: in their interest obviously to make sure that that doesn't happen. 431 00:22:41,600 --> 00:22:43,879 Speaker 4: Another question, actually, but I want to ask it now 432 00:22:43,880 --> 00:22:46,840 Speaker 4: because I might forget otherwise. A headline I think earlier 433 00:22:46,880 --> 00:22:49,720 Speaker 4: this week again headline is almost every day. Can you 434 00:22:49,800 --> 00:22:54,560 Speaker 4: explain why Meta, which is building out which obviously hyperscaler, 435 00:22:54,680 --> 00:22:57,680 Speaker 4: they have tons of data center capacity, why do they 436 00:22:57,760 --> 00:22:59,800 Speaker 4: need to do a deal with Corewave? So I saw 437 00:22:59,840 --> 00:23:04,040 Speaker 4: that Meta is going to pay a few billion, fourteen billion, 438 00:23:04,040 --> 00:23:07,080 Speaker 4: it's some number for capacity on core Weave. What can 439 00:23:07,119 --> 00:23:09,280 Speaker 4: the what do the neo cloudes bring to the table 440 00:23:09,359 --> 00:23:11,080 Speaker 4: such that the legacy is like, you know what, we 441 00:23:11,520 --> 00:23:12,639 Speaker 4: launched some of your capacity. 442 00:23:13,680 --> 00:23:18,399 Speaker 5: It's basically I would say, anything you can do to 443 00:23:18,480 --> 00:23:21,800 Speaker 5: shift that CAPEX away from yourself is great. Right if 444 00:23:22,119 --> 00:23:24,320 Speaker 5: you're if you're right now, a lot of this stuff 445 00:23:24,359 --> 00:23:26,280 Speaker 5: is funded from cash flow, and yeah, we're starting to 446 00:23:26,320 --> 00:23:31,600 Speaker 5: see that financing take its place. But if you the 447 00:23:31,600 --> 00:23:33,760 Speaker 5: the you're running a company that's existed for a while 448 00:23:33,760 --> 00:23:35,320 Speaker 5: and is one of the biggest companies in the world, 449 00:23:35,480 --> 00:23:38,560 Speaker 5: and yes, you could justify taking an ordinate amounts of 450 00:23:38,640 --> 00:23:41,879 Speaker 5: risk to do this, But just like anyone else, if 451 00:23:41,920 --> 00:23:43,520 Speaker 5: you have a goal and you can kind of shift 452 00:23:43,560 --> 00:23:46,359 Speaker 5: away some of that risk to because if you know, 453 00:23:46,440 --> 00:23:49,560 Speaker 5: if core Weeve goes bankrupt, Meta is not it's not 454 00:23:49,600 --> 00:23:51,280 Speaker 5: It's not the worst thing in the world for Meta. 455 00:23:51,520 --> 00:23:54,000 Speaker 2: What would you be looking out for for signs of 456 00:23:54,040 --> 00:23:58,280 Speaker 2: the bubble bursting or I guess the capex FAWCET starting 457 00:23:58,320 --> 00:23:58,920 Speaker 2: to turn off. 458 00:23:59,800 --> 00:24:06,440 Speaker 5: So in terms of like the warning signs first, monitoring 459 00:24:06,760 --> 00:24:09,000 Speaker 5: used GPU prices I think is a great way to 460 00:24:09,320 --> 00:24:13,600 Speaker 5: see what's going on, seeing like contract renegotiations, waiver headlines 461 00:24:13,600 --> 00:24:17,240 Speaker 5: from private credit. And also a big one would be 462 00:24:17,280 --> 00:24:20,679 Speaker 5: any delays in delivering power make it so that you 463 00:24:20,760 --> 00:24:22,520 Speaker 5: finance this thing and then it's not going to be 464 00:24:22,560 --> 00:24:25,960 Speaker 5: able to work, right. I do think that another you 465 00:24:26,080 --> 00:24:30,120 Speaker 5: really do have to kind of index to the revenue 466 00:24:30,119 --> 00:24:32,640 Speaker 5: that's being generated by the AI companies as well, and 467 00:24:33,280 --> 00:24:36,639 Speaker 5: you saw open Ai recently announced that they're basically allowing 468 00:24:36,680 --> 00:24:39,399 Speaker 5: you to buy things with one click from the inside 469 00:24:39,440 --> 00:24:42,080 Speaker 5: the app. And I think that that