We came to this world. To reach the stars. We came to this world. To shape our future. We came to this world. To reach the stars. To shape a future. Reaching to feel the new. Sam. You're watching TVPN. Today's Tuesday, December 9, 2025. Just a few days until Christmas. We're so excited. We're live from the TVPN Ultradome, the temple of technology, the fortress of finance, the capital. Capital. No travel. Christmas, baby. No travel until the new year. We're in the ultra dome, hanging out. We're monitoring the situation. We're monitoring always the Paramount, Netflix, Warner Brothers situation. Because this is one of the most fascinating deals. The deal has gone hostile. Paramount has launched a hostile bid to acquire Warner Brothers. It's an all cash offer. 76.9 billion. Really? Really. And it's a fun one because I feel like it's. It's obviously tech adjacent, but it's not the story that people have been monitoring all year. We've been talking about foundation models. That whole story has just gotten a little bit. Legacy media. They got tired of not getting enough attention. Exactly. Let's spice it up a little. Spice it up. Here's something new to learn about. So everyone's having fun. Everyone's learning new things. And I could tell because when I came in today, I had a bunch of questions that people were throwing at me about how all of this works. Why isn't Warner Brothers just going with the highest price? Like, when I sell a stock, I don't care if Citadel or Jane Street's buying it. If I'm selling 80 bucks a stock, just give me the best price. But when you're selling an $80 billion company, there are other considerations, and it goes beyond just maximizing shareholder value. And so I wanted to break down a few of those. So I did that in today's newsletter. We can run through that and then we can run through some of the news. Perfect. But first, let me tell you about ramp time is money save both easy to use, corporate cards, bill payments, accounting, and a whole lot more all in one place. So this question, you know, it seems obvious. The board has a fiduciary duty to maximize shareholder value. That's legal requirement to take the higher offer. And yet that's not what's happening. Like yesterday, we saw that, that Ellison and Paramount came in with over 100 billion. Of course, that included the CNN, the TV. But even when you broke it out, it seemed like it was very clear that Ellison was willing to pay more money and make a higher offer. So under what circumstances can a board whose job it is to maximize shareholder value, not take the higher offer? It can't just be whoever we had the better dinner with. Right? Which is part of the news, of course. They went out to dinner and Paramount CEO David Ellison sent a text to David Zaslav, who's running Warner Brothers, after they made a hostile bid today to buy Warner Brothers. And David text the other David and says, david, but I guess he misspelled his name, even though that's his name, which is just funny. But anyway, he says, I appreciate your underwater today, so I wanted to send you a quick text. Know that despite the noise of the last 24 hours, I have nothing but respect and admiration for you and the company. It would be the honor of a lifetime to be your partner and to work and to be the owner of these iconic assets. He's talking about Foghorn Leghorn. He's talking about Porky Pig. He's also talking about Batman and Superman, obviously, and Harry Potter and Lord of the Rings and a million other iconic assets, which is true. If we have the privilege to work together, you will see that my father and I are the people you had dinner with, which I like that. I think that's cool. They had dinner. It's a fantastic text. Yes, yes, yes. It's a great one. So, quickly, let me tell you about Julius AI, the AI data analyst that works for you. Join millions who use Julius to connect their data, ask questions, and get insights in seconds. So there are two main reasons why you don't just take. Why your size might not be size. Directors at a company like Warner Brothers. They have to maximize shareholder value, but maximizing shareholder value is an expected value calculation. So if you come in with $100 billion offer, and I think there's a 75% chance that you're going to deliver that. And someone else comes in with an $80 billion offer, and I think there's 100% chance that they're going to deliver that. Well, the expected value of Your bid is 75 billion. The expected value of their bid is 80 billion. I go with the 80 billion. Even though it's like a lower headline number, it has a higher expected value. How much of the calculus do you think is just a deal that will actually get done? Right, A deal that is going to get. So a deal that doesn't get done could still have value because of a breakup fee. So you could, in theory, go into a deal that you know is impossible. And the example that I gave is, what if Bytedance came in and they were like, hey, you know, we're $400 billion media companies. We'd love to own Porky Pants. We'd love to own these iconic assets. Let's pick up Warner Brothers. We'd love to own CNN too. You know, who knows what we'd put on the TV. We already own TikTok. We already own TikT. Why not CNN also? Why not Shark Week? Not Shark Week. They'd have to carve out Shark Week in my book. But obviously the government would block that. And we have cfius, which is an organization in the US government that determines whether or not an international buyer can take an American company because of intellectual property. All sorts of geopolitical considerations. We don't want another country cornering a market on a really key piece of the supply chain like the Nvidia H200 for example, which we might be talking about later in the show with Aaron Ginn, co founder of Hydra, host, repeat guest on the TVPN show. But then there's also just the ftc. So certain deals, like I gave you the example yesterday, like Disney, Disney would be so blocked, I think that that deal doesn't even get kicked around. No one even talks about it. And because it would get blocked by the ftc, so it doesn't even happen. But it would maximize shareholder value if Disney came and said, yeah, we'll give you a $10 billion breakup fee and we want to try and buy you for 200 billion or a trillion or 200 million. Like you would immediately say yes because you just want the breakup fee. But being in this turmoil and being in this limbo and not being able to sign deals with other, that has an opportunity cost. Right. And so you might want to back off of that. And so basically the, yeah, you look. Back at the FIGMA acquisition, Adobe paid a billion dollar breakup fe effectively like non dilutive financing for figma that has a happy ending because they were able to, you know, re accelerate, get out into the public markets. But there was another situation where they went through a rough patch and actually really needed that capital to get through. Yeah, so, so there's, so there's, there's basically two buckets of risk that I think most, most dealmakers would be considering in this situation. First is financing risk. So. Will they come through with the money and what money are they paying with? Because the initial offer, this is the history. There's actually been six offers put forward by Paramount. David Ellison is putting on a clinic of just not taking no for an answer. Because all the way back on September 14, he offered $19 a share, and 60% of that was cash. So 60% cash offer. Then September 30, two weeks later, he comes back $22 a share and ups the amount of cash to 66.7%. Zaslav is doing a masterclass in making your opponent negotiate against themselves. 100%. It's actually great. And so October 3, 23, 50 a share, 80% cash. Then November 20, 25, 50 a share, 85% cash. Then finally, December 1, $26.50 per share, 100% cash. December 4, $30 a share, 100% cash. Now, why does this matter? Well, it's because you get locked into owning the shares of 40% of the shares is paramount. And then while it's closing, the stock trades down, you wind up getting less value. And so the 80 billion today might wind up being 70 billion later. And that doesn't maximize shareholder value. And so there's also just the question of, like, can you actually marshal the cash? Just, like, if somebody comes to you and you're about to sell your house to them and they say, yeah, I'm definitely going to get a loan for this, well, that's a financing risk. Maybe they don't. Maybe they back out of the deal. Why cash offers often times. Exactly. If you can prove you have the cash, that makes difference. And I mean, he's effectively gone and done that because he's teamed up with Jared Kushner, reportedly, and gone around the world, got a whole bunch of different sovereign funds. And just really, people, anyone with big, deep pockets has kind of said, yes, I'm down to come along on this ride and put up a bunch of the capital. The capital has been marshaled. It seems like it's ready to go. And then also you have Larry Ellison, who has three times as much money, I think, than the whole deal value, something like that, 275 billion to his name. And he's only not cash assets, not cash. But he should not just, I know they want it, but do not market. Sell your Oracle position, please. Yeah. I'm actually going all cash now. But he's effectively acted as backstop. And so everyone who's come in and said, okay, I guess I'm good for 10 billion of that, 80 billion, but only if everyone else is in. Larry Ellison reportedly has come in and said, well, you know, if there's one person in the bunch that backs out of their slug at the last second, I'll jump in and get that. He's not saying he's going to put up the whole 80. He's just saying he's backstopping it. Yeah. And that's the term that's been thrown around. So. David Ellison feels very good because he's done what Zaslav wanted him to. He said, hey, they wanted an all cash offer. I brought them an all cash offer. I brought them an offer that's higher. And we believe that it's. There's no reason why factor number two should come into play in factor number two is regulatory approval. And so there's been this question about will Netflix get approved? If there was a different buyer, it would have even more regulatory risk. And so you don't want to. Even if the price is higher, you don't want to accept a higher price with a lower chance of actual, actual conversion. Right. Bobby thinks NLE Choppa might up his offer to 100k of cash and get back in the. In that scenario, at least you know that the 100k cash is real and it's going to be delivered on the. Day of maybe in actual cash. Probably. I think that was the. That is funny because that was the joke of that meme was that he was giving like actual cash. When we're saying cash offer here, of course we are not talking about physical cash. Hey, you never know. Are you as well over there? Yeah. Is Tyler doing money spread? What does he got? Did Christmas come early for you? So I had a question though. What makes the Paramount offer, like, hostile? Is it that it's like reacting to. The Netflix office at the Netflix board already or, sorry, the Warner. Warner Brothers board went with Netflix and they're coming. It's hostile because they're saying, screw the board decision. We want to go to the shareholders. Okay. Yeah. Because the. It sounded like they had a nice dinner and the texts were like, kind, yes, yes, yes. No, no, no, no, no. Hostile in the sense of like, of like not accepting the final, like the final decision there and then. And then also there is a. There is a point where you can appeal directly to the shareholders and make the. You can actually make the legal case that the board is not acting in their fiduciary duty and in their. In the interest of shareholders. Because. But again, these things need to be argued in shareholder lawsuits because it is fuzzy because, like, if I say I'm offering you 100 billion and someone else is offering 80 billion and the board says, yeah, but that hundred billion has only a 75% chance of going through, you have to discount that. Now the board says it's worth 75. But whoever made that offer can say, no, we're actually good for it. You should apply a 10% discount rate and treat our offer as if it's 90% or 100%. Good. Right. And so, and so all of those arguments, those can be made in the court and the shareholders can, you know, react to that. And if the shareholders align and say, yeah, actually we think we'd be getting more money here with this, with this deal, then they can push back against the board at a certain point. But yeah, Friend of the show, high powered media exec says Zaslav has been architecting this situation for the last 12 months. Sitting pretty and he's a wily old fox and has what he has been shooting for. Two heavy hitters fighting over a deal. Don't doubt there will be a couple more turns here with the price or even another bidder coming in sideways. Don't quote me. Which is remarkable because the price is so high already compared to where it was six months ago. Zaslav. It really feels like it's coming together that he will be remembered as an incredible business executive for this deal. I want to know more. I'm still learning about this particular industry and so I've been, I've been having fun digging in, but possible deal guy. Of the year, possibly. He's in the conversation. He's definitely in the Was Sam looked like he was running away with. Yes, summer. That's true. Early fall. And he lost ball control. He lost ball control, yeah. Well, quickly, let me tell you about Fin AI, the number one agent for customer service Automate, the most complex customer service queries on every channel. But. There is a wrinkle that people haven't really been considering, which is the fact that Warner Bros. Discovery holds a massive, often underrated vault of masculine cinema. And if this falls into the wrong. Hands, what is masculine cinema? Masculine cinema being dudes off film. So they own, they own a bunch of Clint Eastwood films. Dirty Harry, Magnum Force, the Enforcer, Heartbreak Ridge, American Sniper, Letters From Iwo Jima, Unforgiven, Gran Torino. They also own the classic Steven Seagal Run. Basically every movie where Steven Seagal was a massive theatrical star before going to direct to video. We're talking about above the Law, Hard to Kill, Marked for Death, out for Justice, Under Siege, which is Die Hard, On a Boat, Under Siege, Two On Deadly Ground, Fire Down Below, Exit Wounds. They also own Rambo, the Whole Sebaster, Stallone, the Cobra Wing, Cobra, Demolition man, the Specialist, Tango and Cash Bullet to the Head, get Carter the 2000 remake. They also own Mad Max, Lethal Weapons, all of the Lethal Weapons. The Matrix blade, Mortal Kombat, 