1 00:00:02,800 --> 00:00:07,400 Speaker 1: Z Media. Hell, welcome to this week's Better Offline Monologue. 2 00:00:07,440 --> 00:00:10,160 Speaker 1: I'm your host ed Zeitrono. As I wrote while drafting this, 3 00:00:10,240 --> 00:00:12,320 Speaker 1: I'm the house house, a line I'm leaving in as 4 00:00:12,360 --> 00:00:26,920 Speaker 1: I'm being accountable to you, the customer. Anyway, you've heard 5 00:00:26,920 --> 00:00:29,120 Speaker 1: about the in Video Open Ai deal. By now that 6 00:00:29,280 --> 00:00:31,920 Speaker 1: InVideo will and I put this in air quotes invest 7 00:00:31,960 --> 00:00:34,280 Speaker 1: one hundred billion dollars in open Ai. And I want 8 00:00:34,280 --> 00:00:36,960 Speaker 1: to start out by being explicit that Invidia is not 9 00:00:37,120 --> 00:00:40,440 Speaker 1: wired open Ai as single goddamn dollar. All right, So 10 00:00:40,600 --> 00:00:42,960 Speaker 1: the deal is that in VideA is investing ten billion 11 00:00:43,000 --> 00:00:45,960 Speaker 1: dollars up front, according to CNBC, despite the fact that 12 00:00:46,200 --> 00:00:49,280 Speaker 1: the date before it said that they would be investing 13 00:00:49,320 --> 00:00:52,520 Speaker 1: ten billion dollars with the first gig. But anyway, they're 14 00:00:52,560 --> 00:00:55,160 Speaker 1: investing ten billion dollars in a month, allegedly, and I 15 00:00:55,240 --> 00:00:58,120 Speaker 1: quote here once the deal is finalized, which means in 16 00:00:58,200 --> 00:01:01,480 Speaker 1: simpler terms, that nothing has actually been and Video will 17 00:01:01,480 --> 00:01:04,640 Speaker 1: then invest ten billion dollars in tranches for each Gigawa 18 00:01:04,720 --> 00:01:08,039 Speaker 1: have capacity to that open Ai built. That's right, That's right. Everyone. 19 00:01:08,319 --> 00:01:10,679 Speaker 1: Most of this funding is gated behind open Ai, a 20 00:01:10,720 --> 00:01:13,400 Speaker 1: company that burns billions of dollars building huge amounts of 21 00:01:13,480 --> 00:01:17,480 Speaker 1: data centers, something they have never done, and that also 22 00:01:17,600 --> 00:01:20,839 Speaker 1: costs billions of dollars. And just as an aside, open 23 00:01:20,840 --> 00:01:23,679 Speaker 1: ai really has built nothing themselves. Everyone else is building 24 00:01:23,680 --> 00:01:27,000 Speaker 1: it for them. These fucking silicon value libertarians. They love 25 00:01:27,080 --> 00:01:31,039 Speaker 1: themselves and welfare, don't they? Anyway? How does open ai 26 00:01:31,160 --> 00:01:34,640 Speaker 1: afford to build these data centers? I don't know. Nobody 27 00:01:34,640 --> 00:01:37,720 Speaker 1: else seems to know. Where are they being built? Who knows? 28 00:01:37,760 --> 00:01:40,680 Speaker 1: It's unclear. Has the land been purchased? Also unclear. There's 29 00:01:40,720 --> 00:01:43,960 Speaker 1: some sort of crap announcement about oh stargate, We've got 30 00:01:43,959 --> 00:01:47,480 Speaker 1: a list of different fucking new places, and I had 31 00:01:47,480 --> 00:01:50,160 Speaker 1: someone on the reddit suggests I was wrong. I just 32 00:01:50,200 --> 00:01:52,320 Speaker 1: want to be clear. Of the list of the five 33 00:01:52,440 --> 00:01:56,000 Speaker 1: new stargate locations, one of them is in Shackleford, Texas, 34 00:01:56,080 --> 00:01:59,080 Speaker 1: which was already chosen to be one of Oracle's data centers. 