Bezos' $100B AI Plan, Nvida Chip Smuggling, The Mansion Section | Diet TBPN

0:00John Coogan:First, I wanted to chat more about the breaking news from yesterday. Jeff Bezos is in talks to raise$100 billion for an AI manufacturing fund. There was a funny post that was like, he has$200 billion. Why does he need to raise more money? No conviction? No conviction? Oh, yeah. It was funny. It was like it definitely this this broke containment and got into kind of the anti-capitalist people are not part of acts and they were like wait why are you asking people for money you have all the money yeah he wants to let other people in on the action well i don't want to let it i i do wonder what his like gp commit will be that's what i was saying yesterday i was like i would expect him to be like yeah i'm good for 20.

0:4020 or 30 or something like that i don't know maybe he'll go

0:43John Coogan:even longer but in general i think the i think the idea of going around the globe because he's an international person at this point uh hoovering up money from all over uh different sovereign wealth funds and then deploying that to help rebuild the American manufacturing base is a good thing. I wrote about it today in the newsletter, tbpn.com, and I also tried to do some searching and think about what do you actually buy for$100 billion? What does a portfolio look like if it's a private equity style roll-up? If you're just taking out the market caps and you're sort of assuming the debt and continue to operate the business but then bring more efficiency to bear, what are some of the top options?

1:22John Coogan:What do some of the options look like? The other thing is I saw a very viral post about someone that seemingly read the headline and thought, don't you have enough money already? Why do you have to buy all these companies and fire? Like they saw automation and they just assumed like job loss. And when I look at this, I'm like, okay, if he's successful, we will actually start adding a bunch of new manufacturing jobs, which we've been shedding over the last year. Yeah, exactly. Exactly. So headline number's funny. If he had done$99 billion or$101 billion, I don't think people would be making as many SoftBank comparisons.

1:57John Coogan:But of course, as soon as you hear$100 billion for a mega fund, you instantly think SoftBank. You think this guy's got vision. He's got vision. It's a vision fund for America. I like the idea of an American SoftBank, an American vision fund. So I'm generally pro. But of course, people are going to make comparisons to SoftBank, which has had some huge winners and a few black eyes. Moss is a high vol guy. He's been the world's richest man, briefly, during the dot-com boom. It was reported that he was worth more than Bill Gates. It was pretty temporary, unclear exactly how much. Of course, there was no liquidity at the time.

2:30John Coogan:But he's also held the crown for the man who lost the most money in human history after losing$76 billion in two years. You're supposed to do the wah-wah for that. But he made it all back. I was preemptively hitting it. Okay, yeah, yeah. He did make it all back. He's made almost$100 billion twice. I think maybe there was a moment when he had the title of making$100 billion twice. One on Alibaba and then the second one on Arm. He's made money on Nvidia and lots of other things. The idea of an American SoftBank, especially one focused on revitalizing American manufacturing. I like that idea. There are a million reasons why this is an important project.

3:06John Coogan:Job creation, economic independence, national security, and you might just get more cool stuff. Like as manufacturing gets cheaper, like the tools that you use, the cars that you drive, the washing machine in your home, the tires on your cars, all these different things get better, not just cheaper, but actually unlock new capabilities and new things that people enjoy. So I'm in favor of more manufacturing, more economic development, basically. And so the question about Jeff Bezos' role, he's a great operator, and I think he's the right person to take the reins and actually deliver here. So like Masa, Jeff Bezos has been through the ups and downs of the dot-com boom and bust.

3:45John Coogan:At one point, Jeff Bezos lost 85 % of his net worth in like two years. But he made it back. He did make it back and more. He was worth like$8 or$9 billion and got down to his last$1 billion, basically, because of the stock drop in Amazon. But he built it. But he not only kept Amazon alive, which I think everyone knows that, you know, Amazon went up during the dot-com bubble and then crashed and then built back up. But he also kept Blue Origin alive during that time because he founded Blue Origin in 2000. Wow. Before the crash. I didn't know that. Yeah, before the crash. So he was... Was it before SpaceX?

4:20John Coogan:Yeah, before SpaceX. Whoa. Yeah. So... Because I... Elon sold PayPal. For some reason, I assume that Blue Origin was just mimetic with Elon. No, no, no. He did it earlier. Wow. And so he kept Blue Origin alive. And can you imagine how stressful it is? you're like okay i'm worth 10 billion certainly i can have a little side project as a treat and you're like okay i just lost 85 of my money i deserve what was i deserve i deserve a side project that loses yes 20 million a month exactly yeah i don't know how much they were losing but like but that doesn't seem unreasonable but he was able to keep it going uh of course blue origin was a slow story all the way up until like last year when they landed new glenn um and uh and so uh he He kept both alive and he's never given up on the project Blue Origin that always has felt behind SpaceX for the two decades that it's been operating, but is now sort of unlocking new capabilities and feels more real.

5:16They were going turtle mode.