was also a 470 00:24:42,119 --> 00:24:45,919 Speaker 5: big driving factor in the router where they kind of 471 00:24:45,960 --> 00:24:49,320 Speaker 5: route your model to the it's setting it up for advertising, 472 00:24:49,320 --> 00:24:51,720 Speaker 5: because the other day I asked Google a pretty simple 473 00:24:51,800 --> 00:24:54,399 Speaker 5: question and it just gave me the answer with no 474 00:24:54,520 --> 00:24:58,680 Speaker 5: advertisements delivered to me. So we need to I think 475 00:24:58,720 --> 00:25:01,639 Speaker 5: that there is the killer apps kind of are here 476 00:25:01,760 --> 00:25:05,119 Speaker 5: or coming, but that has to happen fast. 477 00:25:05,440 --> 00:25:08,639 Speaker 2: Yeah, the fact that the Internet kind of runs on 478 00:25:08,760 --> 00:25:12,960 Speaker 2: ads is an underappreciated reality of our world, and we 479 00:25:13,000 --> 00:25:15,560 Speaker 2: did an episode on it way way back in the day. 480 00:25:15,960 --> 00:25:18,800 Speaker 2: Is there anyway if I wanted to get exposure to 481 00:25:19,400 --> 00:25:22,680 Speaker 2: like the good stuff of AI, the actual money being spent, 482 00:25:23,080 --> 00:25:26,199 Speaker 2: but then I don't like all the circular financing arrangements. 483 00:25:26,400 --> 00:25:28,720 Speaker 2: Is there a way to hedge that particular risk? 484 00:25:29,280 --> 00:25:33,280 Speaker 5: Yeah, you could buy all the real companies that are 485 00:25:33,280 --> 00:25:36,320 Speaker 5: building this kind of data center, and then you could 486 00:25:36,320 --> 00:25:39,360 Speaker 5: short Blue Owl and the private credit companies that are 487 00:25:39,400 --> 00:25:42,399 Speaker 5: doing the kind of financing aspect of it, because if 488 00:25:42,440 --> 00:25:45,240 Speaker 5: it goes bad, they have limited upside and kind of 489 00:25:45,359 --> 00:25:49,040 Speaker 5: unlimited downside, and if it continues happening, the companies that 490 00:25:49,080 --> 00:25:52,359 Speaker 5: are actually building it have unlimited upside and they're still, 491 00:25:52,480 --> 00:25:55,520 Speaker 5: like you said, these oldline kind of industrial that like, 492 00:25:55,560 --> 00:25:57,159 Speaker 5: they still have a core business to go back to. 493 00:25:57,720 --> 00:26:02,360 Speaker 4: What's your take on so open ai announced its new 494 00:26:02,440 --> 00:26:06,080 Speaker 4: Sora thing that Tracy has a code for, Meta announced 495 00:26:06,080 --> 00:26:08,080 Speaker 4: it's Meta AI, which a lot of people is like, 496 00:26:08,119 --> 00:26:09,600 Speaker 4: this is slop and I don't we don't need to 497 00:26:09,600 --> 00:26:12,240 Speaker 4: get into a debate about whether this is garbage. I 498 00:26:12,280 --> 00:26:14,000 Speaker 4: kind of think it is, but that's not actually the 499 00:26:14,080 --> 00:26:16,520 Speaker 4: question I'm interested in. The question I'm interested in it 500 00:26:16,560 --> 00:26:19,600 Speaker 4: is one way to view these things is you know, 501 00:26:19,640 --> 00:26:22,000 Speaker 4: what these are going to be, These video, these AI 502 00:26:22,080 --> 00:26:24,200 Speaker 4: videos are going to be really important sort of business 503 00:26:24,320 --> 00:26:28,320 Speaker 4: revenue drivers in our pursuit to AGI. And the other 504 00:26:28,480 --> 00:26:31,359 Speaker 4: view is, look, if they were close to AGI, they 505 00:26:31,359 --> 00:26:34,600 Speaker 4: would not be devoting time to the slop factories. And 506 00:26:34,640 --> 00:26:37,000 Speaker 4: this is the sort of question and I'm curious from 507 00:26:37,119 --> 00:26:39,880 Speaker 4: your perspective, what do you think is the real logic 508 00:26:40,160 --> 00:26:42,280 Speaker 4: behind all of this effort to