300, Rush Hour, Pacific Rim and a number of dad and military shows. Band of Brothers, the Pacific Generation. Kill. There's a whole host. They should release the full DVD set. Just called Guys being Dudes. Yes, they also own some Jason Statham properties. The Meg franchise where Jason Statham fights a big shark. Whoa. Wrath of Man. I guess most of the Guy Ritchie films are actually with Lionsgate or mgm, but there are a variety. And I think this could be a, you know, a big political hot button of what happens to the Rambo franchises. The Hollywood hunks. No one's talking about the Hollywood hunks. Everyone's focused on the Looney Tunes characters. But if the Hollywood hunks fall into the wrong hands, it could be. It could swing our entire culture potentially. Anyway, let me tell you about the. Anthony Michael on our team worked on Wrath of Man. Oh really? No way. First job. Very cool. Very cool. Automate compliance and security AI that powers everything from evidence collection and continuous monitoring to security reviews and vendor risks. You're hiding inorderate amounts of Middle Eastern funding sources in your takeover bid, are you not says boring business. And not hiding it. It's out in the open. Like this is, this is. This is like the playbook for these. But again double. Double that of the amount of equity that the Allison's are putting in. Putting it around. Yeah. I mean I think everyone assumes that there'd be a lot of Middle Eastern funding in this. That has become the standard funding instrument these days. I don't know like EA Games. We just followed that story a little bit. That was Obviously it's like 93% Saudi money and that is part of the modern deal making playbook these days. Every. I'm pretty sure most of the. A lot of venture funds have raised money over there. A lot of the big AI companies have raised money over there. Like that seems like a foregone conclusion that if you need money, you go where the money is. Why do bank robbers rob banks? Because that's where the money is. Well, Dylan Byers says. David, I appreciate your underwater today. So sent you a quick text. He's talking about the dinner. Oh, so that was the correct. That was the misspelling D A I V D. He must have been texting very quickly. It does give you. Let's. Let's turn this into a copy Pasta. Yes, I enjoyed the dinner. So Exec Sum has an overview of who's actually Putting in the money here. Filings reveal that Paramount's bid for Warner Brothers is backed by equity commitments including 11.8 billion from the Ellison family, 24 billion from Saudi Arabia's, Qatar's and Abu Dhabi sovereign wealth funds, 1 billion from China's Tencent, additional commitments from Redbird Capital Partners and Jared Kushner's Affinity Partners. That doesn't add up to 100 billion, so there must be cash coming from somewhere else. Maybe there needs to be some more reporting until. Oh, they did give an updated version of this post but doesn't quite add up. So will be interesting to follow where you know how the deal comes together. And we have a few more people joining this week to help us break down the deal. Ben Smith from Semaphore tomorrow and then Dylan Byers. Oh, Dylan's coming. Puck on Thursday in person. I'm excited. Oh, he's coming in person. Yes. That's great. Well, let me tell you about numeral compliance. Handled numeral worries about sales tax and VAT compliance so you can focus on growth. Trump says that the United States will allow Nvidia H200 chip sales to China and get a 25% cut. This is a pretty big change. I mean we were talking about. Not selling. I mean, I guess it's not Blackwell, right. It's Hopper. So we're still a generation behind. But was this, it was like a pretty nerfed chip before. We're getting less nerfed. We seem to move, we seem to be moving in a, in a more thumbs up direction. More. Let's, let's actually get the chips over to China at the same time. AI is fake so it doesn't matter. Right. That's the take, the pro China take is that it's like it's SaaS. It's SaaS. Or it's not like this nuclear bomb that's going to destroy everyone. So it's harder, it's getting harder to make the. The Manhattan Project argument. Right. And it feels impossible to make the argument that we're going to get them addicted to the US AI supply chain and they'll never develop their own capabilities and so we want them. Well, I disagree with that. I think that, that, I think that that argument actually holds. That argument holds for sure if we. In, in the next. I mean, I totally disagree. I think they're, I think China's smart enough to know they don't want to be dependent on any foreign country to produce any critical. Yes. But I still think there's enough of like a market force within China that if we flood the market with cheap Nvidia chips. It'Ll just be expensive for them to keep propping up their local industry, even if they are aware of it, even if they know that they have to. It's a cost. And it's something that a lot of AI researchers over there will just say, you know what? I'm already. It's so much easier. The Nvidia ecosystem is so great. I'm just going to stick with that. I don't know. I'm a little bit skeptical that, like, there's no advantage to selling Nvidia chips there. I've become more. I've become more receptive to that argument, particularly, even though you have not. But that's fine. Tyler, what were you going to say about this? Yeah, I was just going to say, I mean, we kind of joke about this, but it is kind of crazy the extent to which you can basically bake down all, like, AI policy questions to, like, if you are AGI pilled. Yeah. Because, like, if you are. If you're Dario. Yeah. I mean, you shouldn't be giving stuff to China. You also shouldn't, like, if you are a GPI pill, you should be thinking about safety, all these things. Like, we had Keith, her boy on. He doesn't seem super AGI pilled. He's also like, oh, safety is a hoax, et cetera, stuff like this. I think basically all of these questions, you can just. Well, it depends, because if you're super fast takeoff pilled and you're super like, asi, it's like. But you think that it can somehow be contained in whoever gets it first. You actually don't mind if China's six months behind you, because as soon as you hit that inflection point, you're 1,000 years ahead. And so you might as well just do whatever it takes, fund whatever, make as much money as possible, as long as it accelerates you to be the first one to hit that inflection point. So there is an argument where you can be extremely AGI pelt, even super intelligent. Well, Truth Social Media is creating a prediction markets product within the app. I thought you were going to say AI product. Did they launch an AI thing last week? I think that was fake news. Oh, that was fake. Okay. But it's possible Trump is going to use the prediction markets on Truth Social in order to kind of set AI policy going forward, say, H2 hundreds to China. Good or bad. That's wild. I mean, that's not that. Like Robin Hanson says that that's like a good idea. That was the whole, that was the whole, that was the whole pitch originally. There is definitely a bull case, you could say like a steel man for that. Yeah, there's always a steelman. Yeah. I mean, where's your helmet, John? Yeah, yeah, yeah. Put on the helmet. Well, let's read through this Wall Street Journal article on the, on the details of the new Nvidia deal. But first, let me tell you about figma. Think bigger, build faster. Figma helps design and development teams build great products together. So notable that this seems to have been completely priced in already because Nvidia is actually down half a point today. Yes, it's been such a dramatic story all year, and yet it's always felt like a complete footnote in the overall financial performance of Nvidia. It's never felt like, okay, yeah, if this goes Nvidia's way, they're going to double their revenue or double their profits or like, you know, really move the needle. They're growing so fast that, you know, I don't know how many we're going to get into figuring out how many H2 hundreds they sell, but they would need to sell a lot to actually move the needle on this behemoth of a business. What is the world's largest company in the world? So President Trump said he would let Nvidia export its H200 chip to China and that the US would receive a 25% cut, his latest bid to make money for the government in an unusual agreement with a private company. I have informed President Xi Jinping of China that the United States will allow Nvidia to ship H200 products to approved customers in China and other countries under conditions that allow for strong national security. 25% will be paid to the US of A. The move is a boon for Nvidia, which has fought for months to maintain access to the world's second largest economy. The company had agreed earlier this year to give the US 15% of China's sales from a lower performing chip, only for the Chinese to scuttle those plans as part of continuing trade talks between the two sides. Chips from the world's most valuable company have become a prized geopolitical tool. The H200 has higher performance than the H20 that Nvidia was previously allowed to sell. But this isn't as powerful as the company's top Blackwell products released this year, nor the Rubin generation chips generation of chips coming next year. The move follows a meeting between Trump and Jensen Huang last week where the pair discussed H200 exports, people familiar with the matter said Nvidia shares added nearly 2% after hours. Not too bad. Even with the US government taking a cut. The decision could be worth billions of dollars in sales to Nvidia, which enjoys comfortable margins on its AI chips. In the most recent fiscal quarter, Nvidia reported gross margins of 73.4% on $57 billion in sales. That is crazy. They can totally afford 25% for the big guy. I have no idea if it'll be legal. We'll have to figure out if this gets approved because in general, the United States does not import enforce export duties. That's not something America's done historically. We had some folks on the show to contextualize that and give us some background when this was first floated. Feels like it is happening now, but we'll see where it lands in August. Nvidia CFO Colette Krass said that if geopolitical issues subside, the company could ship between 2 billion and 5 billion of chips to China per quarter, which increase which could increase if orders pick up. So they're doing. Yeah. Right around that time, Lutnick was on CNBC and this was the quote that originally ticked off the Chinese. He said, we don't sell them our best stuff, not our second best stuff, not even our third best. And he said, you want to sell the Chinese enough that their developers get addicted to the American technology stack. That's the thinking. And of course they CCP basically immediately said we don't want any of them now because we're offended. But it does seem like, yeah, I don't know, we'll have to have Bill Bishop or someone on the show to contextualize how China is receiving this information, whether or not they will actually buy it. Of course we are going to be joined by Aaron ginn in just 30 minutes, so stay tuned for that. Also, let me tell you about LINEAR Meet the System for modern software development. LINEAR streamlines work across the entire development cycle from roadmap to release. So the exports could help Chinese tech giants that have struggled to get top chips to train their models. Huang has argued Nvidia should be allowed to compete in the Chinese market because China has many of the world's top AI researchers and the US should want them using American technology. If you're you are not going to replace China, wong said at an event. At a think tank event, Trump said that the government would take a similar approach to exports from Nvidia competitor AMD as well as intel, in which the government now has a 10% stake. The approval comes just weeks after administration advanced officials, including Secretary of State Marco Rubio torpedoed a push from Nvidia to sell a slimmed down Blackwell chip to China before a recent trade meeting between Trump and Xi Jinping. Some officials, including Aizar David Sachs and Commerce Secretary Howard Lutnick, have backed exporting the H200 because it could be a good compromise that allows Nvidia to compete with China's Huawei Technologies without vaulting China past the US in AI. People familiar with the discussion said, I don't know. It's such an interesting question. I still wrestle with like, we exported a ton of Teslas to China and BYD and Huawei have now arguably completely leapfrogged. Completely leapfrogged. So they did not become addicted to the American. That's what I've got. That was the point I was making earlier. You can make the art, like almost with every single product. They've said, we'll work with you to make this thing that you want to make. We're really good at making things. And then they ultimately just make a better version of said product and make it for cheaper. Yes. And in the case of cars right now, we're obviously not allowing these cars to be imported into the US but they can simultaneously say, we're happy to keep making you these things. We're also going to compete with you directly and constantly try to be better than what you do. I just wonder if you play it back and you don't allow Tesla to export the amazing Model s or Model 3. Does that slow down BYD's development of their car, or Huawei's or Xiaomi's development? My understanding is they were able to basically get a paid education making Teslas and they were able to leverage that into making. But that's about making the car there. If you don't make the car there. Which is because the Nvidia has a Shanghai. But that's research. Not made there. That's not made there. It's very different when you're saying, okay, we're gonna go and we're gonna go and produce this product there and like, you are gonna get educated there. And the world's different. But you remember, we've gone through this before. China's like five year plans to create a domestic chip industry. They've been doing five year plans for 60 years. I know they're building. That's what I'm saying. And I would argue that it's working. They're not caught up, but they are Certainly made, you know, massive amount of progress. They've made more progress than any other country on earth. Yeah. I don't know. What do you think, Tyler? I think if you want to think about like getting the Chinese addicted to our chips, like, how addictive are the chips? Yeah. Because if you can make a comparison between like, you have the Nvidia chips and then you have the Chinese or like tpu. Right. Because the TPU is like, in some ways it's harder to use. The