5:17John Coogan:Yeah, turtle mode for sure. But it's working. Regardless of the fact that SpaceX is still ahead of Blue Origin, you have to give him credit for operating that company efficiently and actually delivering a product that we all saw go up and come back. And so like the rockets did go up and land and come back and he's delivered people to space and back And so he's he's he's like achieved the goal and not lost all the money and not put the company in jeopardy And so like clearly there's some operational expert. That's a narrative violation. It is a narrative violation What makes Bezos uniquely equipped to take on this type of project is the nature of his business over the last few decades?

5:55John Coogan:So he's never really had the zero marginal cost luxuries afforded to other pure play internet founders Amazon's just always interfaced with the physical world So it's been even though it's a software company and a tech company. It's always interfaced with the physical world And they've had to focus on operational efficiencies to scale. So in 2012 Amazon bought Kiva Systems which turned into Amazon robotics and That's a big risk to buying a big company And we think that if this$100 billion fund turns into sort of a private equity rollup, you're going to ask the question, can Bezos buy a company for$1 billion,$5 billion,$10 billion,$50 billion, who knows?

6:37John Coogan:And operate it and scale it and actually continue to deliver value. And he certainly did that with Kiva Robotics, which turned into Amazon Robotics, which delivered over a million robots that are deployed across the operations network. So from fulfillment centers, warehouse automation, inventory flow, last mile delivery, All of these require tight integration between hardware and software to run smoothly. Bezos clearly loves this stuff. And you can actually see it when you look at his face. Whenever he gives a tour of an Amazon warehouse or a Blue Origin facility, he's just beaming. And he's absolutely obsessed with this technology and the physical world.

7:10John Coogan:And so the question I had was like, what should he buy? Yeah, the other thing is like, you know, seeing all the different volume across Amazon and being like, well, it'd be pretty convenient if that was made here or this was made here, or at least some of the components were made here. And it wasn't, you know, you didn't just have assembly here. Yeah. I mean, yeah, there isn't, I hadn't considered that narrative, but there is a different flow where you're - He's not sitting there being like, great, I love that so much of our volume is having to be shipped or flown across the ocean. A ton of that, yeah.

7:41John Coogan:And then also he's seen attacks from Timu and Shein and Alibaba and AliExpress and a whole bunch of direct from manufacturing plant to customer. And how do you respond to that? Maybe getting into the deeper in the supply chain makes sense. Although I'm not entirely sure that there's so little reporting at this point. It's the only line is that he'll be buying manufacturing companies. So there's a lot in the supply chain, so I wanted to dig through some stuff that could potentially be in the crosshairs for Bezos. The first is Lear. They manufacture seats and electronic systems for the automotive industry.

8:21John Coogan:And a lot of these companies have huge revenues and small market caps. So Lear did$23 billion in 2025 revenue. The market cap is around$5 billion. And so the business is operationally complex. I mean, on a P.E. ratio, it trades at, like, I think, 13 priced earnings. And so, like, you have these high revenues because you're in the manufacturing business, but you're low margin. Yeah, and this has been some of the concern around VCs investing in manufacturing businesses. Yep, are they used to it? And just a general concern that, you know, they might today be getting valued on a revenue multiple. Yep.

8:55But as they get to the public markets, people will expect them to be valued on earnings. And that is a very different story.

9:02John Coogan:BorgWarner is another example. They're a scaled auto supplier. $14 billion in 2025 revenue. Market cap is 9.5 billion. They're more focused on propulsion and power-related systems. They literally make turbochargers for internal combustion engines, although they have shifted their business a lot to be focused on electronic components, both in terms of making battery packs, motors. But they're actually already getting into selling turbine generator systems for data centers. They're part of the AI boom already. Hexcel is another company. They're guiding to$2 billion in revenue, but the market cap is$5 billion.

9:40John Coogan:So a little pricier on a price to sales ratio. Here's a fun one, Goodyear. Not only do you get blimps, which has gotta be important if you're trying to go blimp for blimp with Sergey Brin, you just jump to the front of the line. It's only a$1.8 billion company. This revenue multiple is criminal. Yeah, it's only, so the market cap for Goodyear is$1.8 billion and revenue was$18 billion. end so they're they're training at a point somewhere out there there's like a vc funded uh tire company it's like 100x revenue and they're like we're gonna go through a thousand x multiple compression over the next decade yes you will uh but uh goodyear is sort of the classic ai for manufacturing they gotta focus they gotta get investors thinking about their advertising blimp monopoly yep and scaling that business yep uh so quality control is really critical uh there's opportunities for downtime optimization.

10:35John Coogan:So if you can have better systems that limit downtime, you can produce more tires. The tire market is extremely competitive. You see Chinese tires on cars more and more. And so these are thin, thin, thin margin companies. And that's why the multiple is so low. But getting good here. Scariest moment of my life in a car. Chinese tires. Old Chinese tires. Old Chinese tires. But a lot of people go in and they say, I need new tires. Just give me whatever's cheapest and so if you're competing on price and you're a low-cost supplier every dollar counts. On the bigger side you have Rockwell automation.

