Nvidia’s $500B Compute Deal, Paramount Threatens CA Exit, Musk’s “Shortcut” to $1T Payday | Diet TBPN

0:01John Coogan:Is that the NVIDIA Compute Deal Alarm? Huge deal on the cover of the Financial Times today. Wall Street, big names, join NVIDIA to build a$500 billion AI financing package, a landmark lending plan. Jensen really lined up the murderer's row of financiers for his new AI financing package. Legendary setup over on CNBC, too. One of the greatest roundtables of all time. Yeah, incredible. Only Larry Fink was remote. They got David Solomon, CEO of Goldman Sachs, in studio, John Gray from Blackstone, Jim Zelter from Apollo, Bruce Flatt from Brookfield, and of course, Jensen breaking it all down for everyone.

0:42John Coogan:I'm sure the Hollywood starlets were pounding on the glass outside looking for some new arm candy, don't you think? 100%. 100%. Because it's very rare to see that many individuals. Capital allocators. specifically capital allocators, put it all on the line every day in the markets. Some would call them bad boys. Some might. Some might. Some have. So I'm sure they were lined up. But it's a great segment. I wanted to play this one clip of them discussing just the profitability, why they see this as an investable asset. There were a couple of quotes. The sheer size of the AI infrastructure build out is unprecedented.

1:18John Coogan:The president of Apollo said more than$8 trillion of capital is expected to be invested, a staggering sum. We see an enormous opportunity for private capital to finance a portion of this along with public capital. So Wall Street's not taking their foot off the gas. Anyway, let's play this clip from CNBC. More of your margin of safety advantage. So are these concerns about whether we can meet this demand overdone at this point? Do you think, Jensen, that from where you see things, the demand level and how we're building up around it, that it's going to be okay? It'll all work out. We're going to be constrained for some time and pretty much across the board from chips to memories to packaging to systems, photonics, connectors, land, power, construction workers, the whole thing.

2:05John Coogan:The entire supply chain up and down behind me upstream all the way downstream. And this is happening at a time when AI has become useful because it's starting to do productive work and it's happening all over the world. And AI tokens are profitable, incredibly profitable. When you have something profitable, everybody wants to make more of it. Great demand, great profitability. The conditions are exactly right for the work that we're doing right now. Jensen, why these companies? And did you go to any partners who said no? No one said no, but this is the sixth premier, world's premier institutional financiers for infrastructure.

2:49John Coogan:This is the best of the best. What John said, that right now you'll be less likely to have public capital that comes into this because a lot of these are companies that aren't making money yet. Is he right on that, or are there going to be big banks and others that kind of step up? I believe within months you're going to realize that these companies are extremely profitable. These are the fastest growing technology companies in history. Your customers, you mean? That's right. These are fastest growing technology companies in history, and the tokens they're generating are incredibly profitable.

3:21John Coogan:You know, if the wafers that we buy from TSMC are incredibly profitable, there's incredible demand for it, I'm going to want to buy a lot more. By the way, who are we talking about? Your customers? Which customers will have access to these? AI labs. AI labs. ALAN SEALES - And AI Labs are the ones. That's the ones that you think are profitable. But will this financing go? ALAN SEALES - AI Labs, AI startups, as you know, this last six months, the world put in about$500 billion in AI startups. $500 billion is the largest investing period, probably in recent history. And these companies need compute.

3:55John Coogan:And so we now have the vehicle to do so. ALAN SEALES - That's great stuff. $500 billion, huge number. but it feels a lot smaller when he lays out the actual compute calculation 50 billion per gigawatt 50 to 60 50 to 60 so you're looking at 10 billion or 10 gigawatts of powered compute which is you know the labs are at like three combined across a few of them and we've been on this 3x scaling so this is really just like next year's compute basically i think meta individually has a 10 gigawatt plan or something like that He's having to talk about profitability and the profitability of incremental tokens.

4:35And I think one of the reasons that a lot of people are just very uncomfortable with this is that the two leading companies in the space are private. And their various numbers leak out from time to time, but you definitely don't get a complete picture. And it's very unusual to have the two companies driving the public markets are actually private.