create these sort of 509 00:26:42,440 --> 00:26:45,199 Speaker 4: you know, dom like, oh, a video of a astronaut 510 00:26:45,320 --> 00:26:48,320 Speaker 4: riding a horse through space with a you know, cowboy hat. 511 00:26:48,200 --> 00:26:50,399 Speaker 5: On or whatever. Video models are going to become the 512 00:26:50,440 --> 00:26:55,400 Speaker 5: most important part of this. You you need, for example, 513 00:26:55,720 --> 00:26:58,840 Speaker 5: lms don't really interact with the world, but we're starting 514 00:26:58,840 --> 00:27:02,000 Speaker 5: to see in robotics, AI models interact with the world. 515 00:27:02,280 --> 00:27:05,800 Speaker 5: So you have these video language action models, and you 516 00:27:05,840 --> 00:27:08,600 Speaker 5: can structure it so that there was an interesting paper 517 00:27:08,600 --> 00:27:13,040 Speaker 5: on Archive recently that and you know this is video 518 00:27:13,119 --> 00:27:17,840 Speaker 5: models like VO three are emergent zero shot learners and understanders. 519 00:27:17,880 --> 00:27:20,040 Speaker 5: And what that just means is you can give it 520 00:27:20,560 --> 00:27:23,280 Speaker 5: basically like a second of context, and then it can 521 00:27:23,320 --> 00:27:27,399 Speaker 5: extrapolate about what's going on beyond that. That becomes extremely 522 00:27:27,440 --> 00:27:31,040 Speaker 5: important for robotics. And you can structure these vision language 523 00:27:31,040 --> 00:27:33,840 Speaker 5: action models where you have like a slow reasoning model 524 00:27:34,200 --> 00:27:36,840 Speaker 5: that is kind of like your conscious mind and determines 525 00:27:36,880 --> 00:27:38,320 Speaker 5: what it needs to do, and then you have like 526 00:27:38,400 --> 00:27:40,840 Speaker 5: a fast model that actually does the movements and stuff. 527 00:27:40,960 --> 00:27:45,440 Speaker 5: AI kind of extends into the realm of you need 528 00:27:45,480 --> 00:27:48,240 Speaker 5: to generate as much data as possible, and having you know, 529 00:27:48,480 --> 00:27:50,840 Speaker 5: video data and knowing what's good and what's bad is 530 00:27:50,880 --> 00:27:53,840 Speaker 5: also very important. But yeah, to your question, open AI 531 00:27:54,800 --> 00:27:58,200 Speaker 5: probably does want to show that they're making more money, 532 00:27:58,600 --> 00:28:01,000 Speaker 5: and I don't doubt that they're people there that truly 533 00:28:01,040 --> 00:28:03,520 Speaker 5: believe that they're building the machine GUD. And I also 534 00:28:03,600 --> 00:28:06,200 Speaker 5: don't doubt that they realize if they have a shot 535 00:28:06,200 --> 00:28:07,880 Speaker 5: at doing that, they need the market to play long 536 00:28:07,960 --> 00:28:11,159 Speaker 5: and they need to keep growing their revenues. So I 537 00:28:11,160 --> 00:28:15,879 Speaker 5: think there's a lot of really good counter arguments about 538 00:28:15,920 --> 00:28:17,680 Speaker 5: like not necessarily are we in a bubble? 539 00:28:17,800 --> 00:28:17,960 Speaker 2: Yes? 540 00:28:18,880 --> 00:28:21,359 Speaker 5: Is the bubble over? Like? There are good arguments for that. 541 00:28:21,400 --> 00:28:23,880 Speaker 5: I don't think one of the I don't think that, oh, well, 542 00:28:23,880 --> 00:28:26,480 Speaker 5: they're leaning into like kind of this slops that you know, 543 00:28:26,520 --> 00:28:28,199 Speaker 5: it's a company. They're going to do whatever they can 544 00:28:28,240 --> 00:28:31,080 Speaker 5: to make money. That's capitalism, you know. 545 00:28:31,200 --> 00:28:33,680 Speaker 2: One of the things I think about another reality of 546 00:28:33,720 --> 00:28:36,760 Speaker 2: our current time. This is kind of related, but if 547 00:28:36,800 --> 00:28:41,080 Speaker 2: we're talking about video data, YouTube staying power has been 548 00:28:41,200 --> 00:28:44,800 Speaker 2: like phenomenal, right, It's been around for decades now, and 549 00:28:44,840 --> 00:28:49,320 Speaker 2: we haven't really seen any successful competitors anyway. 