open source is not as good. But if it's just more economical, like if the actual hardware is just like a little bit cheaper, then it doesn't really matter how worse the software is, people will eventually move to it. So it's like, I don't think there is actually that much like soft power in the kind of general open source stuff like that in regards to, like Nvidia. It doesn't seem very addictive outside of it just being cheaper. Yeah, yeah. I mean, at the same time, like, like we are America is very much like addicted to Chinese solar panels right now. Like, the Chinese solar panels come here, they're cheap. And so we don't wind up buying or building a domestic solar panel industry because to get anything off the ground, you have to go in and say, okay, we're going to deal with having no margin forever and no venture capitalist can underwrite it and no private equity firm can underwrite it. And so it just doesn't really happen. Yeah, but the Chinese, they do do that. I guess they do bear the cost more. So it doesn't matter. But it's like they wouldn't have to bear the cost. So we're imposing a little bit of a cost on them. It's like they actually have to pay it. Whereas they don't. Whereas they wouldn't if we didn't do that. Right. Because if we don't go in and we don't sell them Nvidia chips, then they're like, okay, well, it makes economic sense to back Huawei and fund them. Whereas now they're like, it's not economically logical to back Huawei, but we have to, because there's so much pressure from better chips flooding the market like Nvidia chips, that if we don't, if we don't fund Huawei uneconomically, if we don't lose money on Huawei, like, we will never get to. We will never get, we will never be competitive. Yeah. I think it seems to me that like, either way they're going to fund it. The government, the Chinese government. I think that's Jordan's point, too. Really? I don't think it actually matters that much. Yeah, I don't know. Well, let's. Let's move on. But let's also tell you about Adeo, the AI native CRM. Adeo builds scales and grows your company to the next level. The United states will allow Nvidia's H200 processors its second best. So this is a. This is a change from what Ludnick said, right? Ludnik said, we're not selling the best. We're not selling the second best. Turns out we are selling the second best now. Maybe we'll sell the first best soon. Who knows? Well, we should play this clip from Patrick o' Shaughnessy's fantastic interview with Gavin Baker on Invest like the Best. He's talking about space data centers. We had Dallian on the show yesterday talking about space data centers. Dalian still thinks Gavin. It's too far out. Gavin knew we needed a new narrative. He said after the fateful BG2 episode ended, the AI trade, we needed a new trade. And now we're getting the space data center trade. Okay, let's hear the space data center. I think the most important thing that's going to happen in the world in the next three to four years is data centers in space. If you think about it from first principles, data centers should be in space. What are the fundamental inputs to running a data center? They are power and they are cooling. And then there are the chips, the inputs to making the tokens come out of the magic machines. So in space, you can keep a satellite in the sun 24 hours a day. That's pretty cool. And the sun is 30% more intense. And this results in six times more irradiance in outer space than on planet Earth. So you're getting a lot of solar energy. Point number two, because you're in the sun 24 hours a day, you don't need a battery. This is a giant percentage of the cost. So the lowest cost energy available in our solar system is solar energy and space. Okay. Second, for cooling in one of these racks, a majority of the mass in the weight is cooling. And the cooling in these data centers is incredibly complicated. You know, I mean, the H vac, the CDUs, the liquid cooling. In space, cooling is free. You just put a radiator on the dark side of the satellite. It's gold, and it's as close to absolute zero as you could get. So all that goes away, and that is a vast amount of cost. And Then how. This isn't the problem A bit more complicated than that. The dissipating heat in space, this is. What everyone's fighting about. I'm not exactly sure how it plays out, like, the physics of it, but a lot of people are saying, like, yes, it's in a vacuum, but because it's in a vacuum, you can't just. You can't. Like, it's not as passive of cooling. It's more of a material science problem. Like, there's more physics problems there also. Like, I don't know, I feel like you might need some. I feel like you might need some batteries, because I think there are orbits that are effectively in the sun all the day, but I think that the Earth's shadow, like, hits them not like, every day, but, like, every once in a while. You'Ll be caught in a shadow almost in, like, in a. What's it called? I don't know, some sort of. You just get caught in the shadow. Not every day, but, like, every 12 days or every 30 days or something like that. And so in that scenario, I mean, maybe it'd be fine if you're doing inference and you don't and you can just turn off the whole system for a little bit and there's some redundancy. I just texted our resident space man, Dalian. Yeah, I said, answer this. I just said, quick take. Yeah, on this. It's urgent. It's urgent. We'll get back from him. There's more to that clip, though. We should play this because there's a funny part in here. What, are you going to connect those racks? Well, it's funny. In the data center, the racks are over a certain distance connected with fiber optics, and that just means a laser going through a cable. The only thing faster than a laser going through a fiber optic cable is a laser going through absolute vacuum. So if you can link these satellites in space together using lasers, you actually have a faster and more coherent network than in a data center on Earth. Okay. For training, that's going to take a long time just because it's so big. But for inference, let's think about the user experience. When I asked Rock about you, and it gave the nice answer. This is crazy, though. They've done a podcast together five times. Why is he asking about Patrick? I was like, what is going on? Went to some sort of Metro aggregation facility in New York, probably within 10 blocks of here. There's a small little Metro router that's routed those packets to a big XAI data center somewhere. And Then the computation was done and it came back over the same path. If the satellites can communicate directly with the phone and Starlink has demonstrated direct. To cell phone this whole thing. I would like to know how much space X exposure Gavin has, because if this does become a dominant, at least narrative over the next few years. Yeah. A lot of the people that we know that have computers in space are not exactly bullish on data centers in space on a short time horizon. Yep. We have not. We had the founder of Star Cloud on. He's obviously bullish, but I don't believe he has any satellites in orbit. Yep. We've had Dalian on. He puts things into orbit, he brings him back down. He's talked about having. He technically has. I mean, Delian's in a hilarious position because Founders Fund has a huge position in SpaceX, obviously. But also Delian has had to do the hard work of like getting something in space and probably understands like how much of a grunt work it is and how much of a grind it is to get even. Just like something as small as like the pod into space. And then he's also seen at Crusoe, like what a real data center looks like. And the idea of putting something so massive into space, like it's bigger than the Hubble, it'd be bigger than the iss. What are you thinking about the space data center thing? Have you evolved on this? Yeah, I mean, I don't know. To me it seems just kind of impractical. Like if you think about, if you're doing inference or something, like how much of the cost of the token is actually in the energy. Right. Because I feel like one of the main benefits of having the data center in space is that energy is basically free. Right? It's right there. Sure. But if you think about, like, I think Casey Anwar has talked about this. Like, if you think about an inference, when you're doing inference, the portion of the cost that's actually the energy, it's like very tiny because it's like the chip is expensive or doing like, it's not that expensive. So it's like, okay, the cost gains overall are like, not that great. And then it's like, well, what if you want to change the chip to the next generation? Or what if you need to do some kind of like, mechanical work? Are we ignoring how broken Cluster would take some sort of intervention to fix as well as the cost per kg in fuel costs? Yeah, just. Just for context. So Gavin was at Fidelity and led a $900 million investment in SpaceX at a $12 billion post money valuation. That's amazing. Wow, what a go. That's incredible. Okay. Actually. Yeah, so maybe they didn't do the full 900 million. Yeah, Grok isn't. Sure. Sure, sure, sure. But, but I mean, either way, like, you can. You can be pro or con on space data. And grok also says SpaceX has been a tradie's largest venture capital position as of 2022, comprising an estimated 30 to 40% of its VC portfolio. Feels wild. So. So anyways, Gavin is set to. On one hand, Gavin could know things about space data centers that the rest of us don't know because of his position. And proximity to SpaceX. At the same time, SpaceX, if they go public next year, would benefit a lot from having a space data center kind of like, narrative as part of this. It's crazy because the business is so solid. I feel like you don't need a fourth act or whatever. Okay, wait, so what about this? So if a lot of the gains of space and center are just because the energy is free, why don't we just. But then there's this hard part about having, like, different satellites everywhere. Put the data center on the moon, there's no atmosphere, so you don't get the 30% reduction in energy from. From the sun and just put data. Center on the moon. Well, don't you get a 50%. Reduction because there's like, there's a dark side of the moon. Like the moon you wouldn't put on. The dark side of the moon. I know, I know, but what I mean is that our dark side sometimes gets light from when it's not facing us. Right. And so if you're on the moon, you only get light 50% of the time. Whereas if you're a satellite and you're flying out in space and you're kind of orbiting the Earth, but at a weird axis that you're not constantly in shadow all the time. You're getting sun, he says, 24 hours a day. I think it's. I think it's like 99.9% of the time. But regardless, it's like almost always. Whereas if you're. If you're actually, like, on the moon, you don't get sunlight 100% of the time. You get a 50% of the time. I'm pretty sure. I don't think it's because maybe there's some. Yeah, the sun's not. The sun's not like. Or the moon's not locked. It's Earth locked. Pull up this post, I put it at the bottom of the stack from Joe Morrison. He's over at Umbris. While we pull that up. Let me tell you about Restream 1 livestream 30 plus destinations. If you want to multi stream go to restream.com and we will pull up this next post. Joe says you just put a radiator on the dark side of the satellite and he says thermal engineers in absolute shambles right now. Yeah. People are not. I don't know. People are so scared. Casey Hammer says to be fair this is a big satellite. Joe says many big satellites. Big satellite makes it easier. I don't know. I don't know. Yeah, I think you need like, like a lot of like you need. You need basically a lot of mass in order to dissipate enough. Sure. Heat from a, from a GB300 or something. GB200.200 somehow AST Space Mobile Inc. Is up 27% in the last five days. So. People are. Quite excited. Yeah. Well, should we head over to Gulfstream News? There's some massive news from Gulfstream Aerospace. The G400 introduces next gen Gulfstream tech to its class. Sound on to learn more with Gulfstream president Mark Burns. How has Mark Burns not been on the show yet? Here we go. Clearly don't have a no render policy over here. They're. They're render racing. I think they're just not out. You think they could just be adding special effects? What do you mean? This is the most rendered video I've ever seen. This looks like it was rendered okay. But here he is 2010. You think this is real though? Bring Gulfstream performance standards. This is on a green screen cabin class fulfilling our customers needs for a product line. I knew it was too good to. Be true and enhance every journey. Like this is real. This what it looks like. It's far away. It does look like a nice plane though. With Gulfstream's signature combination of range, speed and cabin comfort, the G400 will provide unrivaled efficiency thanks to the combination of the advanced practical. This is going to be a hot Christmas gift this year. Absolutely. People have been wondering. I think this will be top of. This will be under the tree for a lot of people. It's not ideal to do like a pre order as a gift. You know it's nice to have something under the tree but I think people would make an exception here. So funny. I was talking, I was talking with a mutual friend of ours, a guest who's been on the show many times and they're Planning to purchase a jet at some point in the relatively near future. And he was so fixated on like needing to be able to. It's his first, it will be his first jet. He was so fixated on like I need to be able to stand up at like it can't have, like the cabin needs to be big enough that I can like walk around freely. I was like, come on, for your first jet, can you not just like, you know, tolerate like slightly lower height in the cabin? He was like, no, I gotta be able to. I gotta be able to move freely. Go straight for the 747, the BBJ. Yeah, get the VVIP package. The one that only the guitarist have for some reason. Well, Casey Hanmark. Okay, I got it back. I got a text back from Delian. I won't read all of it. Okay, there's a line here. This is from someone else. Dude, all this data centers in space stuff also I'm not going to read that you need 1 million 100 kilowatt sats to generate 100 gigawatts in orbit. 