4:55John Coogan:Yes. Well, there is the SpaceX factor here, too, which recently announced that they're going all in on NVIDIA. But again, they are turning into a neocloud in many ways and licensing out compute. So there's this world where you might see Wall Street banks. Yeah, but I'm looking at SpaceX more on the supply side, right? No, no, no. Not driving. No, exactly. So it's like a six-layer cake, and you're seeing the finance guys there. than NVIDIA making the chips, marshalling all this capital. A lot of it's going to go to a Colossus 5 data center and then be rented by a lab. That will be one of the potential outcomes.

5:32John Coogan:But then they're just in the supply chain. Because I don't think the semi-analysis forecast was for Grok token demand specifically. It was just overall inference and compute demand because they're capable of building data centers very quickly. Anyway. Jensen also took to X to post a long essay, 12 ,000 likes. He's like fully on X now. I haven't seen him. It's almost like he was lurking, studying. You think he was lurking the whole time? Well, I don't know if he was studying enough because people are getting mad at him for using the forbidden phrase. He says, NVIDIA compute is not just a chip. It is a complete AI factory platform, including accelerated computing, networking, system software, AI frameworks, and a global developer ecosystem.

6:19John Coogan:That doesn't sound that AI to me. I don't know. Somebody will run this through Pangram, and we'll get to the bottom of it. Jensen is completing the circle, says Prakash. Bankers don't like GPUs as collateral because the depreciation is unpredictable. It's unpredictable because a new GPU can obsolete an old one. But Jensen knows his own roadmap. so he's offering depreciation insurance to the banks. The depreciation insurance up to 25 % helps the banks get marginal deals over time. Speculation here from Prakash. NVIDIA will also advise the banks on reference designs for data centers that will make them fungible, so you will know this is a one gigawatt data center, but it's in this particular class, this particular configuration.

6:58John Coogan:It's Blackwell data center. It's powered this way, and so if you can put it in this bucket, you can underwrite it a particular way because it becomes more fungible. Having them be fungible means the debt can be repackaged into asset-backed securities, collateralized loan obligations, and collateralized debt obligations, ABS-CLOs and CDOs from 2008. I'm sure there will be a lot of folks upset about all of the comparisons to the mortgage-backed security build-out in 2008. This allows tranching to get investment-grade ratings on the debt so it can to be sold and resold to pension funds and insurance firms.

7:34John Coogan:It also allows the banks to trade idiosyncratic project-specific credit risk for sector-wide credit risk. So Jensen is trying to get his customers the same cost of financing as real estate rather than venture equity. This is going to move the data center game out of the VCs and into the big leagues. And so I'm sure people will be speculating all sorts of different things on what happens next. There are some other stories. Paramount is threatening to leave California by October 1st if the state refuses to negotiate a settlement in the legal fight over its Warner Brothers Discovery merger. We've been covering the story back and forth for a while, but according to Variety, Paramount CEO David Ellison wants a quick resolution to the antitrust lawsuit brought by 12 state attorneys general seeking to block the deal.

8:20John Coogan:He has now told senior executives that Paramount is prepared to begin the process of moving its operations out of California if Attorney General Rob Bonta doesn't enter settlement talks. Ellison reportedly told his leadership team last week that the Paramount Skydance board has already approved the relocation plans. If negotiations with Bonta haven't begun by August 1st, the company would start preparing its exit with Paramount's Los Angeles headquarters potentially moving out of the state as early as October. The threat dramatically raises the stakes in Paramount's fight with Bonta, who has emerged as a leading opponent of the Warner Brothers Discovery acquisition.

8:58John Coogan:Bonta hasn't publicly detailed what Paramount would need to offer to resolve the case, but he has said that any acceptable remedies would likely need to be structural, such as asset divestitures rather than behavioral commitments, like maintaining certain levels of production. So it's not going to be enough for Ellis to say, hey, we're still going to do 12 movies a year, 24 movies a year. It needs to be something specific about the actual structure of the company. Paramount is also racing against an expensive clock. Beginning October 1st, the company will owe Warner Brothers Discovery shareholders a$7 million per day ticking fee until the transaction closes.

