Musk said he would do it from space. Now SpaceX wants 80,000 cell sites

Elon Musk and SpaceX, they're, you know, reportedly going to be uh building out a terrestrial network for their their direct to sell mobile service here. Uh Bloomberg reporting they're going to be uh giving uh access to 80,000 tower sites uh needed for that buildout here in this country. And in Crown Castle CEO, they're big tower operator saying you can assume we're talking to all parties involved, right? >> And Crown Castle shares were spiking. So we're seeing this happening. Musk has said I'm going to build out this whole terrestrial network for our direct to sell service, etc. But but again, can you challenge these these top three mobile carriers? >> But the other problem for us is the money it takes to build a mobile terrestrial network. And look, what we have been saying collectively is >> you're probably not going to build it. You're going to buy it because it would be so expensive and take so many years to get a return on that, >> right? And on this report yesterday, not only did Crown Castle shares go up, but again, all the carrier shares came down. And so as we've talked about, as you've theorized, like at some point, do you put enough pressure on one of those carriers with >> So I mean, what's funny is it's like Musk wants to like a katamari wants to collect every low margin business. >> Great game, by the way. >> Oh, yeah. Probably makes more money than a cell carrier. I mean, I was looking into a lot of junk debt few weeks ago, >> and always the cell carriers are on it. It's always they're just capex heavy businesses. He wants data centers. He wants rockets. He wants cell carriers. What's he going to move into? Like >> consumer package goods. Like what's the next? He's going to become a wholesaler. Like it's just a very strange thing. And I think that >> I think he just wants to do a bit of everything. He wants to do Cena Weebo, I guess. Like I think he wants he really wants to do the everything app, but if you look at the quality of like Twitter and like X money or what have you, these are rough products to say the least. I mean, I guess the current cell carriers are pretty crap themselves, so it's not really a huge challenge. Why? >> Well, there is this hubris, right? There is this like this it's this whole Silicon Valley first principles kind of thing. I anything you can do, I can do better kind of thing. If I start from the beginning, you the examples that you brought up, but there's also the example of Doge. I can come in. What is a more legacy operation than the United States federal government >> and this thinking that I can come in and do it differently and fix it? >> The most complicated bureaucracy in the world. What's funny is he could have just gone and dealt with Medicaid fraud, >> the huge amounts of doctors over billing Medicare, an obvious >> clear set well doumented place like no I'm going to f get rid of US aid like the the we do that so that the countries like us so we have soft power right and relatively cost nothing right uh but back to the whole cell >> sorry no back real quick on on the carrier stuff is that you know he did have this first principal thing where we're going to come in for over the top from space get you direct cell service down down to from satellites. Oh, that doesn't work. We have to get the terrestrial network. So, you're just doing it all over again. What AT&T and Verizon how is it any different, right? >> I mean, but the stuff that they're doing now does work very well for the for the use case that exists. We've said this, Starlink internet, I've said this, if you're out in like a remote part of the world, Starlink internet is unbelievably, >> but that is a different proposition than running a cell phone in the New York City subway. Right. >> Yeah. >> Yeah. That's a different proposition. >> One little thing that I that Julie you you spotted this was that and you might like this at in SpaceX 10Q they talk about you know part of that gain that they had was a $561 million gain from their investment in Elon Musk's boring company right which he also owns. So it's like forget circular financing. You have the circular musconomy here, right? With >> we got to we got to at the end of this we got to reeregulate accounting standards and stuff like we got to work out a way to fix this net income boosting from company valuations. The things that Google did with SpaceX's valuation, the gains on anthropic and all that. We need to change the equity. Like we actually have to do something about this because companies should not be able to post massive profits based on oh yeah on paper I tweak this here and there. I don't know. I think it's if we do that though, we open up Pandora's box with all of the net asset values of things like Blackstone's funds and Blue Owl and all the I mean, we have a problem with private credit that we could spend an hour talking about. I know it's just frustrating to watch everyone fall for Musk again, though, because you say, "Oh, AT&T stock crash." It's just like, >> are the markets piloted by actual children? It's like that. Oh, they're piloted by bots. >> They're piloted by line and then and then it gets I mean and I don't mean bots in the sense of AI quantitative good old James of the world who quant trade. >> Remember that Catrini thing from a few months about the global intelligence crisis? Yeah. The only global intelligence crisis was the bunch of idiots read it and freaked out. It's like oh agents are going to take on Mastercard. Shut up. >> So, okay. So Ed, I want to ask you about something else because I was reading your most late your latest note and I know you've been asked about this before. You are super emotional about this stuff >> and you really paint the people who are doing all of this as villains. In your most recent note, you single out Jensen Wong in particular, but do you think that that sort of like