100 kilowatt sats Wang. 1000 kilogram including payload would mean a 5x increase in power density. And you can't just not have batteries and slap a radiator on the back. Lol. And then there was some choice words exchanged that I won't read on the show. But anyways, I think we should host the data center in space debate. That sounds good. We should get a few people on here. Somebody to be bullish, somebody to be bearish. I think we want somebody to be bullish that has put things into orbit and taken them down many times. I think it's hard to. It's hard to lean on somebody's kind of take or opinion if they're not actually participating. In space. They're just sort of speculating. Yeah. Let's read through a little bit of what Casey Hanmer had to say. He said here's one idea about SpaceX's next big thing. AI computing, inference. Again, not training on orbit. But how the hell can SpaceX do this? Cheaper than just building more data centers on the ground from first principles. It's an attractive proposition because the GPUs have extremely high value per kilogram and extreme revenue per kilowatt, both of which are relatively expensive. That is the value prop somewhat washes out the pain of operating in space. So I took a closer look. If anyone can make this work, it is a Starlink derived system. So I started with the Starlink V3 satellites. With some high fidelity CEO CAD below orbital parameters Pick a Sun synchronous orbit so we're in full 1400 kilowatt per meter squared sunlight at all times. No need for batteries. Deploy the solar array in sun slicer mode facing full sun, but the edge is pointing in the orbital direction bottom right in these images to minimize drag. But the inference Starlink star thought satellites don't have to scrape the atmosphere being in sun synchronous orbit. They'll need to use the rest of the Starlink constellation for backhaul via laser links anyway and higher orbits actually improve worst case latency very slightly too high though, and SSO is relatively full of debris. Let's pick 560km. A Starlink satellite in this orbit has full sun, so the back half is always shaded and relatively cool. The next hottest thing in the sky is the Earth taking up almost half the sky to the bottom left in these images. And so what does he conclude? He says, I've seen a bunch of high inclination Starlink launches from the Vandenberg recently, but I don't think any of them were going to sun synchronous orbit. In any case, a ring of Inference satellites visible at dawn and dusk running north and south will be awesome. If one Starship can launch 100 tons to Leo, then that gets close to 30 megawatts of inference per launch. 1,000 watts is 300 gigawatts. Now we're talking real scale. We're doing one launch of Falcon 9 per day. There was delions loosely. We're now one a day. So 300 in a year so you get to 1,000 over three years if you're on a one a day cadence or if you can wind up getting more the math is not that crazy, but it does seem like Delian was pushing more towards like hey, this is maybe 10 or 20 years, not the next three years. But it certainly will be fun to follow the story. In the meantime, let me tell you about cognition, the team behind the AI software engineer Devin Crush your backlog with your personal AI engineering team. Jordi has survived. I keep like getting the podcast in a can like down the wrong they. Should really make a make a drink specifically for podcasters that has no carbonation that cannot be choked on in any way. Mark Benioff says LLMs are the new disk drives commodity infrastructure you hot swap for whoever's cheapest plus best the fantasy that the model is the moat just expired. Marc Benioff having fun on the timeline. I love that he's having fun. I Love that he's taking shots. Don't they have an AI lab? To be clear, they were working on Einstein for a while, but I think that they are very much happy to be a rapper, happy to be a buyer of LLMs at this point. It certainly has penciled out that way. We can talk more to Mark about that. It's feeling Good. Stock's up 10% in the last five days. Here for. Best salesman history. What. What do you think about Michael Burry? The whole industry needs a $500 billion IPO ASAP. Said any. Everyone that knows anything knows this. OpenAI is the next Netscape doomed and hemorrhaging cash. Microsoft is still trying to keep it afloat while keeping it off balance sheet and sucking out the ip. So why do they keep getting funded? The whole industry needs a $500 billion IPO ASAP. The whole industry needs to go away and sit in a corner for a couple of centuries and think about what's done, says Jeffrey Miller. Wow. I don't know. This is. People want extreme takes. This post would go super hard if you don't understand the Microsoft OpenAI relationship. Yeah, but. Yeah, it's just. It's just. What was Netscape revenue? What was Netscape, I remember correctly, they. IPO'd with like 15 million of revenue. That feels a little bit different. In Netscape's peak revenue, 1997 was 500 million. I mean. We'Re up at 20 billion this year for OpenAI. Yeah. Their total revenue before IPO for first two operating quarters, Netscape reported total revenue of 16.6 million. Just odd. I mean, did Netscape have like an enterprise business? That's Mark. Yeah. I don't know. I mean, even if there's a. Yeah, yeah. Like, even if there's more commoditization, like it's going to take. It's going to take time. And I don't know, the dynamics of the competition here feels like there will still be a lot of value, even if. Even if it is somewhat commodity infrastructure. It's like, you know what else is commodity infrastructure? Like aws, gcp, Azure, like you get like a server with some hard disk on it. Like that is commodity. And yet they all have 30% margins and they're all massive businesses. And when AWS, when Amazon broke out AWS as its own line, it was like an IPO of its own company because it was so big, it was such a massive business and it should be completely commodity because it's just servers in a data center. And yet. Yeah, it's Interesting too. I mean Burry has positioned himself of just hating any company that's overheated. But ChatGPT having close to a billion weekly actives and ultimately even if they just compete in search. Right. At least it's a multi trillion dollar opportunity. Whether or not they fully execute against that is another thing. Are you buying the IPO timeline of Cohere, then Anthropic, then OpenAI or do you think they will be sorted in a different pattern? What was the order? Cohere first. That's Aiden Gomez's foundation model lab. He is of course on the transformer paper also a dripped out technology brother. That's right, with a fantastic set of outfits, then anthropic and then OpenAI. OpenAI third. That's the rumor. Hard to say. I have no idea how the market will react to cohere. Yeah, Cohere has shifted more into the business, the enterprise market. Fully enterprise. Not in the time line at all. That could very well be intentional. But it would be. It'Ll be interesting how much excitement they can build around the ipo. Assuming they're losing a lot of money. Yeah, assuming the enterprise is really competitive and they're competing with anthropic and open AI and open source models. If the narrative changes to the point where you know this, this idea of, of every enterprise is going to need a basically a custom LLM that's fine tuned or pre trained on their data. Like what AWS CEO Matt Garman was talking to us about, about these pre training checkpoints where you can go in and say okay, I have a business and I absolutely do not want upload my data to Anthropic. I don't want to update it, I don't want to upload it to OpenAI, but I want a model that understands my business's data at a core level in the pre training step. And Cohere can offer that and show traction there. That feels like that could be the next wave. We've talked to a number of founders that are building those sort of like custom pipelines and there's not really a public company that is even anywhere in that narrative. And that narrative does feel very nascent. It feels like it has not percolated through the public market markets yet. So I don't know, it could be, you know, exciting for them. Well, let me tell you, they've only raised 1.4 billion. I did not realize they raised, sorry, 1.5, 1.5 billion. 1.54 billion. Very modest amount, especially when you have seed stage companies. Raising half a billion. We got one of those on the show today. Dean Rao, we're excited joining. Let's go later. He kicked the hornet's nest on the timeline. What do you do? A little bit. People were, people were pushing back because. Of his, his, his racing history. No, no, no. Oh people, people. Show respect. Love that. Show massive respect. No, but I think, I think Jack at SLOW was saying that Nvidia was valued at 500 million when they IPO'd. Of course that was very long time ago. But anyways, people, people are just saying like it feels pretty potentially overheated. But we have Naveen coming on the show today and I'm excited to hear about the opportunity from him. I'm ready to say it's overvalued unless he's built a company previously and sold it twice. Maybe. Yeah, twice. If he's done that then he's off the hook. But he's gotta prove it to us because if he's just some new grad. Some dropout like a Waterloo. Yeah. If he's just like. If he shows up on the show and he's just like yeah. I had this idea in my garage, I didn't go to school, I've never done in business before and now then I'm going to call him out and I'm going to say you're overvalued, you raised too much money. But if he can prove to us that he's done it once or twice before, then I'll let him off the hook. I'll let him off the hook anyway. I think that's fair. Let me tell you about Privy. Privy makes it easy to build on crypto rail, securely spin up white label wallets, sign transactions and integrate on chain infrastructure all through one simple API. Let's read the semi analysis post says important a common misconception about OpenAI's upcoming custom chip is that since it's a custom chip it won't be flexible and will be a data flow machine. OpenAI has recognized that the 100x efficiency gains for training and inference happen at the algorithm layer and the hardware chip needs to be flexible enough to accommodate these algorithm changes. We went from just pre training transformers to now doing RL post training on transformers. We went from dense transformers to mixture of expert transformers and soon ultra sparse transformers that will have four active experts per token out of 2,048 total experts. We went from casual MHA attention to MQA and GQA to attention, sync sliding attention to now even learn sparse attention despite AI tourists think the chip OpenAI is building with Broadcom will be far more flexible than TPUs. Interesting. Despite most of OpenAI's chip team being poached from Google's TPU team. The semiconductor horse race continues to fascinate me. And I'm so glad that we have the good folks over at Semianalysis coming on the show regularly to help us understand it. I always enjoy talking to them. And we have. Two folks from Semianalysis joining the show next week, so we will be having a great time. Let me tell you about Turbo Puffer, serverless Vector and full text search built from first principles and object storage. Fast 10x cheaper and extremely scalable. Of course. That's right, it's scalable. A hedge fund was ordered to pay a bonus to a trader who made 97% of its revenues. This is hilarious because when I read this at first I thought it was he had made like his target bonus was let's say 10 million and he brought in 9.7 million. And so he was 97% of the way to his bonus. And they were like, you didn't hit your bonus, buddy. You don't get the bonus at all. And I was like, oh, that sucks. But like that's kind of the deal that kind of makes sense. But Evolution Capital Management has to pay him because. Yeah, so a hedge fund that was sued by a trader for refusing to pay a performance related bonus despite him making 97% of its revenue, has been ordered to pay him 5.4 million plus interest by the High Court in London. When I read the headline of the story, I expected it to not be in the mid seven figures. But would you expect. I mean I was hoping at least 8. Not getting a G400 off of this. Not even close. Not even close. Robert Gagliardi sued his former employer, Evolution Capital Management in London, alleging that it actually acted in bad faith by denying him a $7.5 million discretionary bonus after he had generated more than 60 million for the firm. Wow. Wait, does this mean that every other trader lost money was just something like that? I don't know. Heavily. So he's at the fund brief stint, he made 12 million, he was already paid 7 million, including a $625,000 signing bonus. Base salary was 425k and the $6 million new issue bonus. So he needs to get out of this. He needs to go into AI research ASAP if he wants to be putting up real numbers. This is rough, but there are some harsh words here. He said Gagliardi, a block trading specialist alleged that he was told in early 2021 that a return of $10 million over the rest of the year would be an excellent result. When Gagliardi asked the fund's founder, Michael Lurch, for the payout in 2022, he responded, I'm not going to pay you the bonus if you sue me. So Gagliardi did, and he won. He did. He did. Although Evolution did not dispute his extraordinary returns, the fund argued that the damage done to its reputation as a result of dealing with a US SEC probe into some of Gaglieri's trades at his previous employer outweighed his performance and that the bonus did not need to be paid because it was discretionary. There's no question that Mr. Gagliardi made exceptional profits for the funds, and Mr. Lurch frequently praised his performance in that respect. Had Evolution properly performed the contract, Mr. Gagliardi would have received a discretionary bonus of 5.3 million. The judge also said that Gagliardi was Lurch's prize asset at Evolution. That's a good goal for everyone that, you know, maybe working at a company become the prize asset. Hopefully that doesn't result in a lawsuit like this, but they're feeling. They're feeling. Gagliardi and his lawyer are feeling very vindicated. Well, we have our next guest here in the Restream waiting room. Let me tell you about Gemini 3 Pro. First, Google's most intelligent model yet. State of the art reasoning. Next level vibe coding and deep multimodal understanding. We have Aaron Ginn, the CEO and co founder of Hydra Host. Good to see you. It's been too long. How are you doing? Welcome back. Been all right? Yeah, Yeah. I just got back from Southeast Asia. Oh, yeah? Yeah. And despite what some of the critics were saying, I wasn't smuggling anything I shouldn't have been smuggling. Getting accused of smuggling, that is wild. Well. Give us the update on your interpretation of the H200 deal. Where things landed, where you think the next debate points will be. I'd love to hear that. Yeah. Like the. I mean, as you know, I've said a number of times that the. Even though I am a China Hawk, that we have to realize that Trump fundamentally is a pragmatist. Sure. And he is not interested in the ideological warfare. And I think that him taking over the Republican Party proves that. And the orientation of reshoring to America, which y' all know, like I wrote one of the first arguments for a new Monroe Doctrine. So now you saw that the Department of War has been.