550 00:28:49,200 --> 00:28:51,800 Speaker 5: Well, since we wanted to talk about stocks that are 551 00:28:51,800 --> 00:28:57,080 Speaker 5: going up, and I know that's another one people kind 552 00:28:57,120 --> 00:28:59,440 Speaker 5: of forgot that video takes up a lot more storage 553 00:28:59,440 --> 00:29:02,040 Speaker 5: than textas and if we're going to start just like 554 00:29:03,160 --> 00:29:06,320 Speaker 5: throwing out the slop on video the part. Basically, one 555 00:29:06,320 --> 00:29:09,440 Speaker 5: of the only areas of semiconductors that has gotten ignored 556 00:29:09,760 --> 00:29:13,040 Speaker 5: has been storage, not like not like memory like like 557 00:29:13,200 --> 00:29:16,480 Speaker 5: uh RAM and stuff, but like hard drives and solid 558 00:29:16,520 --> 00:29:20,600 Speaker 5: state discs like Western Digital and uh okay. 559 00:29:20,360 --> 00:29:23,080 Speaker 4: So Seagate, Yes, that's what we're talking about. So this 560 00:29:23,240 --> 00:29:26,200 Speaker 4: was a this was an eighty four dollars stock in 561 00:29:26,760 --> 00:29:28,760 Speaker 4: late March, and now it's a two hundred and fifty four. 562 00:29:28,800 --> 00:29:31,960 Speaker 4: That's where you're going, Yeah, okay, yeah, yeah, there's a crazy. 563 00:29:31,960 --> 00:29:34,520 Speaker 2: There's a crazy, but it has gotten attention. But you 564 00:29:34,560 --> 00:29:36,120 Speaker 2: think it could go up more? 565 00:29:36,360 --> 00:29:39,640 Speaker 5: I think so, and I think that uh to illustrate 566 00:29:39,680 --> 00:29:42,200 Speaker 5: that dynamic that's going to keep happening. Basically, you just 567 00:29:42,280 --> 00:29:45,240 Speaker 5: keep looking at stuff where there's people say, oh, well, 568 00:29:45,280 --> 00:29:48,160 Speaker 5: you know there's a there's a glut of X and 569 00:29:48,200 --> 00:29:50,000 Speaker 5: that's why this isn't going to win from AI. And 570 00:29:50,000 --> 00:29:52,680 Speaker 5: then eventually the glut of X doesn't matter. It's a 571 00:29:52,960 --> 00:29:55,200 Speaker 5: That's what happened to Nvidia in the beginning of all this. 572 00:29:55,480 --> 00:29:58,840 Speaker 5: There was a glut of crypto GPUs and now nobody care, like, 573 00:29:58,920 --> 00:30:00,200 Speaker 5: wh when's the last time you heard some want to 574 00:30:00,240 --> 00:30:02,640 Speaker 5: be barrassed on a video because there's too many crypto minors. 575 00:30:02,920 --> 00:30:05,800 Speaker 4: Sand Disc was forty eight dollars in April second, that's 576 00:30:05,800 --> 00:30:07,360 Speaker 4: one hundred and twenty six dollars stock. 577 00:30:07,520 --> 00:30:08,080 Speaker 5: Pretty crazy. 578 00:30:08,160 --> 00:30:10,000 Speaker 4: These are just these are like just the tools are 579 00:30:10,040 --> 00:30:11,120 Speaker 4: just store just basic. 580 00:30:11,320 --> 00:30:12,320 Speaker 3: Yeah, yeah, crazy. 581 00:30:12,600 --> 00:30:17,040 Speaker 5: It's it is and it's it. I guess you could 582 00:30:17,120 --> 00:30:21,120 Speaker 5: be worried by that, or you could say this isn't uh, 583 00:30:21,240 --> 00:30:24,000 Speaker 5: this isn't like in Nvidia doubling in a month necessarily right, 584 00:30:24,040 --> 00:30:27,240 Speaker 5: that it's just people discovering. I think it's a good barometer, 585 00:30:27,880 --> 00:30:30,880 Speaker 5: which is important in a bubble, that people are still 586 00:30:30,920 --> 00:30:34,680 Speaker 5: paying very close attention. And