35 00:01:59,200 --> 00:02:02,200 Speaker 1: This was already an announced. They're re announcing stuff as 36 00:02:02,200 --> 00:02:05,160 Speaker 1: a pr professional, When you're re announcing shit, it's because 37 00:02:05,160 --> 00:02:08,519 Speaker 1: you run out of stuff to announce this real. But look, 38 00:02:09,080 --> 00:02:13,079 Speaker 1: let's let's be explicit about one thing. Data centers take 39 00:02:13,120 --> 00:02:15,840 Speaker 1: a long long time to build, right, now, Crusoe is 40 00:02:15,840 --> 00:02:18,840 Speaker 1: building a data center in Abilene, Texas, made up of 41 00:02:18,880 --> 00:02:21,200 Speaker 1: eight buildings, which will amount to about one point two 42 00:02:21,240 --> 00:02:23,640 Speaker 1: gigawats of compute capacity. This is the one for Oracle 43 00:02:23,639 --> 00:02:26,200 Speaker 1: and open AI, and at the current rate of development, 44 00:02:26,200 --> 00:02:30,000 Speaker 1: it's taken about two and a half years per giga one. Look, 45 00:02:30,160 --> 00:02:32,760 Speaker 1: I realize it's easy to get drowned in big numbers 46 00:02:32,760 --> 00:02:35,280 Speaker 1: and ridiculous promises and media headlines where the writers don't 47 00:02:35,280 --> 00:02:37,440 Speaker 1: seem to bother to learn even the simplest fucking thinks. 48 00:02:37,600 --> 00:02:40,440 Speaker 1: But data centers are expensive and take forever, and this 49 00:02:40,639 --> 00:02:46,680 Speaker 1: entire plan hinges upon open AI building data centers. In fact, look, 50 00:02:46,880 --> 00:02:50,000 Speaker 1: I've seen lots of people suggest that the ten billion 51 00:02:50,040 --> 00:02:52,520 Speaker 1: dollars the open AI is getting from Nvideo will let 52 00:02:52,600 --> 00:02:56,320 Speaker 1: them build data centers. And again, can somebody please actually 53 00:02:56,360 --> 00:03:00,600 Speaker 1: learn the basics here? The Abilene data center venture requires Crusoe, 54 00:03:00,639 --> 00:03:04,040 Speaker 1: Blue Oul and Primary Digital Infrastructure to raise fifteen billion 55 00:03:04,080 --> 00:03:07,000 Speaker 1: dollars in debt just for the data center itself, the construction, 56 00:03:07,080 --> 00:03:09,400 Speaker 1: the people, the water, the power, and all that good stuff, 57 00:03:09,520 --> 00:03:12,040 Speaker 1: with Oracle footing the bill for the hundreds of thousands 58 00:03:12,040 --> 00:03:14,560 Speaker 1: of GPUs to go inside. At a price of I 59 00:03:14,560 --> 00:03:17,080 Speaker 1: think about twenty four billion dollars. It works out to 60 00:03:17,680 --> 00:03:20,320 Speaker 1: by my calculations, open ai will need one hundred and 61 00:03:20,360 --> 00:03:22,919 Speaker 1: twenty five billion dollars just to build the data centers 62 00:03:22,960 --> 00:03:25,600 Speaker 1: to get the funding from in Video, which will require 63 00:03:25,639 --> 00:03:28,320 Speaker 1: another two hundred billion dollars in GPUs. By the way, 64 00:03:28,639 --> 00:03:31,200 Speaker 1: the information also reports that in Video is going to 65 00:03:31,240 --> 00:03:35,320 Speaker 1: be potentially leasing GPUs to open ai, setting up a 66 00:03:35,400 --> 00:03:39,040 Speaker 1: special purpose entity to buy the GPUs from themselves that 67 00:03:39,040 --> 00:03:43,240 Speaker 1: they would then lease to open AI. I've really hesitated 68 00:03:43,320 --> 00:03:47,200 Speaker 1: to bring up any Enron comparisons. I've kind of flirted 69 00:03:47,200 --> 00:03:51,080 Speaker 1: with one with core We even then that's that's straight 70 00:03:51,240 --> 00:03:53,640 Speaker 1: Enron shit. Like, man, they haven't conferred they're going to 71 00:03:53,680 --> 00:03:55,880 Speaker 1: do it. It's an information report or put a link 72 00:03:55,960 --> 00:03:58,080 Speaker 1: in there. You'll complain that you have to pay for it. 73 00:03:58,200 --> 00:04:03,560 Speaker 1: I recommend the information. I just every day, every day, 74 00:04:03,680 --> 00:04:05,800 Speaker 1: these new pieces of news where everyone kind of like 75 00:04:05,920 --> 00:04:08,960 Speaker 1: trumpets them like Jesus just woke up from behind that 76 00:04:09,000 --> 00:04:13,040 Speaker 1: big fucking rock, when in fact, I think something far 77 00:04:13,120 --> 00:04:16,159 Speaker 1: worse is happening. But I really want to break something 78 00:04:16,200 --> 00:04:18,320 Speaker 1: down for you, and you'll likely hear this again in 79 00:04:18,360 --> 00:04:20,760 Speaker 1: a future episode. I want to get into the actual 80 00:04:20,800 --> 00:04:25,000 Speaker 1: economics here, Cruso's one point two gigawatts of compute for 81 00:04:25,040 --> 00:04:27,800 Speaker 1: open Ai is a fifteen billion dollar joint venture, which 82 00:04:27,880 --> 00:04:30,560 Speaker 1: means a gigawa of compute runs about twelve point five 83 00:04:30,640 --> 00:04:34,080 Speaker 1: billion dollars. Abilene's eight buildings are meant to hold about 84 00:04:34,120 --> 00:04:37,080 Speaker 1: fifty thousand in video GB two hundred GPUs in their 85 00:04:37,080 --> 00:04:40,480 Speaker 1: associated networking infrastructure. So let's say a giga what is 86 00:04:40,520 --> 00:04:45,120 Speaker 1: about three hundred and three thousand, three hundred and thirty 87 00:04:45,120 --> 00:04:48,000 Speaker 1: three Blackwell GPUs. Yeah, this math is a little bit 88 00:04:48,080 --> 00:04:50,560 Speaker 1: funky due to in Video promising to install theirny Reubin 89 00:04:50,680 --> 00:04:54,920 Speaker 1: GPUs in these theoretical data centers. But I really do 90 00:04:54,960 --> 00:04:57,560 Speaker 1: think that they'll require about just a little under two 91 00:04:57,640 --> 00:05:00,279 Speaker 1: hundred billion dollars worth of GPUs. And this is on 92 00:05:00,360 --> 00:05:02,800 Speaker 1: top of the four hundred billion dollars the open Ai 93 00:05:02,920 --> 00:05:05,560 Speaker 1: has promised to spend with Oracle over the next five years, 94 00:05:05,920 --> 00:05:09,040 Speaker 1: when open ai had already leaked that they'd be burning 95 00:05:09,080 --> 00:05:11,560 Speaker 1: one hundred and fifteen billion dollars over the next four years. 96 00:05:11,800 --> 00:05:15,400 Speaker 1: This math does not make any sense. It's complete nonsense. 97 00:05:15,839 --> 00:05:18,000 Speaker 1: For open ai to survive, they're going to need around 98 00:05:18,080 --> 00:05:21,920 Speaker 1: six hundred billion dollars in the next five years. It's ridiculous, astonishing, 99 00:05:21,960 --> 00:05:24,400 Speaker 1: and insulting to the general public that the media keeps 100 00:05:24,440 --> 00:05:27,120 Speaker 1: acting as if these deals are signed or promoted in 101 00:05:27,120 --> 00:05:29,600 Speaker 1: good faith. And I'm shocked that more people aren't doing 102 00:05:29,640 --> 00:05:32,960 Speaker 1: the very simple maths that And I've seen a disgraceful 103 00:05:32,960 --> 00:05:36,280 Speaker 1: amount of people talking about Open Aiy's revenue projections, which 104 00:05:36,320 --> 00:05:39,000 Speaker 1: involve them making one hundred billion dollars in twenty twenty 105 00:05:39,040 --> 00:05:41,080 Speaker 1: eight and one hundred and forty five billion dollars in 106 00:05:41,120 --> 00:05:44,279 Speaker 1: twenty twenty nine, and that's revenue, not profit. They're losing 107 00:05:44,440 --> 00:05:46,880 Speaker 1: tens of billions of dollars both those years, and they're 108 00:05:46,920 --> 00:05:49,880 Speaker 1: acting as if these are serious, meaningful statements rather than 109 00:05:50,000 --> 00:05:54,120 Speaker 1: egregious lies from a gaggle of Charlatan's. It's a fucking joke. 110 00:05:54,440 --> 00:05:57,200 Speaker 1: It's sickening. Oh, by the way, Open Eyes projected to 111 00:05:57,200 --> 00:06:00,520 Speaker 1: make thirteen billion this year. From Mike calculs, they made 112 00:06:00,520 --> 00:06:04,560 Speaker 1: about six point twenty six billion. Is it's wank And 113 00:06:04,600 --> 00:06:07,719 Speaker 1: that's the technical term they use it. The financial analysts 114 00:06:07,760 --> 00:06:09,840 Speaker 1: put it. When you pay JP Morgan to see their 115 00:06:09,839 --> 00:06:12,240 Speaker 1: private shit, it's or yeah, this is all wank. This 