diminishes your case because you are so vehement about the villain of the of these people? And is it possible that they actually believe in this stuff and that's why they're selling it? >> I think that we have insufficient contempt as a society for these people and I think it's one of the reasons that they get away with so much. And I actually think people say, "Oh, you're too emotional." I think that people are insufficiently emotional about this. The stakes are huge. The damage like Musk's data center, I think, is in Mississippi, uh, gassing a black neighborhood. People are just like, "Oh, yeah. I don't feel anything about that. Oh, it feels I don't know how. I don't think it diminishes my case at all. And I think what people's reaction to stuff like that and the I can't repeat what I said about Jansen Huang in the newsletter, right? >> Familyfriendly show. >> Yes. >> Here's the thing. I think it's really easy for people to see that and go, "Who cares about the rest of it?" Because it's a really easy way of just not thinking. It's a comfortable way of getting away from the actual real problems because Jensen Hong is channel stuffing. It's obvious. He's forcing GPUs down the channel, saying, "You have to buy these now or they won't be available in a few months." I think that's morally reprehensible. I think the hyperscalers involving themselves in it are morally reprehensible because it's driving up the price of every imaginable consumer electronics. It's making everyone's life worse so that Jensen Hang's life can be better and so that whoever owns Nvidia stock can feel slightly better. I mean, the gains in Nvidia stock have not been >> on on the channel stuffing point from from Nvidia. >> Is there a is there any legitimacy in your eye to the argument that it is just simply a demand signal? Jensen's going to the market and saying, "Look, if you want these, you should buy them now because we can see where demand is. We see the three, four, five, sixyear lead times for products in the semiconductor space, and if you want to get in, you should get in now." >> Legitimacy there. >> How is their scarcity of a product that they have too many of? How and if the scarcity is because I think 70% of their accounts receivable for the first half of the year was five companies. One company made up 16% of their revenue and everyone's taking months early delivery of these GPUs. There's not scarcity and other than that which is created by Jensen Hong. This isn't like 100 companies and they're saying yeah you got to pre-order because there's 99 other companies. No, they all know each other. They all >> right. But all of those companies are building out compute as fast as they want to build it out. I mean, what? I'm sorry. Like that. What does more data centers actually get us? We talk about these things all the time. What does more data centers do other than allow them to charge anthropic and open AAI more? It isn't really clear what the actual point of more of them is. Amazon just canled a data center in Luden County. Like what do we we need more clarity at a minimum into what these data centers are like how many of them are online because Bloomberg reported a few week a month ago actually that Microsoft only has 2 GW of AI compute capacity and that's out of 12 G of reported. What is going on here? And the thing is a demand signal would suggest that there was demand for something cuz you need it right now versus I literally don't need this right now but Jensen saying oh well you're going to take it right now. the early rack receipt program to quote Eugene Kim over at Business Insider. >> And so this isn't demand in the traditional sense. It's not like they're getting them a month early. There was one quote in that Business Insider story that said that there was an email from an NG saying, "We have no foreseeable path to installing these." That's insane. >> But that's one story. Has there been others that have supported this whole idea of channel stuffing? >> I reported with the Guardian that Microsoft has only got 2 gigawatts worth of capacity before Bloomberg. And I have read the list of Microsoft's usable GPUs. They have spent $250 something billion in capex. They have about $50 billion worth of chips in operation. So yes, the answer is yes. >> So real quick, you know, we talk about you have these chips that are sitting in storage. They're not like they actually stay in time and they're good to go next year. Like they can age, right? They can actually age out tech technically speaking >> technically speak in years, right? So that's also potentially a concern if you have chips that are just not being used and aging out. Uh but you know talking about SpaceX and Musk again you know they've sort of showing that their data center or their sort of their business with Macroheart is making some money right the question is I guess your question would be or your answer would be they're not going to keep be able to do that in the future we're seeing them monetize it right now >> but most of that's coming from anthropic it's all of this is just a chain of venture capital pays Microsoft Google Amazon Meta or not Meta didn't work out SpaceX and who pays anthropic 80% of the revenue per ramp comes from 1% of the customers who are mostly ventureback startups. Who funds them? Venture capital. Who funds anthropics so they can pay the c the uh hyperscalers? Venture capital. Everything's funded by chains of debt and venture capital funded by pension funds. Who funds pension fund? Well, the pension funds also fund the private credit firms who are also building the data centers. >> And we've kind of seen commentary in the last 24 hours around that maybe there's not enough money to fund this. At some point, the will is going to dry up somewhere along this chain and suddenly all of it's not really going to work out. Sure, the hyperscalers can raise again, but can Anthropic and Open AI get the cash to pay