you start to get worried 587 00:30:34,680 --> 00:30:37,240 Speaker 5: when everybody just goes into the cisco of it all. 588 00:30:37,440 --> 00:30:40,560 Speaker 5: But these are real companies that have real products. Of course, 589 00:30:40,560 --> 00:30:44,440 Speaker 5: there's some silly stuff that's happening, but it's not only 590 00:30:44,520 --> 00:30:49,160 Speaker 5: silly stuff. It's not twenty twenty one with the you know, yeah, 591 00:30:49,240 --> 00:30:49,840 Speaker 5: I don't need to. 592 00:30:51,480 --> 00:30:54,360 Speaker 2: So speaking of on the ground research, or maybe I 593 00:30:54,360 --> 00:30:57,680 Speaker 2: should say in the sky research. When it comes to robotics, 594 00:30:58,000 --> 00:31:00,360 Speaker 2: you had a robot dog, right, yeah, and of those 595 00:31:00,440 --> 00:31:03,600 Speaker 2: like really advanced ones, we got a unitary robot dog. 596 00:31:03,840 --> 00:31:05,640 Speaker 5: It is super impressive. 597 00:31:05,920 --> 00:31:09,120 Speaker 2: It's also incredibly creepy when it stands up on its 598 00:31:09,160 --> 00:31:09,760 Speaker 2: hind legs. 599 00:31:09,880 --> 00:31:10,920 Speaker 5: Yeah, have you seen the video? 600 00:31:11,040 --> 00:31:12,080 Speaker 3: Yeah, I saw your video of it. 601 00:31:12,200 --> 00:31:17,720 Speaker 5: Yeah, it's uh. And Unitry we went and visited Unitry too, 602 00:31:17,840 --> 00:31:22,440 Speaker 5: and it's I've seen Boston Dynamics. You know spot robot, 603 00:31:22,800 --> 00:31:26,280 Speaker 5: which is actually more impressive. Also could probably murder you 604 00:31:27,240 --> 00:31:30,440 Speaker 5: like and and the interesting thing about Unitry is, yeah, 605 00:31:30,480 --> 00:31:32,440 Speaker 5: this the robot kind of feels like a weapon. It's 606 00:31:32,520 --> 00:31:35,560 Speaker 5: very heavy and stuff, but it's even they're humanoid, like 607 00:31:35,600 --> 00:31:38,920 Speaker 5: it can't lift more than eight pounds. And also Unitree 608 00:31:38,960 --> 00:31:41,200 Speaker 5: is not focused on like making the brain. They just 609 00:31:41,240 --> 00:31:43,040 Speaker 5: want to make the hardware. And you go there and 610 00:31:44,560 --> 00:31:48,000 Speaker 5: they were like, where's your light are from? Oh, well 611 00:31:48,280 --> 00:31:51,520 Speaker 5: we needed this specific thing, so we just made it ourselves. Okay, 612 00:31:51,840 --> 00:31:54,680 Speaker 5: what about your actuators, Well we made those ourselves. Do 613 00:31:55,000 --> 00:31:57,080 Speaker 5: anything that they don't make themselves. They get within a 614 00:31:57,160 --> 00:31:59,880 Speaker 5: ninety kilometer radius of their headquarters. But where's their head 615 00:32:00,480 --> 00:32:07,720 Speaker 5: It's in hangzhoune So it's something where you're really that's 616 00:32:07,760 --> 00:32:09,800 Speaker 5: the competition that we have stacked up against us in 617 00:32:09,840 --> 00:32:14,080 Speaker 5: the US versus China, and I think the interesting route 618 00:32:14,080 --> 00:32:17,440 Speaker 5: that they've taken is it's super advanced. You can code 619 00:32:17,520 --> 00:32:19,800 Speaker 5: on it. You can already see videos of people you know, 620 00:32:19,840 --> 00:32:23,360 Speaker 5: buying them and making them do specific things, but they 621 00:32:23,400 --> 00:32:27,160 Speaker 5: made them not super like, it's not going to kill you. 622 00:32:27,720 --> 00:32:29,400 Speaker 5: And that's a great way to get a bunch of 623 00:32:29,400 --> 00:32:32,320 Speaker 5: training data because we don't have enough training data of 624 00:32:32,360 --> 00:32:34,000 Speaker 5: what it's like to move around the world. We have 625 00:32:34,040 --> 00:32:36,640 Speaker 5: the sum history of all human knowledge that's been recorded 626 00:32:36,720 --> 00:32:40,920 Speaker 5: for text models, but we need more data for robotics. 