9:37John Coogan:The state's antitrust trial isn't scheduled to begin until May 2nd of 2027, so almost nine months from now, meaning Paramount could rack up roughly$1.2 billion in payments by the time the case is expected to conclude. Ellison is effectively putting pressure on California from the other direction. If the state won't help find a path to closing the merger, Paramount is prepared to start taking jobs elsewhere. Yeah, so he's trying to force the issue. This gets extremely painful for the Ellisons if this antitrust thing just drags on. You could imagine it dragging on for a couple years. That puts a pause on all of their integration plans.

10:22It just makes everything a lot harder. I don't think he has this. This seems like his one option, right? It's kind of the nuclear option. It's going to piss off a lot of the industry here in L.A.

10:35John Coogan:It is the most symbolic move because, according to Variety, the company's L.A. headquarters would be the first operation to leave. and Ellison already has a five-year plan to move most studio jobs out of California. No destination has been selected, but Georgia, Texas, and Tennessee reportedly under consideration. So it's like if you're moving the L.A. headquarters, the iconic Paramount headquarters, everyone's seen the water tower. That is a huge shot across the bow as opposed to something that might actually be more economically impactful. Like just, oh, for this production, we were going to make it in L.A.

11:10John Coogan:We're going to do it in Atlanta. that might actually move more dollars around, but maintaining that headquarters is so symbolic, right? So Ellison and much of Paramount's senior leadership currently work from the historic Paramount Pictures studio lot in Hollywood, but maybe it will be moving to Georgia, Texas, or Tennessee. I wonder where they will go. Anyway, speaking of publicly traded companies, Elon Musk may have found a shortcut to unlocking a trillion-dollar payday. He needed a win. You asked for it. He's delivering. Yeah. The idea is having SpaceX buy Tesla. This has been rumored for a long time.

11:48John Coogan:Lots of speculation on when these two companies will merge, if they will merge. But the Wall Street Journal on the front page outlines a very odd scenario where Elon could wind up making an incremental$1 trillion. It's really complicated. And it's not as simple as just, oh, he'll just buy the companies and then he just gets another trillion dollars. It's more complicated, but there is a potential outcome, but there's some mitigating factors. But it's worth understanding the mechanics of the deal or the mechanics of potential deal as it might play out over the next few years. Because this is something that couldn't happen right now, but in the future it is possible.

12:28John Coogan:So the Wall Street Journal reports that there's an obscure provision in Musk's 2025 Tesla pay agreement that's already been approved by the shareholders. And it was that crazy one we talked about where it was like a million robo taxis by this time. And like the Optimus needs to be shipping. And it was a lot of really bold claims. But if he hits all those and the stock goes to like$10 trillion, then he gets a huge unlock of new stock. And it was all like, OK, well, like it's a lot of money, but that's a lot of progress for this company. Because there's a lot of projects at Tesla that just are sort of slow and steady, not really advancing that quickly.

13:08John Coogan:And so he was sort of throwing down the gauntlet saying, okay, give me another couple years, and I'm going to deliver in a really big way. And if I do, I want to be compensated for that. And the shareholders approved. But the obscure provision, there's an obscure provision in the already approved Tesla pay agreement that could eliminate half of those performance requirements attached to the stock award if Tesla is acquired. So change of control affects those pay packages. What are you laughing about? that mark in the chat says talk about zuckerberg again i like feisty geordie feisty geordie is based af yeah i guess i woke up on the wrong side of the bed yesterday i had some strong opinions we'll see somebody had somebody had to say it hey hey he fired back he called out he alex heath he said something big is coming he was vague posting directly to alex heath and again i just felt like I think that was so memetic with all the people that are actually at the frontier.

14:05And it's just the whole thing. I'm like, yeah, I'll believe it when I see it.