We came to this world. To shape a future. Reaching to feel the new. Sam. You're watching TVPN. Today's Tuesday, December 9, 2025. Just a few days until Christmas. We're so excited. We're live from the TVPN Ultradome, the temple of technology, the fortress of finance, the capital. No travel until Christmas, baby. No travel until the new year. We're in the ultra dome, hanging out. We're monitoring the situation. We're monitoring the Paramount, Netflix, Warner Brothers situation because this is one of the most fascinating deals. The deal has gone hostile. Paramount has launched a hostile bid to acquire Warner Brothers. It's an all cash offer. 77 year, 9 billion. Really? Really. And it's a fun one because I feel like it's. It's obviously tech adjacent, but it's not. It's not the story that people have been monitoring all year. We've been talking about foundation models. That whole story has just gotten a little bit. Legacy media, they got tired of not getting enough attention. Exactly. Let's spice it up a little. Spice it up. Here's something new to learn about. So everyone's having fun. Everyone's learning new things. And I could tell because when I came in today, I had a bunch of questions that people were throwing at me about how all of this works. Why isn't Warner Brothers just going with the highest price? When I sell a stock, I don't care if Citadel or Jane Street's buying it. If I'm selling 80 bucks a stock, just give me the best price. But when you're selling an $80 billion company, there are other considerations. And it goes beyond just maximizing shareholder value. And so I wanted to break down a few of those. So I did that in today's newsletter. We can run through that and then we can run through some of the news. Perfect. But first, let me tell you about ramp time is money save. Both easy to use, corporate cards, bill payments, accounting, and a whole lot more all in one place. So. This question, you know, it seems obvious. The board has a fiduciary duty to maximize shareholder value. That's legal requirement to take the higher offer. And yet that's not what's happening. Like yesterday, we saw that Ellison and Paramount came in with over 100 billion. Of course, that included the CNN, the TV assets. But even when you broke it out, it seemed like it was very clear that Ellison was willing to pay more money and make a higher offer. So under what circumstances can a board whose job it is to maximize shareholder value, not take the higher offer? It can't just be whoever we had the better dinner with, Right? Which is part of the news, of course. They went out to dinner, and Paramount CEO David Ellison sent a text to David Zaslav, who's running Warner Brothers, after they made a hostile bid today to buy Warner Brothers. And David text the other David and says, david, but I guess he misspelled his name, even though that's his name, which is just funny. But anyway, he says, I appreciate your underwater today, so I wanted to send you a quick text. Know that despite the noise of the last 24 hours, I have nothing but respect and admiration for you and the company. It would be the honor of a lifetime to be your partner and to work and to be the owner of these iconic assets. He's talking about Foghorn Leghorn. He's talking about Porky Pig. He's also talking about Batman and Superman, obviously, and Harry Potter and Lord of the Rings and a million other iconic assets, which is true. If we have the privilege to work together, you will see that my father and I are the people you had dinner with, which I like that. I think that's cool. They had dinner. It's a fantastic text. Yes, yes, yes. It's a great one. So, quickly, let me tell you about Julius AI, the AI data analyst that works for you. Join millions who use Julius to connect their data, ask questions, and get insights in seconds. So there are two main reasons why you don't just take. Why your size might not be size. Directors at a company like Warner Brothers, they have to maximize shareholder value, but maximizing shareholder value is an expected value calculation. So if you come in with $100 billion offer, and I think there's a 75% chance that you're gonna deliver that, and someone else comes in with an $80 billion offer, and I think there's 100% chance that they're gonna deliver that well. The expected value of Your bid is 75 billion. The expected value of their bid is 80 billion. I go with the 80 billion. Even though it's like a lower headline number, it has a higher expected value. How much of the calculus do you think is just a deal that will actually get done? A deal that is going to get. So a deal that doesn't get done could still have value because of a breakup fee. So you could, in theory, go into a deal that you know is impossible. And the example that I gave is, what if bytedance came in and they were like, hey, you know, we're $400 million media companies. We'd love to own Porky. We'd love to own these iconic assets. Let's pick up Warner Brothers. We'd love to own CNN too. You know, who knows what we'd put on the TV? We already own TikTok. We already own TikTok. Why not CNN also? Why not Shark Week? Not Shark Week. They'd have to carve out Shark Week in my book. But obviously the government would block that. And we have, we have cfius, which is an organization in the US Government that determines whether or not an international buyer can take an American company because of intellectual property. All sorts of geopolitical considerations. We don't want another country cornering a market on a really key piece of the supply chain like the Nvidia H200 for example, which we might be talking about later in the show with Aaron Ginn, co founder of hydrahost, repeat guest on the TV Pan show. But then there's also just the ftc. So certain deals, like I gave the example yesterday, like Disney, Disney would be so blocked, I think that, that that deal doesn't even get kicked around. No one even talks about it. And because it would get blocked by the ftc, so it doesn't even happen. But it would maximize shareholder value. If Disney came and said, yeah, we'll give you a $10 billion breakup fee and we want to try and buy you for 200 billion or a trillion or 200 million. Like you would immediately say yes, because you just want the breakup fee. Yeah. But having being in this turmoil and being in this limbo and not being able to sign deals with other. That has an opportunity cost. Right? Yeah. And so you might want to back off of that. And so basically the, yeah, look, you. Look back at the FIGMA acquisition, Adobe paid a billion dollar breakup fee. So it's effectively like non dilutive financing for figma that has a happy ending because they were able to, you know, re accelerate, get out into the public markets. But there was another situation where they went through a rough patch and actually really needed that capital to get through. Yeah. So there's basically two buckets of risk that I think most dealmakers would be considering in this situation. First is financing risk. So will they come through with the money and what money are they paying with? Because the initial offer, this is the history, there's actually been six offers put forward by Paramount. David Ellison is putting on a clinic of just not taking no for an answer. Because all the way back on September 14th he offered $19 a share and 60% of that was cash. So 60% cash offer. Then September 30th, two weeks later, he comes back $22 a share and ups the amount of cash to 66.7%. Zaslav is doing a masterclass in making your opponent negotiate against themselves. 100%. It's actually great. October 3rd, $23.50 a share, 80% cash. Then November 20th, 25, 50 a share, 85% cash. Then finally, December 1st, $26.50 per share, 100% cash. December 4th, $30 a share, 100% cash. Now, why does this matter? Well, it's because you get locked into owning the shares of 40% of the shares is paramount. And then while it's closing, the stock trades down, you wind up getting less value. And so the 80 billion today might wind up being 70 billion later. And that doesn't maximize shareholder value. And so there's also just the question of, like, can you actually marshal the cash? Just like if somebody comes to you and you're about to sell your house to them and they say, yeah, I'm definitely going to get a loan for this. Well, that's a financing risk. Maybe they don't. Maybe they back out of the deal. That's why cash offers often times. Exactly. If you can prove that you have the cash, that makes difference. And I mean, he's effectively gone and done that because he's teamed up with Jared Kushner reportedly and gone around the world, got a whole bunch of different sovereign funds. And just really people, anyone with big deep pockets has kind of said, yes, I'm down to come along on this ride and put up a bunch of the capital. The capital has been marshaled. It seems like it's ready to go. And then also you have Larry Ellison, who has three times as much money, I think, than the whole deal value something that 275 billion to his name. And he's asset together, not cash assets, not cash. But he should not just, I know they want it, but do not market. Sell your Oracle position, please. Yeah, I'm actually, actually going all cash now. But, but, but he's, he's effectively acted as backstop. And so, and so everyone who's come in and said, okay, who. I guess I'm good for 10 billion of that 80 billion, but only if everyone else is in. Larry Ellison reportedly has come in and said, well, you know, if there's one person in the bunch that backs out of their slug at the last second, I'll jump in and get that. He's not saying he's going to put up the whole 80. He's just saying he's backstopping it. Yeah. And that's the term that's been thrown around. So. So David Ellison feels very, you know, good, because he's done what Zaslav wanted him to. He said, hey, they wanted an all cash offer. I brought them an all cash offer. I brought them an offer that's higher. And we believe that it's. There's no reason why factor number two should come into play. And factor number two is regulatory approval. And so there's been this question about will Netflix get approved? If there was a different buyer, it would have even more regulatory risk. And so you don't want to. Even if the price is higher, you don't want to accept a higher price with a lower chance of actual, actual conversion. Right. Bobby thinks NLE Choppa might up his offer to 100k of cash and get back in the mint. In that scenario, at least you know that the 100k cash is real and it's going to be delivered on the. Day of close, maybe in actual cash. Probably. I think that was the. That is funny because that was the joke of that meme was that he was giving like actual cash. When we're saying cash offer here, of course we are not talking about physical cash. Cash. Hey, you never know. Are you as well over there? Yeah. Is Tyler doing money spread? What does he got? Did Christmas come early for you? So I had a question though. What makes the Paramount offer like, hostile? Is it that it's like reacting to. The Netflix office at the Netflix board already or, sorry, the Warner. Warner Brothers board went with Netflix and they're coming. It's hostile because they're saying, screw the board decision. We want to go to the shareholders. Okay. Yeah. Because the. It sounded like they had a nice dinner and the texts were like, kind. Yes, yes, yes. No, no, hostile. No, no, no. Hostile in the sense of like, of like not accepting the final, like the final decision there and then. And then also there is a. There is a point where you can appeal directly to the shareholders and make the. You can actually make the legal case that the board is not acting in their fiduciary duty and in their. In the interest of shareholders. Because. But again, these things need to be argued in shareholder lawsuits because it is fuzzy because, like, if I say I'm offering you 100 billion and someone else is offering 80 billion and the board says, yeah, but that hundred billion has only a 75% chance of going through, you have to discount that. Now the board says it's worth 75. But whoever made that offer can say, no, we're actually good for it. You should apply a 10% discount rate and treat our offer as if it's 90% or 100%. Good. Right. And so, and so all of those arguments, those can be made in the court and the shareholders can, you know, react to that. And if the shareholders align and say, yeah, actually, we think we'd be getting more money here with this, with this deal, then they can push back against the board at a certain point. But, yeah. Friend of the show, high powered media exec says Zaslav has been architecting this situation for the last 12 months. Sitting pretty and he's a wily old fox and has what he has been shooting for. Two heavy hitters fighting over a deal. Don't doubt there will be a couple more turns here with the price or even another bidder coming in sideways. Don't quote me. Which is remarkable because the price is so high already compared to where it was six months ago. Zaslav. It really feels like it's coming together that he will be remembered as an incredible business executive for this deal. I want to know more. I'm still learning about this particular industry and so I've been. I've been having fun digging in, but. Possible deal guy of the year. He's in the conversation. He's definitely in the was. Sam looked like he was running away with it over the summer. That's true. Early fall and he lost ball control. He lost ball control, yeah. Well, quickly, let me tell you about Fin AI, the number one AI agent for customer service Automate, the most complex customer service queries on every channel. But. There is a wrinkle that people haven't really been considering, which is the fact that Warner Brothers Discovery holds a massive, often underrated vault of masculine cinema. And if this falls into the wrong. Hands, what is masculine cinema? Masculine cinema being dudes on film. So they own. They own a bunch of Clint Eastwood films. Dirty Harry, Magnum Force, the Enforcer, Heartbreak Ridge, American Sniper, Letters From Iwo Jima, Unforgiven, Gran Torino. They also own the classic Steven Seagal Run. Basically every movie where Steven Seagal was a massive theatrical star before going to direct to video. We're talking about above the Law, Hard to Kill, Marked for