627 00:32:41,360 --> 00:32:44,840 Speaker 5: And I think that that's like an interesting kind of 628 00:32:45,200 --> 00:32:47,160 Speaker 5: thing where they're just kind of sending these out into 629 00:32:47,200 --> 00:32:49,320 Speaker 5: the world so that they can get enough data to 630 00:32:49,320 --> 00:32:51,800 Speaker 5: actually make it safe enough to have a robot in 631 00:32:51,800 --> 00:32:52,160 Speaker 5: your house. 632 00:32:53,160 --> 00:32:54,600 Speaker 3: Can anyone buy a unitary dog? 633 00:32:54,640 --> 00:32:54,760 Speaker 2: Like? 634 00:32:54,760 --> 00:32:56,640 Speaker 3: Are they for sale online? 635 00:32:56,760 --> 00:32:59,040 Speaker 5: Yeah, there's a lead time, but I can get around 636 00:32:59,040 --> 00:33:00,680 Speaker 5: it if you wait. 637 00:33:00,720 --> 00:33:01,360 Speaker 2: How much are they? 638 00:33:02,080 --> 00:33:05,320 Speaker 5: So they start at three thousand, and then we got 639 00:33:05,400 --> 00:33:08,719 Speaker 5: the super advanced one that you can attach a robotic 640 00:33:08,800 --> 00:33:10,800 Speaker 5: arm to so it can manipulate its environment, and that 641 00:33:10,840 --> 00:33:13,800 Speaker 5: was eight thousand dollars, and then they're humanoid is sixteen, 642 00:33:13,880 --> 00:33:15,680 Speaker 5: which is like less than the cost of the used 643 00:33:15,720 --> 00:33:16,360 Speaker 5: time to sither. 644 00:33:17,360 --> 00:33:19,840 Speaker 2: Well, three thousand. Also the price for like a really 645 00:33:19,880 --> 00:33:23,040 Speaker 2: good pure bread dogs, I don't know what that. 646 00:33:23,120 --> 00:33:25,719 Speaker 3: People pay a lot of money for actual dogs. 647 00:33:25,760 --> 00:33:27,760 Speaker 2: So yeah, it seems why are you looking at me? 648 00:33:28,400 --> 00:33:29,560 Speaker 3: Why do you think I'm looking at you? 649 00:33:30,120 --> 00:33:32,880 Speaker 2: Okay, James, that was so much fun. Thank you so 650 00:33:32,960 --> 00:33:35,920 Speaker 2: much for coming in and sharing with us your your 651 00:33:35,960 --> 00:33:37,600 Speaker 2: field trip observations. 652 00:33:37,800 --> 00:33:52,080 Speaker 5: Thanks Levan, Joe. 653 00:33:52,160 --> 00:33:55,480 Speaker 2: I think maybe it's the cynical journalist in me, but 654 00:33:56,360 --> 00:33:58,680 Speaker 2: I get really nervous when we start saying it's different 655 00:33:58,720 --> 00:33:59,200 Speaker 2: this time. 656 00:34:00,280 --> 00:34:02,920 Speaker 4: Yeah, I mean it's always different. It's always the same 657 00:34:02,960 --> 00:34:04,320 Speaker 4: as sort of there are. 658 00:34:04,240 --> 00:34:05,640 Speaker 3: Certainly you know. 659 00:34:06,080 --> 00:34:09,440 Speaker 4: Yeah, I mean, look, it's crazy and it's a lot 660 00:34:09,440 --> 00:34:12,160 Speaker 4: of money, and it's still you know, we've been asking 661 00:34:12,200 --> 00:34:14,680 Speaker 4: people on the podcast what kind of use value are 662 00:34:14,719 --> 00:34:18,279 Speaker 4: you getting out of this technology? And it's still very ambiguous. 663 00:34:19,800 --> 00:34:20,239 Speaker 5: I don't know. 664 00:34:20,360 --> 00:34:23,040 Speaker 4: I don't have It's not my job to time the market, 665 00:34:23,160 --> 00:34:25,120 Speaker 4: so I'm glad I don't have to have a view 666 00:34:25,280 --> 00:34:26,240 Speaker 4: on all of this stuff. 667 00:34:26,280 --> 00:34:28,520 Speaker 2: But I do think, cop out, Joe, No, I do think. 668 00:34:29,160 --> 00:34:29,680 Speaker 5: I do think. 669 00:34:29,719 --> 00:34:31,759 Speaker 4: For Okay, what's your okay, give us your give us 670 00:34:31,800 --> 00:34:33,040 Speaker 4: your price target. 