14:11John Coogan:Yeah, yeah. It's kind of fun. What was the actual? See, see, Mark, now you got me started. Now you snapped him. Now I snapped. Over the weekend, Zuck called me via his metaglasses while fishing to discuss the 6 ,000-word pro-AI manifesto he published this morning. Yeah, so he's trying to balance. I love that aesthetic. That's fantastic. Good old American. Fishing. American. American Mark. He told me the immediate reason he decided to publish his philosophy now is that he believes Meta is very close to having substantially stronger models. Probably true. And he wants people to understand his values before those models arrive.

14:55We know your values. we know your values which uh are connecting the world it's fine instagram's fine no i mean the whatever whatever the stated values versus the the lived values the lived values are

15:13John Coogan:if there is a hot product help small businesses grow what about that value come on uh instagram has birthed so many companies we know so many people that their companies would not exist if not for meta platforms. How about great hours of enjoyable content? Yes, yes. For everyday Americans to watch every day. Would we be even remotely familiar with Professor Sandy and the creation of the Wamba without Mark Zuckerberg? You would have never seen that he turned it down. You would have never seen that he turned it down. Or the Dabitah. Yeah, you would not be aware of that. All these iconic moments.

15:49John Coogan:You also wouldn't know that once you go to Ibiza, you must go back to Ibiza, And you would not know that there's a 21-year-old bodybuilder who looks like he's 35 who's on his way. See? It's enjoyable. Admit it. I don't have a problem with Mark. Yeah. And I don't have a problem with Meta products. I do get enjoyment out of them. In the business. The ads are good. I have built companies that are dependent on Meta platforms. I've invested in many companies that exist in their current form entirely because of Meta platforms. There you go. But I know what Mark's values are. What are those? Mark cares about...

16:31John Coogan:Delighting customers. Profitable advertising. These are good values. I wish he cared more about advertising. That's true. Yeah. I wish he cared more about advertising. That would be good. But, you know, you just look at the historical behavior. It's like buy or copy or chase the hot thing. Yeah, but that's just a game. Mike, you're talking about a$10 billion startup that's coming after you with some new form factor. I don't know. You copy it. I think you're discounting how good MSL is. Everyone is saying that MSL is clearly in third place. That means that they're ahead of XAI and Dmine. I mean, if they flip Dmine, it's crazy.

17:08John Coogan:If they actually get the next version of Spark out before Gemini 4 and it's better, and then Gemini 4 launches and stays in fourth place, that is a crazy, crazy reversal. They're executing well, coming from incredibly far behind. They are approaching the frontier. I just think that if he wants to be the pick me lab, you don't like the comms. I don't. Yeah, I just think it's I think it's not. I just don't think it's very authentic. OK, I think it's the opposite. I think I think it is authentic. I think it is. He is thinking about this stuff. He might not be the, it is just a little bit of a tough voice because there's so much attention from the social reckoning and the social network and stuff.

17:58John Coogan:And then like in terms of the overall AI industry, attaching him, it's like attaching Demis to the AI industry was definitely better for the voices of the AI industry. Because Demis would stay on message and just talk purely about science. He won the Nobel Prize. And so if I was like, okay, I'm dealing with an AI skeptic who is paranoid about water use and surveillance. Like, who can I put them in the room with? I'm like, yeah, go sit down with Demis, Sir Demis. He's going to put on a good show and, like, walk you through this and give you a really optimistic vision. And he's not going to have the baggage of anything else, any lawsuits or anything else, right?

18:40John Coogan:And so, yeah, it's a little bit of a tough go. Anyway, back to Elon Musk. Anyways, I just think, like, again. Let's get back to less controversial tech leaders. Let's talk about Elon Musk and how he's going to make his next trillion. Because this is what's in the Wall Street Journal today. I wish. I'll just end it here. Please. I wish that Zuck came out with, like, a five-point plan to get his own trillion-dollar pay packet. Which was just, like, sell, like, trillions of dollars worth of ads. Ooh. And I would just be praising that all day. That would be sick. I would be praising that all day versus being like, oh, we're doing open source.