116 00:06:12,400 --> 00:06:15,960 Speaker 1: is wink. Okay, I'll keep going now for the next 117 00:06:15,960 --> 00:06:18,720 Speaker 1: three part episode I'm gonna put out next week. I've 118 00:06:18,760 --> 00:06:20,800 Speaker 1: done a lot of digging into how much actual money 119 00:06:20,880 --> 00:06:24,640 Speaker 1: exists to invest in general, and we're genuinely running up 120 00:06:24,640 --> 00:06:27,800 Speaker 1: against the limits of private and public capital. By my calculations, 121 00:06:27,839 --> 00:06:30,600 Speaker 1: there's only about four hundred and seventy seven billion dollars 122 00:06:30,640 --> 00:06:33,479 Speaker 1: of available capital across the top ten private equity funds, 123 00:06:33,800 --> 00:06:36,440 Speaker 1: and I'm deadly serious. I will be publishing the links 124 00:06:36,720 --> 00:06:38,680 Speaker 1: and maybe one hundred and sixty four billion dollars of 125 00:06:38,800 --> 00:06:41,440 Speaker 1: US venture capital left, and maybe let's call that another 126 00:06:41,480 --> 00:06:44,840 Speaker 1: one hundred and fifty BILLI in other funds sloshing around. 127 00:06:46,279 --> 00:06:49,039 Speaker 1: Are we really going to carve out half to two 128 00:06:49,080 --> 00:06:52,560 Speaker 1: thirds of available global capital for one fucking company that 129 00:06:52,600 --> 00:06:56,159 Speaker 1: burns a billion fucking dollars every two fucking seconds. I 130 00:06:56,240 --> 00:07:00,600 Speaker 1: know it's slower than that. Shut up. I'm just I'm 131 00:07:00,640 --> 00:07:02,880 Speaker 1: a little bit ornery. I'm not going to lie. You 132 00:07:02,920 --> 00:07:04,680 Speaker 1: can hear it in my voice. And I am even 133 00:07:04,760 --> 00:07:08,080 Speaker 1: more ornery than ever, more ornery than ornery, as Rob 134 00:07:08,200 --> 00:07:11,760 Speaker 1: Zombie once sung. I'm just pissed off because when you 135 00:07:11,800 --> 00:07:14,440 Speaker 1: look at these numbers, they do not make sense. There's 136 00:07:14,560 --> 00:07:18,640 Speaker 1: no parallel in history for any company or any industry 137 00:07:18,680 --> 00:07:21,520 Speaker 1: that's ever spent this much money. There's never been one 138 00:07:21,640 --> 00:07:26,040 Speaker 1: company that has been the focus like this. And honestly, 139 00:07:26,080 --> 00:07:29,040 Speaker 1: I've never seen a company less deserving of any investment 140 00:07:29,080 --> 00:07:31,880 Speaker 1: than open AI. They've burned billions of dollars. Their product 141 00:07:31,920 --> 00:07:36,600 Speaker 1: is questionably useful, it's incredibly commoditized. They have yet to 142 00:07:37,080 --> 00:07:39,720 Speaker 1: create anything really exciting or new since And if I'm 143 00:07:39,760 --> 00:07:43,520 Speaker 1: being kind reasoning, but let's be honest. GPT four, Oh 144 00:07:43,760 --> 00:07:48,080 Speaker 1: they are washed. And yet everybody's sitting around them clapping 145 00:07:48,160 --> 00:07:51,360 Speaker 1: like they're a gifted child and insult to children and 146 00:07:51,520 --> 00:07:54,120 Speaker 1: startup founders. And that really is something that pisses me 147 00:07:54,160 --> 00:07:56,720 Speaker 1: off as well, because AI took thirty three percent of 148 00:07:56,840 --> 00:07:58,960 Speaker 1: US venture capital last year and I think they're probably 149 00:07:58,960 --> 00:08:02,680 Speaker 1: taking more this year. And what's insane is nothing has 150 00:08:02,720 --> 00:08:07,000 Speaker 1: been proven. None of these companies are profit generating, none 151 00:08:07,080 --> 00:08:10,360 Speaker 1: of these products are popular. Every single day, it feels 152 00:08:10,400 --> 00:08:12,640 Speaker 1: like a study comes out saying, yeah, no one really 153 00:08:12,760 --> 00:08:14,720 Speaker 1: likes this stuff, no one wants to pay for it. 154 00:08:15,600 --> 00:08:18,600 