671 00:34:33,280 --> 00:34:34,960 Speaker 3: That's okay, Okay, give us your price time. 672 00:34:35,920 --> 00:34:40,080 Speaker 4: Okay, I do think this is moving so fast and 673 00:34:40,160 --> 00:34:42,680 Speaker 4: every day the numbers are so big, and it's the 674 00:34:42,760 --> 00:34:45,400 Speaker 4: stakes are so high for all of this that it 675 00:34:45,480 --> 00:34:49,480 Speaker 4: does feel like we just need regular updates on the 676 00:34:49,520 --> 00:34:51,440 Speaker 4: state of how much money is being spent in this 677 00:34:51,560 --> 00:34:53,719 Speaker 4: area and what the eventual payoff will be. 678 00:34:54,000 --> 00:34:58,000 Speaker 2: I want someone to draw a massive, like infographic of 679 00:34:58,120 --> 00:35:00,799 Speaker 2: all the financing kills, because I think you could make 680 00:35:01,120 --> 00:35:02,239 Speaker 2: a really interesting one. 681 00:35:02,320 --> 00:35:03,120 Speaker 3: Yeah, you definitely could. 682 00:35:03,200 --> 00:35:06,000 Speaker 2: That would probably be illuminating. On that note, I did 683 00:35:06,040 --> 00:35:10,240 Speaker 2: think James's suggestion of, you know, shorting the private credit 684 00:35:10,400 --> 00:35:13,960 Speaker 2: players and then going along the actual physical hardware ones 685 00:35:14,200 --> 00:35:15,000 Speaker 2: that was interesting. 686 00:35:15,160 --> 00:35:15,359 Speaker 5: Yeah. 687 00:35:15,440 --> 00:35:18,000 Speaker 4: No, it's really interesting. Also, I had not realized. I mean, 688 00:35:18,040 --> 00:35:22,120 Speaker 4: it is very interesting the way booms or bubbles, whatever 689 00:35:22,120 --> 00:35:23,680 Speaker 4: you want to call it. I don't like using the 690 00:35:23,719 --> 00:35:28,480 Speaker 4: word bubble because it applies whatever booms or bubbles, the 691 00:35:28,520 --> 00:35:29,359 Speaker 4: way they spread out. 692 00:35:29,520 --> 00:35:31,200 Speaker 3: I just think that's extremely interesting. 693 00:35:31,280 --> 00:35:31,440 Speaker 5: Right. 694 00:35:31,600 --> 00:35:34,080 Speaker 4: So, in video takes off in late twenty twenty two, 695 00:35:34,160 --> 00:35:37,040 Speaker 4: when people are blown away by chat CHPT and they 696 00:35:37,040 --> 00:35:39,160 Speaker 4: say there's incredible and it's being powered by all these 697 00:35:39,640 --> 00:35:41,960 Speaker 4: in vidio chips, and then people go on the hunt 698 00:35:42,040 --> 00:35:45,200 Speaker 4: right for other things. It's like what else, And even 699 00:35:45,320 --> 00:35:49,799 Speaker 4: still today in October twenty twenty five, that spreading out 700 00:35:49,880 --> 00:35:53,280 Speaker 4: process continues. And so okay, in the last few months, 701 00:35:53,400 --> 00:35:55,799 Speaker 4: maybe it's like a caterpillar, or maybe suddenly people are 702 00:35:55,840 --> 00:36:00,160 Speaker 4: realizing that just sort of old fashioned computer storage is 703 00:36:00,480 --> 00:36:03,080 Speaker 4: to be very hot. But we're still in this phase 704 00:36:03,400 --> 00:36:06,920 Speaker 4: whereas people open their eyes to these things, the spreading 705 00:36:06,920 --> 00:36:08,919 Speaker 4: out continues to this day right. 