19:20And now we're not doing open source. And now we're open sourcing again because I'll get some brownie points. But also we're going to still have closed models.

19:32John Coogan:Yeah, I mean, you go to the Dara thing and it's like the guy's been hard on China. He's been anti-open source since day one. Extremely consistent. You love him. You hate him. But, like, there's consistency there. And that is just reliable and in many ways admirable. I understand what you're saying. Anyway, back to Elon Musk. He's been extremely consistent. He's getting that trillion dollars in Tesla one way or another. And there are multiple ways. He's tasted being a trillionaire before. Yeah. He wants to run it back. He's got to run it up. He's got to double it. They say your first trillion is the hardest.

20:01John Coogan:And this is how he gets his second trillion. So this is how he's going to do it. So in Musk's 2025 Tesla pay agreement, if he hit certain milestones, he can eliminate a lot of those performance requirements that were previously attached to a stock award if the company's acquired. And so change of control at this scale, normally you would think that's impossible. You cannot take Tesla private. He tried. It was too big of a company. And then also once you're a trillion dollar. Funding was secured. Funding was secured. But there was a trillion dollar, once you're a trillion dollar company, like you can't just get acquired, except we're in this very unique case where SpaceX is also a trillion dollar company.

20:42John Coogan:So Tesla shareholders approved the compensation plan in November, last November. Under its normal terms, Musk can earn as many as 423 million Tesla shares across 12 tranches. But each tranche requires Tesla to hit both a market cap and an operational milestone. So the goals are deliberately enormous. Tesla would eventually need to reach an$8.5 trillion market cap while accomplishing targets, including delivering 20 million vehicles. Remember, I think the number of vehicle deliveries is actually declining this year. It was a very, very bold plan. They need 10 million active FSD subscriptions. That actually seems easy.

21:23John Coogan:FSD is really good. They need to deliver 1 million Optimus robots. That seems crazy because Optimus is still so early as a project. We should try to interview someone with a Tesla that doesn't get the FSD. They turn it down. They turn it down. And I want to understand why. I understand that if you just bought an older Tesla, you like the ease of charging and you haven't upgraded to the newer hardware package. Because the FSD is available on the older hardware three technology, but I think it's best on the hardware four, which is like 2024 onward. Not everyone can upgrade. If it's a financial decision.

21:59John Coogan:I understand it. But the last one was they need to get a million robo-taxis into commercial operation. That also seems pretty doable. I saw a robo-taxi driving around LA recently in gold. And just from using FSD, it seems like it's ready. There's probably some legal stuff. But in general, I think they could roll out the robo-taxis very quickly. They can make a million cars pretty quickly. And they have the technology. So it's just about putting those Yeah, I haven't spent a ton of time in Teslas, but the times recently where human drivers had to take over was getting into a driveway. Oh, yeah.

22:37It's effectively like private property. Yeah. It's a robot taxi you can just pull over to the side of the street and say, like, you're walking the last 100 feet, buddy.

22:44John Coogan:Yep, totally. And so, yeah, I mean, that one doesn't seem that difficult, although obviously it is a lot. I mean, I think that's like 100 times as many Waymos. I think there's like 10 ,000 Waymos out there. So it would be a big move. But over a couple of years, is that possible? Yeah, so vehicle deliveries were falling in 2024 and 2025, but seemed to be rebounding. But they're not far off from like a million vehicles, right? 838 in each one this year. 838 ,000. Yeah. And so over a year or two, they could probably manufacture a million robo-taxis. And I think the technology is pretty much there.

23:24John Coogan:When you actually look online and you see the reviews of people talking about Tesla Ubers, they're like, I wish there was a Tesla product or an Uber product where I could demand that if it's a Tesla, they have to stay in full self-driving mode. because many Uber riders regard the full self-driving experience as smoother and less likely to cause indigestion and sickness in the back seat versus watching a driver who has a Tesla who isn't that experienced and doesn't understand how to use the regenerative braking properly. And so it's much more jerky when a human's driving it because FSD is actually superhuman relative to a newbie Uber driver with a Tesla.