Death, out for Justice, Under Siege, which is Die Hard, On a Boat, Under Siege, Two On Deadly Ground, Fire Down Below, Exit Wounds. They also own Rambo, the host of Wrestler Stallone, the Cobra Wing Cobra, Demolition man, the Specialist Tango, and Cash Bullet to the Head, get Carter, the 2000 remake. They also own Mad Max, Lethal Weapons, all of the Lethal Weapons. The Matrix, blade, Mortal Kombat, 300, Rush Hour, Pacific Rim and a number of dad and military shows. Band of Brothers, the Pacific Generation Kill. There's a whole host. They should release the full DVD set. Just called Guys being Dudes. Yes. They also own some Jason Statham properties. The Meg franchise where Jason Statham fights a big shark. Whoa. Wrath of Man. I guess most of the Guy Ritchie films are actually with Lionsgate or mgm but there are a variety. And I think this could be a, you know, a big political hot button of what happens to the Rambo franchises. The Hollywood hunks. No one's talking about the Hollywood hunks. Everyone's focused on the Looney Tunes characters. But if the Hollywood hunks fall into the wrong hands, it could be. It could swing our entire culture potentially. Anyway, let me tell you about. Apparently Michael on our team worked on Wrath of Man. Oh really? No. First job. Very cool. Very cool Dan. To automate compliance and security AI that powers everything from evidence collection and continuous monitoring to security reviews and vendor risks. You're hiding in order to amounts of Middle Eastern funding sources in your takeover bid, are you not says boring business and not hiding it. It's out in the open like this is. This is. This is like the playbook for these deals. Double. Double that of the amount of equity that the Allisons are putting in. Putting it around. Yeah. I mean I think everyone assumes that there'd be a lot of Middle Eastern funding in this. That has become the standard funding instrument these days. I don't know like EA Games. We just followed that story a little bit. That was Obviously it's like 93% Saudi money and that is part of the modern deal making playbook these days. Every. I'm pretty sure most of the. A lot of venture funds have raised money over there. A lot of the big AI companies have raised money over there. Like that seems like a foregone conclusion that if you need money you go where the money is. Why do bank robbers rob banks? Because that's where the money is. Well, Dylan Byers says. David I appreciate your underwater today so sent you a quick text. He's talking about the dinner. Oh so that was the correct. That was the misspelling D A I V D. He must have been texting very quickly. It does give you. Let's. Let's turn this into a copy pasta. Yes. I enjoyed the dinner. So Exec Sum has an overview of who's actually putting in the Money here. Filings reveal that Paramount's bid for Warner Brothers is backed by equity commitments including 11.8 billion from the Ellison family, 24 billion from Saudi Arabia's, Qatar's and Abu Dhabi sovereign wealth funds, 1 billion from China's Tencent, additional commitments from Redbird Capital Partners and Jared Kushner's Affinity partners. That doesn't add up to 100 billion so there must be cash coming from somewhere else. Maybe there needs to be some more reporting until. Oh, they did give an updated version of this post but doesn't quite add up. So will be interesting to follow where you know how the deal comes together. And we have a few more people joining this week to help us break down the deal. Ben Smith from Semaphore tomorrow and then Dylan Byers. Oh, Puck on Thursday in person. I'm excited. Oh, he's coming in person. Yes. That's great. Well, let me tell you about numeral compliance. Handled numeral worries about sales tax and VAT compliance so you can focus on growth. Trump says that the United States will allow Nvidia H200 chip sales to China and get a 25% cut. This is a pretty big change. I mean we were talking about. Not selling. I mean, I guess it's not Blackwell, right? It's Hopper. So we're still a generation behind. But was this, it was like a pretty nerfed chip before. We're getting less nerfed. We seem to move, we seem to be moving in a more thumbs up direction. More. Let's actually get the chips over to China at the same time. AI is fake so it doesn't matter. Right. That's the take, the pro China take is that it's like SaaS. It's SaaS. Or it's not like this nuclear bomb that's going to destroy everyone. So it's harder, it's getting harder to make the. The Manhattan Project argument. Right. And it feels impossible to make the argument that we're going to get them addicted to the US AI supply chain and they'll never develop their own capabilities. And so we want that. I disagree with that. I think that, that, I think that that argument actually holds. That argument holds for sure if we. In, in the next. I mean, I totally disagree. I think they're, I think China's smart enough to know they don't want to be dependent on any foreign country to produce any critical. Yes. But I still think there's enough of like a market force within China that if we flood the market with cheap Nvidia chips. It'Ll just be expensive for them to keep propping up their local industry, even if they are aware of it, even if they know that they have to. It's a cost. And it's something that a lot of AI researchers over there will just say, you know what? I'm already. It's so much easier. The Nvidia ecosystem is so great. I'm just going to stick with that. I don't know. I'm a little bit skeptical that, like, there's no advantage to selling Nvidia chips there. I've become more. I've become more receptive to that argument, particularly. Even though you have not. But that's fine. Tyler, what were you going to say about this? Yeah, I was just going to say, I mean, we kind of joke about this, but it is kind of crazy the extent to which you can basically bake down all, like, AI policy questions to, like, if you are AGI pilled. Yeah. Because, like, if you are. If you're Dario. Yeah. I mean, you shouldn't be giving stuff to China. You also shouldn't, like, if you are a GP AGI pill, you should be thinking about safety, all these things. Like, we had Keith, her boy on. He doesn't seem super AGI pilled. He's also like, oh, safety is a hoax, et cetera. Stuff like this. I think basically all of these questions, you can just. Well, it depends, because if you're super fast takeoff pilled and you're super like, asi, it's like. But you think that it can somehow be contained in whoever gets it first, you actually don't mind if China's six months behind you, because as soon as you hit that inflection point, you're 1,000 years ahead. And so you might as well just do whatever it takes, fund whatever, make as much money as possible, as long as it accelerates you to be the first one to hit that inflection point. So there is an argument where you can be extremely AGI pelt, even super intelligent. Well, Truth Social. Truth Social is creating a prediction markets. Product within the app to say AI product. Did they launch an AI thing last week? I think that was fake news. Oh, that was fake. Okay. But it's possible Trump is going to use the prediction markets on Truth Social in order to kind of set a policy going forward, say, H200 to China. Good or bad. That's wild. I mean, that's not that, like Robin Hanson says that that's like a good idea. That was the whole. That was the whole. That was the whole pitch. Originally, there is definitely a bull case, you could say like a SEAL man for that. Yeah, there's always the steel man. Yeah. I mean, it's a helmet. Where's your helmet, John? Yeah, yeah, put on the helmet. Well, let's read through this Wall Street Journal article on the details of the new Nvidia deal. But first, let me tell you about figma. Think bigger, build faster. Figma helps design and development teams build great products together. So notable that this seems to have been completely priced in already because Nvidia is actually down half a point today. Yes, it's been such a dramatic story all year, and yet it's always felt like a complete footnote in the overall financial performance of Nvidia. It's never felt like, okay, yeah, if this goes Nvidia's way, they're going to double their revenue or double their profits or like, you know, really move the needle. They're growing so fast that, you know, I don't know how many we're going to get into figuring out how many H2 hundreds they sell, but they would need to sell a lot to actually move the needle on this behemoth of a business. What is the world's largest company in the world? So President Trump said he would let Nvidia export its H200 chip to China and that the US would receive a 25% cut. His latest bid to make money for the government in an unusual agreement with a private company. I have informed President Xi Jinping of China that the United States will allow Nvidia to ship H200 products to approved customers in China and other countries under conditions that allow for strong national security. 25% will be paid to the US of A. The move is a boon for Nvidia, which has fought for months to maintain access to the world's second largest economy. The company had agreed earlier this year to give the US 15% of China's sales from a lower performing chip, only for the Chinese to scuttle those plans as part of continuing trade talks between the two sides. Chips from the world's most valuable company have become a prized geopolitical tool. The H200 has higher performance than the H20 that Nvidia was previously allowed to sell. But this isn't as powerful as the company's top Blackwell products released this year, nor the Rubin Generation Chips generation of chips coming next year. The move follows a meeting between Trump and Jensen Huang last week where the pair discussed H200 exports. People familiar with the matter said Nvidia shares added nearly 2% after hours. That's not, not too bad. Even with the US Government taking a cut, the decision could be worth billions of dollars in sales to Nvidia, which enjoys comfortable margins on its AI chips. In the most recent fiscal quarter, Nvidia reported gross margins of 73.4% on $57 billion in sales. That is crazy. They can totally afford 25% for the big guy. I have no idea if it'll be legal. We'll have to figure out if this gets approved because in general, the United States does not import enforce export duties. That's not something America's done historically. We had some folks on the show to contextualize that and give us some background when this was first floated. Feels like it is happening now, but we'll see where it lands in August. Nvidia CFO Colette Krass said that if geopolitical issues subside, the company could ship between 2 billion and 5 billion of chips to China per quarter, which increase. Which could increase if orders pick up. So they're doing. Yeah, right around that quarter. Lutnick was on cnbc and this was the quote that originally ticked off the Chinese. He said, we don't sell them our best stuff, not our second best stuff, not even our third best. And he said, you want to sell the Chinese enough that their developers get addicted to the American technology stack. That's the thinking. And of course they CCP basically immediately said, we don't want any of them now because we're offended. But it does seem like, yeah, I don't know. We'll have to have Bill Bishop or someone on the show to contextualize how China is receiving this information, whether or not they will actually buy it. Of course. We are going to be joined by Aaron ginn in just 30 minutes, so stay tuned for that. Also, let me tell you about Linear Meet the System for modern software development. LINEAR streamlines work across the entire development cycle from roadmap to release. So the exports could help Chinese tech giants that have struggled to get top chips to train their models. Huang has argued Nvidia should be allowed to compete in the Chinese market because China has many of the world's top AI researchers and the US should want them using American technology. If you're, you are not going to replace China, Wong said at an event. At a think tank event, Trump said that the government would take a similar approach to exports from Nvidia competitor AMD as well as intel, in which the government now has a 10% stake. The approval comes just weeks after administration officials, including Secretary of State Marco Rubio torpedoed a push from Nvidia to sell a slimmed down Blackwell chip to China before a recent trade meeting between Trump and Xi Jinping. Some officials, including Aizar David Sachs and Commerce Secretary Howard Lutnick, have backed exporting the H200 because it could be a good compromise that allows Nvidia to compete with China's Huawei Technologies without vaulting China past the US in AI. People familiar with the discussion said, I don't know. It's such an interesting question. I still wrestle with. We exported a ton of Teslas to China and BYD and Huawei have now arguably completely leapfrogged. Completely leapfrogged. So they did not become addicted to the American market. That was the point I was making earlier. You can make the art, like almost with every single product. They've said, we'll work with you to make this thing that you want to make. We're really good at making things. And then they ultimately just make a better version of said product and make it for cheaper. Yes. And in the case of cars right now, we're obviously not allowing these cars to be imported into the us but they can simultaneously say, we're happy to keep making you these things. We're also going to compete with you directly and constantly try to be better than what you do. I just wonder if you play it back and you don't allow Tesla to export the amazing Model s or Model 3, does that slow down BYD's development of their car or. Or Huawei's or Xiaomi's development? Yeah. My understanding is they were able to basically get a paid education making Teslas and they were able to leverage that into making. But that's about making the car there. If you don't make the car there. Which is because the Nvidia has a Shanghai. But that's research not made there. That's not made there. Like it's very different when you're saying, okay, we're gonna go and produce this product there and like you are gonna get educated there. And the world, I think it's different. But you remember we've gone through this before, China's like five year plans to create a domestic chip industry. You've been doing five year plans for 60 years. I know they're building. That's what I'm saying. And I would argue that it's working. They're not caught up, but they are certainly made massive amount of progress. They've made more progress than any other country on earth. Yeah. I don't know. What do you