Speaker 1: And by my calculations, I think the whole industry's revenue 155 00:08:18,640 --> 00:08:21,840 Speaker 1: have include core we even all those shits, maybe fifty 156 00:08:21,840 --> 00:08:24,840 Speaker 1: five billion dollars in twenty twenty five. That's on four 157 00:08:24,920 --> 00:08:30,080 Speaker 1: hundred billion plus capex. Is anyone having fun? Does this 158 00:08:30,240 --> 00:08:32,960 Speaker 1: feel like the future or does this feel like the 159 00:08:32,960 --> 00:08:35,360 Speaker 1: future of how we're going to blow hundreds of billions 160 00:08:35,400 --> 00:08:39,640 Speaker 1: of dollars on nothing? And the longer it takes for 161 00:08:39,679 --> 00:08:42,200 Speaker 1: people to walk away from open Ai, and the longer 162 00:08:42,240 --> 00:08:45,840 Speaker 1: it takes for this company to die, the more damage 163 00:08:46,120 --> 00:08:48,880 Speaker 1: that's going to be caused. Because right now in video 164 00:08:48,960 --> 00:08:52,000 Speaker 1: is putting everything on open Ai. If we had functioning 165 00:08:52,040 --> 00:08:55,440 Speaker 1: regulatory boards, we'd actually have something done about this. But instead, 166 00:08:55,480 --> 00:08:57,320 Speaker 1: what's going to happen is they're going to inflate this 167 00:08:57,360 --> 00:09:00,000 Speaker 1: bubble just a little further, and then when it bursts, 168 00:09:00,080 --> 00:09:02,760 Speaker 1: it's going to fuck retail investors and a ton of 169 00:09:02,800 --> 00:09:05,800 Speaker 1: venture capitalists too, ton of private equity people as well, 170 00:09:05,840 --> 00:09:09,160 Speaker 1: which will be kind of funny, but the financial apocalypse 171 00:09:09,200 --> 00:09:13,320 Speaker 1: will hit everyone and it will be grizzly. But I'll 172 00:09:13,360 --> 00:09:16,040 Speaker 1: be here to talk about it. I'm having the time 173 00:09:16,080 --> 00:09:18,199 Speaker 1: of my life doing this show. Next week's three part 174 00:09:18,240 --> 00:09:21,040 Speaker 1: Who is Completely Insane? I wrote it in kind of 175 00:09:21,040 --> 00:09:25,800 Speaker 1: a two day, fifteen thousand word kind of like ejection. 176 00:09:25,880 --> 00:09:26,960 Speaker 1: I was going to say, I was going to say 177 00:09:27,000 --> 00:09:29,559 Speaker 1: another word, which wouldn't have sounded quite as good, But nevertheless, 178 00:09:29,600 --> 00:09:33,439 Speaker 1: that's the fun brandy part of these monologues. But I'm 179 00:09:33,440 --> 00:09:35,080 Speaker 1: really excited for you to hear it because I pulled 180 00:09:35,080 --> 00:09:37,600 Speaker 1: together a lot of numbers that show how fascical this 181 00:09:37,720 --> 00:09:41,040 Speaker 1: industry has become. And I think as the bubble approaches 182 00:09:41,080 --> 00:09:42,839 Speaker 1: its eventual burst, a lot of you are going to 183 00:09:42,880 --> 00:09:45,400 Speaker 1: feel really good about how right you are. Maybe some 184 00:09:45,440 --> 00:09:47,800 Speaker 1: of you still disagree with me, and I will work 185 00:09:47,880 --> 00:09:50,600 Speaker 1: to compel you, and even if I don't, I hope 186 00:09:50,640 --> 00:09:53,520 Speaker 1: you have a good time listening. Got great episodes coming 187 00:09:53,559 --> 00:09:55,480 Speaker 1: up in the future as well. Going to be going 188 00:09:55,559 --> 00:09:58,040 Speaker 1: to see Stephen Burke of Gamers next is, do a 189 00:09:58,040 --> 00:10:01,000 Speaker 1: long interview with him, some fun you a better offline, 190 00:10:01,480 --> 00:10:04,520 Speaker 1: and then I'm already working on some big guests. It's 191 00:10:04,520 --> 00:10:06,560 Speaker 1: going to be a really fun few months. I don't 192 00:10:06,600 --> 00:10:09,720 Speaker 1: know if anyone else, but I'm definitely enjoying doing this podcast. 193 00:10:09,920 --> 00:10:11,120 Speaker 1: Thank you, zever for listening.