706 00:36:08,840 --> 00:36:12,360 Speaker 2: Well, Also there's an expectation that eventually, you know, normal 707 00:36:12,400 --> 00:36:14,799 Speaker 2: companies are going to benefit from it as well and 708 00:36:14,840 --> 00:36:16,520 Speaker 2: become more efficient and all of that. 709 00:36:17,080 --> 00:36:19,160 Speaker 4: I just remembered the one question I was going to 710 00:36:19,160 --> 00:36:21,080 Speaker 4: ask James, but I could just throw it out as 711 00:36:21,160 --> 00:36:23,719 Speaker 4: sort of a spectative question. So he mentioned all of 712 00:36:23,719 --> 00:36:25,960 Speaker 4: the workers. Could you ask all of the workers that 713 00:36:26,000 --> 00:36:28,560 Speaker 4: are out this or that are building these things? And 714 00:36:28,640 --> 00:36:30,640 Speaker 4: it does make me wonder about the sort of crowding 715 00:36:30,640 --> 00:36:33,879 Speaker 4: out possibility, Like imagined if you imagined if you had 716 00:36:33,920 --> 00:36:36,879 Speaker 4: some other business enable Leen, and you're trying to hire 717 00:36:36,880 --> 00:36:41,000 Speaker 4: an HVAC worker right now because HVAC you know, you're like, oh, restaurant. 718 00:36:40,640 --> 00:36:42,279 Speaker 2: And you're competing with like open AI. 719 00:36:42,360 --> 00:36:45,480 Speaker 4: Yeah, you're like, you have a restaurant enabling, and you're like, 720 00:36:45,520 --> 00:36:47,800 Speaker 4: you know what, I need to update my air conditioner 721 00:36:47,960 --> 00:36:51,480 Speaker 4: or whatever or whatever, and what are you going to 722 00:36:51,560 --> 00:36:54,480 Speaker 4: do because all the HVAC workers are at that big 723 00:36:54,600 --> 00:36:57,439 Speaker 4: stargate operation. And I do think this is a very 724 00:36:57,440 --> 00:37:00,200 Speaker 4: interesting macro question on all kinds of things, which is 725 00:37:00,239 --> 00:37:03,280 Speaker 4: the degree to which all of this spending with uncertain 726 00:37:03,320 --> 00:37:06,360 Speaker 4: payoffs will have a crowding out and potentially inflationary effect 727 00:37:06,400 --> 00:37:07,360 Speaker 4: on the rest of the economy. 728 00:37:07,400 --> 00:37:11,960 Speaker 2: Well, presumably machine God will eventually design HVAC robots, so 729 00:37:12,160 --> 00:37:13,160 Speaker 2: problem will be solved. 730 00:37:13,200 --> 00:37:14,000 Speaker 3: Problem will be solved. 731 00:37:14,000 --> 00:37:15,040 Speaker 2: All right, shall we leave it there? 732 00:37:15,080 --> 00:37:15,759 Speaker 3: Let's leave it there. 733 00:37:16,000 --> 00:37:18,320 Speaker 2: This has been another episode of the oud Loots podcast. 734 00:37:18,440 --> 00:37:21,640 Speaker 2: I'm Tracy Alloway. You can follow me at Tracy Alloway. 735 00:37:21,360 --> 00:37:24,080 Speaker 4: And I'm Jill Wisenthal. You can follow me at the Stalwart. 736 00:37:24,200 --> 00:37:27,200 Speaker 4: Follow our guest James Van Gillen, He's at Sutrini seven. 737 00:37:27,440 --> 00:37:30,719 Speaker 4: Follow our producers Carmen Rodriguez at Kerman armand dash O 738 00:37:30,800 --> 00:37:33,759 Speaker 4: Bennett at Dashbod and Kale Brooks and Kale Brooks. For 739 00:37:33,880 --> 00:37:36,320 Speaker 4: more odd Laws content, go to bloomberg dot com Flash 740 00:37:36,400 --> 00:37:38,799 Speaker 4: odd Lots. We have a daily newsletter and all of 741 00:37:38,800 --> 00:37:40,879 Speaker 4: our episodes, and you could chet about all of these 742 00:37:40,880 --> 00:37:44,160 Speaker 4: topics twenty four to seven in our discord discord dot 743 00:37:44,200 --> 00:37:45,279 Speaker 4: gg slash. 744 00:37:44,960 --> 00:37:47,640 Speaker 2: Odd lots And if you enjoy odd Lots, if you 745 00:37:47,840 --> 00:37:50,759 Speaker 2: like it when we talk about the physical realities of AI, 746 00:37:50,920 --> 00:37:53,400 Speaker 2: then please leave us a positive review on your favorite 747 00:37:53,440 --> 00:37:57,120 Speaker 2: podcast platform. And remember, if you are a Bloomberg subscriber, 748 00:37:57,160 --> 00:38:00,160 Speaker 2: you can listen to all of our episodes absolutely add free. 749 00:38:00,440 --> 00:38:02,400 Speaker 2: All you need to do is find the Bloomberg channel 750 00:38:02,440 --> 00:38:06,280 Speaker 2: on Apple Podcasts and follow the instructions there. Thanks for listening.