think, Tyler? I think if you want to think about getting the Chinese addicted to our chips, how addictive are the chips? Yeah. Because if you can make a comparison between, like, you have the Nvidia chips and then you have the Chinese or like tpu. Right. Because the TPU is like, in some ways it's, it's harder to use. The open source is not as good, but if it's just more economical, like if the actual hardware is just like a little bit cheaper, then it doesn't really matter how worse the software is, people will eventually move to it. So it's like the. I don't think there is actually that much like soft power in the kind of general open source stuff like that in regards to, like Nvidia, it doesn't seem very addictive outside of it just being cheaper. Yeah, yeah. I mean, at the same time, like, like we are America is very much like addicted to Chinese solar panels right now. Like, the Chinese solar panels come here, they're cheap. And so we don't wind up buying or building a domestic solar panel industry because just to get anything off the ground, you have to go in and say, okay, we're, we're going to deal with having no margin forever, and no venture capitalist can underwrite it and no private equity firm can underwrite it. And so it just doesn't really happen. Yeah, but the Chinese, they do do that. I guess they do, they do bear the cost more, so it doesn't matter. But it's like they wouldn't have to bear the cost. So we're, we're imposing a little bit of a cost on them. It's like they actually have to pay it. Whereas they don't. Whereas they wouldn't if we didn't do that. Right. Because if we don't go in and we don't sell them Nvidia chips, then they're like, okay, well, it makes economic sense to back Huawei and fund them. Whereas now they're like, it's not economically logical to back Huawei. But we have to because there's so much pressure from better chips flooding the market like Nvidia chips, that if we don't, if we don't fund Huawei uneconomically, if we don't lose money on Huawei, like, we will never get to. We will never get, we will never be competitive. Yeah. I think it seems to me that like, either way they're going to fund it. The government, the Chinese government. I think that's George's point too, really. I don't think it actually matters that much. Yeah, I don't know. Well, let's. Let's move on. But let's also tell you about adeo, the AI native CRM. Adio. Builds scales and grows your company to the next level. The United states will allow Nvidia's H200 processors its second best. So this is a. This is a change from what Ludnick said, right? Ludnick said we're not selling the best. We're not selling the second best. Turns out we are selling the second best now. Maybe we'll sell the first best soon. Who knows? Well, we should play this clip from Patrick o' Shaughnessy's fantastic interview with Gavin Baker on Invest like the Best. He's talking about space data centers. We had Dallian on the show yesterday talking about space data centers. Delian still thinks Gavin. It's too far out. Gavin knew we needed a new narrative. He said after the fateful BG2 episode ended, the AI trade. We needed a new trade. And now we're getting the space data center trade. Okay, let's hear the space data center. I think the most important thing that's going to happen in the world in the next three to four years is data centers in space. If you think about it from first principles, data centers should be in space. What are the fundamental inputs to running a data center? They're power and they're cooling. And then there are the chips, the inputs to making the tokens come out of the magic machines. So in space, you can keep a satellite in the sun 24 hours a day. Pretty cool. And the sun is 30% more intense. And this results in six times more irradiance in outer space than on planet Earth. So you're getting a lot of solar energy. Point number two, because you're in the sun 24 hours a day, you don't need a battery. This is a giant percentage of the cost. So the lowest cost energy available in our solar system is solar energy in space. Second, for cooling in one of these racks, a majority of the mass and the weight is cooling. And the cooling in these data centers is incredibly complicated. You know, I mean, the H vac, the CDUs, the liquid cooling in space, cooling is free. You just put a radiator on the dark side of the satellite. It's fucking gold. And it's as close to absolute zero as you can get. So all that goes away, and that is a vast amount of cost. And then that's like, isn't it? Isn't how this Isn't the problem a. Bit more complicated than that? The dissipating heat and space, this is. What everyone's fighting about. I'm not exactly sure how it. How it plays out, like, the physics of it, but a lot of people are saying, like, yes, it's in a vacuum, but because it's in a vacuum, you can't just. You can't. Like, it's not as passive of cooling. It's more of a material science problem. Like, there's more physics. Physics problems there also. Like, I don't know, I feel like you might need some. I feel like you might need some batteries, because I think there are orbits that are effectively in the sun all the day, but I think that the Earth's shadow, like, hits them not like, every day, but, like, every once in a while. You'Ll be caught in a shadow almost in, like, in a. What's it called? I don't know, some sort of. You just get caught in the shadow. Not every day, but, like, every 12 days or every 30 days or something like that. And so in that scenario, I mean, maybe it'd be fine if you're doing inference and you don't and you can just turn off the whole system for a little bit and there's some redundancy. I just texted our resident spaceman, Dalian. Yeah, I said, answer this. I just said, quick take. Yeah, on this. It's urgent. It's urgent. Well, we'll get back from him. There's more to that clip, though. We should play this because there's a funny part in here. What, are you going to connect those racks? Well, it's funny. In the data center, the racks are over a certain distance connected with fiber optics, and that just means a laser going through a cable. The only thing faster than a laser going through a fiber optic cable is a laser going through absolute vacuum. If you can link these satellites in space together using lasers, you actually have a faster and more coherent network than in a data center on Earth. Okay. For training, that's going to take a long time just because it's so big. But for inference, let's think about the user experience. When I asked Grok about you and it gave the nice answer. This is crazy, though. They've done a podcast together five times. Why is he asking about Patrick? I was like, what is going on? Went to some sort of Metro aggregation facility in New York, probably within 10 blocks of here. There's a small little Metro router that's routed those packets to a big XAI data center somewhere. And then the computation was done and it came back over the same path. If the satellites can communicate directly with the phone and Starlink has demonstrated direct. To cell communication, this whole thing. I would like to know how much use space X exposure Gavin has, because if this does become a dominant, at least narrative over the next few years. Yeah. A lot of the people that we know that have computers in space are not exactly bullish on data centers in space on a short time horizon. Yep. We have not. We had the founder of Star Cloud on. He's obviously bullish, but I don't believe he has any satellites in orbit. Yep. We've had Delian on. He puts things into orbit, he brings him back down. He's talked about having. He technically has. I mean, Dell is a hilarious position because Founders Fund has a huge position in SpaceX obviously. But also Delian has had to do the hard work of like getting something in space and probably understands like how much of a grunt work it is and how much of a grind it is to get even just like something as small as like the pod into space. And then he's also seen at Crusoe like what a real data center looks like. And the idea of putting something so massive into space, like it's bigger than the Hubble, it'd be bigger than the iss. What are you thinking about the space data center thing? Have you evolved on this? Yeah, I mean, I don't know. To me it seems just kind of impractical. Like if you think about, if you're doing inference or something, like how much of the cost of the token is actually in the energy. Right. Because I feel like one of the main benefits of having the data center in space is that energy is basically free. Right? It's right there. Sure. But if you think about like, I think Casey Hanwer has talked about this. Like if you think about an inference, when you're doing inference, the portion of the cost that's actually the energy, it's like very tiny because it's like the chip is expensive or doing like it's not that expensive. So it's like, okay, the cost gains overall are like not that great. And then it's like, well, what if you want to change the chip to the next generation? Or what if you need to do some kind of like mechanical work? Are we ignoring how broken Cluster would take some sort of intervention to fix as well as the cost per kg in fuel costs? Yeah. Just for context. So Gavin was at Fidelity and led a $900 million investment in SpaceX at a $12 billion post money valuation. That's amazing. Wow, what a go. That's incredible. Okay. Actually. Yeah, so maybe they didn't do the full 900 million. Grok isn't. Sure, sure, sure, sure. But I mean, either way, you can be pro or con on space data. And Grok also says SpaceX has been Atreides largest venture capital position as of 2022, comprising an estimated 30 to 40% of its VC portfolio. Feels wild. So anyways, Gavin is set to. On one hand, Gavin could know things about space data centers that the rest of us don't know because of his position. And proximity to SpaceX. At the same time, SpaceX, if they go public next year, would benefit a lot from having a space data center kind of like, narrative as part of this. It's crazy because the business is so solid. I feel like you don't need fourth act or whatever. Okay, wait, so what about this? So if a lot of the gains of space and the center are just because the energy is free, why don't we just. But then there's this hard part about having, like, different satellites everywhere. Put the data center on the moon, there's no atmosphere, so you don't get the 30% reduction in energy from, from the sun and just put data center on the moon. Well, don't you get a 50%. Reduction because there's like, there's a dark side of the moon, like, like the moon. You wouldn't put on the dark side. No, I know, I know, but what I mean is that, is that our, our dark side sometimes gets light from when it's not facing us. Right. And so, so you. Only, if you're on the moon, you only get light 50% of the time. Whereas. Whereas if you're a satellite and you're flying out in space and you're kind of orbiting the Earth, like, but at a weird axis that you're not constantly in shadow all the time. You're getting sun, he says, 24 hours a day. I think it's. I think it's like 99.9% of the time. But regardless, it's like almost always. Whereas if you're, if you're actually like on the moon, you don't get sunlight 100% of the time. You get a 50% of the time. I'm pretty sure. I don't think it's because maybe there's something. Yeah, the sun's not. The sun's not like. Or the moon locked. It's Earth locked. Pull up this post I put at the Bottom of the stack from Joe Morrison. He's over at Umbris. While we pull that up let me tell you about Restream 1 livestream 30 plus destinations. If you want to multi stream go to restream.com and we will pull up this next post. What we got. Joe says you just put a radiator on the dark side of the satellite and he says thermal engineers in absolute shambles right now. Yeah. People are not. I don't know. People are so scared. Stacey Hammer says to be fair this is a big satellite. Joe says many big satellites. Big satellite makes it easier. I don't know. I don't know. Yeah, I think you need like, like a lot of like you need. You need basically a lot of mass in order to dissipate enough. Sure. Heat from a. From a GB300 or something. GB200. 200 somehow AST Space Mobile Inc. Is up 27% in the last five days. So. Excited. People are. Quite excited. Yeah. Well, should we head over to Gulfstream News? There's some massive news from Gulfstream Aerospace. The G400 introduces next gen Gulfstream tech to its class. Sound on to learn more with Gulfstream President Mark Burns. How has Mark Burns not been on the show yet? Here we go. Clearly don't have a no render policy over here. They're. They're render racing. I think they're just not at. You think they could just be adding special effects? What do you mean? This is the most rendered video I've ever seen. This looks like it was rendered okay. But here he is 2010. You think this is real though? Bring Gulfstream performance standards. This is on a green screen cabin class fulfilling our customers needs for a product line. I knew it was too good to be true. Like this is real. This is. It looks like it's far away. It does look like a nice plane though. With Gulfstream's signature combination of range, speed and cabin comfort, the G400 will provide unrivaled efficiency thanks to the combination of the advanced practices. This is going to be a hot Christmas gift this year. Absolutely. People have been wondering. I think this will be top of. This will be under the tree for a lot of people. It's not ideal to do like a pre order as a gift. You know it's nice to have something under the tree but I think people would make an exception here. So funny. I was talking with a mutual friend of ours, a guest who's been on the show many times and they're planning to purchase a jet at some point in the Relatively near future. And he was so fixated on like needing to be able to. It's his first, it will be his first jet. He was so fixated on like I need to be able to stand up at like it can't have, like the cabin needs to be big enough that I can like walk around freely. I was like, come on, for your first jet, can you not just like, you know, tolerate like slightly lower height in the cabin? He was like, no, I got to be able to. I got to be able to move freely. Go straight for the 747. The BBJ. Yeah. Get the VVIP package. The one that only the Qataris have for some reason. Well, Casey Hanmar. Okay, I got it back. I got a text back from Deliance. I won't read all of it. Okay, there's a line here. This is from someone else. Dude, all this data centers in space stuff also I'm not going to read that you need 1 million 100 kilowatt sats to generate 100 gigawatts in orbit. 100 kilowatt sats Wang 1000 kilogram including payload would mean a 5x increase in power density. And you can't just not have batteries and slap a radiator on the back. Lol. And then there were some choice words exchanged that I won't read on the show. But anyways, I think we should host the data center in space debate. That sounds good. We should get a few people on here. Somebody to be bullish, somebody to be bearish. I think we want somebody to be bullish that has put things into orbit and taken them down many times. I think it's hard to. It's hard to lean on somebody's kind of take or opinion if they're not actually participating. In space. They're just sort of speculating. Yeah. Let's read through a little bit of what Casey Hanmer had to say. He said here's one idea about SpaceX's next big thing. AI computing inference. Again, not training on orbit. But how the hell can SpaceX do this cheaper than just building more data centers on the ground from first principles? It's an attractive proposition because the GPUs have extremely high value per kilogram and extreme revenue per kilowatt, both of which are relatively expensive. That is the value Prop somewhat washes out the pain of operating in space. So I took a closer look. If anyone can make this work, it is a Starlink derived system. So I started with the Starlink V3 satellites with some high fidelity CEO CAD below orbital parameters Pick a Sun synchronous orbit so we're in full 1400 kilowatt per meter squared sunlight at all times. No need for batteries. Deploy the solar array in sun slicer mode facing full sun, but the edge is pointing in the orbital direction bottom right in these images to minimize drag. But the Inference Starlink star thought satellites don't have to scrape the atmosphere being in sun synchronous orbit. They'll need to use the rest of the Starlink constellation for backhaul via laser links anyway, and higher orbits actually improve worst case latency very slightly too high though, and SSO is relatively full of debris. Let's pick 560km. A Starlink satellite in this orbit has full sun, so the back half is always shaded and relatively cool. The next hottest thing in the sky is the Earth taking up almost half the sky to the bottom left in these images. And so what does he conclude? He says, I've seen a bunch of high inclination Starlink launches from the Vandenberg recently, but I don't think any of them were going to sun synchronous orbit. In any case, a ring of Inference satellites visible at dawn and dusk running north and south will be awesome. If one Starship can launch 100 tons to Leo, then that gets close to 30 megawatts of inference per launch. 1,000 watts is 300 gigawatts. Now we're talking real scale. We're doing one launch of Falcon 9 per day. There was delions loosely. We're now one a day. So 300 in a year. So you get to 1,000 over three years if you're on a one a day cadence or if you can wind up getting more the math is not that crazy, but it does seem like Delian was pushing more towards like hey, this is maybe 10 or 20 years, not the next three years. But it certainly will be fun to follow the story. In the meantime, let me tell you about Cognition, the team behind the AI software engineer Devin Crush your backlog with your personal AI engineering team. Jordy has survived. I keep like getting the podcast in a can like down the wrong they. Should really make a make a drink specifically for podcasters that has no carbonation that cannot be choked on in any way. Mark Benioff says LLMs are the new disk drives commodity infrastructure you hot swap for whoever's cheapest plus best the fantasy that the model is the moat just expired. Marc Benioff having fun on the timeline. I love that he's having fun. I love that he's taking shots don't. They have an AI lab? To be clear, they were working on Einstein for a while, but I think that they are very much happy to be a rapper, happy to be a buyer of LLMs at this point. It certainly has penciled out that way. We can talk more to Mark about that. It's feeling Good. Stocks up 10% in the last five days. Here for. Best Salesman history. What. What do you think about Michael Burry? The whole industry needs a $500 billion IPO ASAP. Said any. Everyone that knows anything knows this OpenAI is the next Netscape doomed and hemorrhaging cash. Microsoft is still trying to keep it afloat while keeping it off balance sheet and sucking out the ip. So why do they keep getting funded? The whole industry needs a $500 billion IPO ASAP. The whole industry needs to go away and sit in a corner for a couple of centuries and think about what it's done, says Jeffrey Miller. Wow. I don't know. This is. People want extreme takes. This post would go super hard if you don't understand the Microsoft OpenAI relationship. Yeah. Yeah, it's just. It's just. What was Netscape revenue? What was Netscape, I remember correctly, they. IPO'd with like 15 million of revenue. That feels a little bit different. In Netscape's peak revenue, 1997 was 500 million. I mean. We'Re up at 20 billion this year for OpenAI. Yeah. Their total revenue before IPO for first two operating quarters, Netscape reported total revenue of 16.6 million. Just odd. I mean, did Netscape have like an enterprise business? Should ask Mark. Yeah, I don't know. I mean, even if there's a. Yeah. Yeah, like, even if there's more commoditization, like it's going to take time and I don't know, the dynamics of the competition here feels like there will still be a lot of value, even if. Even if it is somewhat commodity infrastructure. It's like, you know, what else is commodity infrastructure? Like aws, gcp, Azure. Like you get like a server with some hard disk on it. Like that is commodity. And yet they all have 30% margins and they're all massive businesses. And when AWS, when Amazon broke out AWS as its own line, it was like an IPO of its own company because it was so big, it was such a massive business and it should be completely commodity because it's just servers in a data center. And yet. Yeah, it's interesting. Yeah, it's interesting too. I mean, Burry has positioned himself of just Hating any company that's overheated. But ChatGPT having close to a billion weekly actives and ultimately even if they just compete in search. Right. At least it's a multi trillion dollar opportunity. Whether or not they fully execute against that is another thing. Are you buying the IPO timeline of Cohere, then Anthropic, then OpenAI or do you think they will be sorted in a different pattern? What was the order? Cohere first. That's Aiden Gomez's foundation model lab. He is of course on the transformer paper also a dripped out technology brother. That's right. With a fantastic set of outfits, then, then anthropic and then OpenAI. OpenAI third. That's, that's the rumor to say. I have no idea how the market will react to Cohere. Yeah, Cohere has shifted more into like the business, the enterprise market. Fully enterprise. Not in the timeline at all. That could very well be intentional. But it would be. It'Ll be interesting how much excitement they can build around the ipo assuming they're losing a lot of money, assuming the enterprise is really competitive and they're competing with Anthropic and OpenAI and open source models. If the narrative changes to the point where, you know, this idea of every enterprise is going to need basically a custom LLM that's fine tuned or pre trained on their data, like what AWS CEO Matt Garman was talking to us about, about these pre training checkpoints where you can go in and say okay, I have a business and I absolutely do not want upload my data to Anthropic, I don't want to update it, I don't want to upload it to OpenAI, but I want a model that understands my business's data at a core level in the pre training step. And Cohere can offer that and show traction there. That feels like that could be the next wave. We've talked to a number of founders that are building those custom pipelines and there's not really a public company that is even anywhere in that narrative. And that narrative does feel very nascent. It feels like it has not percolated through the public markets yet. So I don't know, it could be, it could be, you know, exciting for them. Well, let me tell you, they've only raised 1.4 billion. I did not realize they raised, sorry, 1.5, 1.5 billion. 1.54 billion. Very modest amount. Especially when you have seed stage companies raising half a billion. We got one of those on the show today. Joining later, he kicked the hornet's nest on the timeline. What do you do? A little bit. People were, people were pushing back because. Of his, his, his racing history. No, no, no. Oh people, people. Show respect. Love that. Show massive respect. No, but I think, I think Jack at SLOW was saying that Nvidia was valued at at 500 million when they IPO'd. Of course that was very long time ago. But anyways, people, people are just saying like it feels pretty potentially overheated. But we have Naveen coming on the show today and I'm excited to hear about the opportunity from him. I'm ready to say it's overvalued unless he's built a company previously and sold it twice. Maybe. Yeah, twice. If he's done that then he's, then he's off the hook. But he's got to prove it to us because if he's just some new grad, some dropout like a Waterloo. Yeah, if he's just, if he's just like. If he shows up on the show and he's just like yeah, I had this idea in my garage, I didn't go to school, I've never done in business before and now then I'm going to, I'm going to call him out and I'm going to say you're overvalued, you raised too much money. But if he can prove to us that he's done it once or twice before, then I'll let him off the hook. Yeah, I'll let him off the hook anyway. I think that's fair. Let me tell you about Privy. Privy makes it easy to build on crypto rail, securely spin up white label wallets, sign transactions and integrate on chain infrastructure all through one simple API. Let's read the semi analysis post says important A common misconception about OpenAI's upcoming custom chip is that since it's a custom chip, it won't be flexible and will be a dataflow machine. OpenAI has recognized that the 100x efficiency gains for training and inference happen at the algorithm layer and the hardware chip needs to be flexible enough to accommodate these algorithm changes. We went from just pre training transformers to now doing RL post training on transformers. We went from dense transformers to mixture of expert transformers and soon ultra sparse transformers that will have four active experts per token out of 2,048 total experts. We went from casual MHA attention to MQA and GQA to attention, sync sliding attention to now even learn sparse attention. Despite AI tourists think the chip OpenAI is building with Broadcom will be far more flexible than TPUs. Interesting. Despite most of OpenAI's chip team being poached from Google's TPU team. The semiconductor horse race continues to fascinate me. And I'm so glad that we have the good folks over at Semianalysis coming on the show regularly to help us understand it. I always enjoy talking to them. And we have. Two folks from Semianalysis joining the show next week, so we will be having a great time. Let me tell you about TurboPuffer, serverless vector and full text search built from first principles and object storage. Fast 10x cheaper and extremely scalable. Of course. That's right, it's scalable. A hedge fund was ordered to pay a bonus to a trader who made 97% of its revenues. This is hilarious because when I read this at first I thought it was he had made like his target bonus was let's say 10 million and he brought in 9.7 million. And so he was 97% of the way to his bonus. And they were like, you didn't, you didn't hit your bonus, buddy. You don't get the bonus at all. And I was like, oh, that sucks. But like that's kind of the deal. That kind of makes sense. But Evolution Capital Management has to pay him because. Yeah, so a hedge fund that was sued by a trader for refusing to pay a performance related bonus despite him making 97% of its revenue, has been ordered to pay him 5.4 million plus interest by the High Court in London. When I read the headline of the story, I expected it to not be in the mid seven figures. But would you expect, I mean, I was hoping at least eight. At least eight. He's not getting a G400 off of this. Not even close. Not even close. Robert Gagliardi sued his former employer, Evolution Capital Management in London, alleging that it acted in bad faith by denying him a $7.5 million discretionary bonus after he had generated more than 60 million for the firm. Wow. Wait, does this mean that every other trader lost money was just something like that? I don't know. Heavily. So he was at the fund brief stint, he made 12 million, he was already paid 7 million, including a $625,000 signing bonus. Base salary was 425k and the $6 million new issue bonus. So he needs to get out of this. He needs to go into AI research ASAP if he wants to be putting up real numbers. This is rough, but there are some harsh words here. He said Gagliardi, a block trading specialist alleged that he was told in early 2021 that a return of $10 million over the rest of the year would be an excellent result. When Gagliardi asked the fund's founder, Michael Lurch, for the payout in 2022, he responded, I'm not going to pay you the bonus if you sue me. So Gagliardi did, and he won. He did. He did. Although Evolution did not dispute his extraordinary returns, the fund argued that the damage done to its reputation as a result of dealing with a US SEC probe into some of Gaglieri's trades at his previous employer outweighed his performance and that the bonus did not need to be paid because it was discretionary. There's no question that Mr. Gagliardi made exceptional profits for the funds, and Mr. Lurch frequently praised his performance in that respect. Had Evolution properly performed the contract, Mr. Gagliardi would have received a discretionary bonus of 5.3 million. The judge also said that Gagliardi was Lurch's prize asset at Evolution. That's a good goal for everyone. That is thinking, you know, maybe working at a company become the prize asset. Hopefully that doesn't result in a lawsuit like this, but they're feeling. They're feeling. Gagliardi and his lawyer are feeling very vindicated. Well, we have our next guest here in the Restream waiting room. Let me tell you about Gemini 3 Pro. First, Google's most intelligent model yet. State of the art reasoning, Next level vibe coding and deep multimodal understanding. We have Aaron Ginn, the CEO and co founder of Hydra Host. Good to see you. It's been too long. How are you doing? Welcome back. Been all right? Yeah, yeah. I just got.

NEWSMAX