11:08John Coogan:I don't know if you've ever seen those videos for the Rockwell turbo and cabulator. Have you ever seen these? Look at this. Research has been proceeding to develop a line of automation products that establishes new standards for quality, technological leadership, and operating excellence. With customer success as our primary focus, work has been proceeding on the crudely conceived idea of an instrument that would not only provide inverse reactive current for use in unilateral phase detractors, but would also be capable of automatically synchronizing cardinal gram meters. Such an instrument, comprised of Dodge gears and bearings, Reliance electric motors, Allen Bradley controls, and all monitored by Rockwell software, is Rockwell Automation's retro encabulator.

11:48Now, basically, the only new principle involved is that instead of power being generated by the relative motion of conductors and fluxes, It's produced by the modial interaction of magnetoreluctance and capacitive directance.

12:03John Coogan:Capacitive interaction. Why so reluctant? The original machine had a base plate of pre-famulated amulite, surmounted by a malleable logarithmic casing in such a way that the two sperving bearings were in a direct line with a panometric fan. Panometric fan. The line-up consisted simply of... It's just like such a funny in-joke for electrical engineers because it's just a whole bunch of like slop basically. Like, none of those terms mean anything. Good startup launch video concept. Yeah, yeah, yeah. It really, I mean, it helps you detect, like, are you in group, out group, basically. Very, very fun.

12:34John Coogan:But they are... You're saying in out group. What? Well, were you able to clock it or were you like, oh, that sounds good? Yeah. I mean, it sounded silly. It sounded silly. So Rockwell Automation is on the bigger side, but they sit right at the intersection of factory automation software and control. So although that video is a joke, Rockwell does work on factory automation already. So the thesis with this one is that the business is already dedicated to industrial automation. And so it's less of a turnaround, but it's a control point for pushing AI to thousands of other factories here deeper in the supply chain.

13:09John Coogan:It's expensive at$40 billion, but potentially in the budget. Do you think there's a world where he just buys a Ford Motor Company? That would be crazy. Obviously, I mean, he would have to lever up. What is Ford now? It's 45, but he's not going to just drop, like, raise$100 billion and then buy a company entirely with equity. I feel like it will be deeper in the supply chain than a brand, but I don't know. It's possible. It's possible. But he might say. Isn't he already a big investor in Rivian? Yeah, sure. So, like, looking deeper into that supply chain, and there's a few other car companies that he's been involved in.

13:49John Coogan:I just don't know that that's where the big opportunity is. It might be deeper in the supply chain, like bending the metal that goes into the bumper, this and that, and it's 12 steps deep. It's more boring, but it's even thinner margins, less brand risk and brand value, but really focused on applying that actual commercial excellence. And then there's also the option that we get some sort of Elon-style megacorp with Blue Origin involved at some point, Like, you know, XAI and SpaceX merged. It's possible that Blue Origin and this vehicle come together with Project Prometheus in some way, and then that goes public as a new entity.

14:28John Coogan:A lot of these, he's actually mentioned the whole Space Data Center thing before, I think a while ago. Wasn't it like more than a year ago? Yeah, I think they also just filed something with the FCC. Okay, yeah, for the new cloud. Yeah, it was like 50 ,000 satellites, I think. Yeah, yeah, so he wants like a direct Starlink competitor. And so when you think about the full AI stack, if he's really off-planet pilled and thinks he can get there, even if it's like a year or two behind, or I guess, what's the timeline for his Starlink competitor? That would be like three years behind, four years behind?

15:05John Coogan:Yeah, I mean, for the moon, he's supposed to be ahead, right? Oh, yeah. Maybe this will all go on the moon. Well, either way, it's a big move from an experienced operator with lots of opportunity ahead, and it'll be fun to follow along. And I'm glad he's back in the game in a big way, not just doing random side projects. Bottle service. Or bottle service. He's got to be spending more time and energy securing America's future than securing bottle service. Well, you know, I mean, the guy can take a weekend off, but one photo from his weekend flies a lot further than, you know, he could be completely locked in and just grinding 80 hours and then he goes gets bottle service just one time.

15:46John Coogan:And let's go over to Jensen on the All In podcast. That$500 ,000 engineer, at the end of the year, I'm going to ask him, how much did you spend in tokens? And that person said,$5 ,000. I will go ape something else. Yeah. Right. If that$500 ,000 engineer did not consume at least$250 ,000 worth of tokens, I am going to be deeply alarmed. Okay? And this is no different than one of our chip designers who says, guess what? I'm just going to use paper and pencil. I don't think I'm going to need any CAD tools. This is a real paradigm shift to start thinking about these all-star employees. It almost reminds me of what we learned in the NBA when LeBron James started spending a million dollars a year just on his health of his body and maintaining it.

16:33That's right. Here he is at age 41 still playing. It really is, hey, if these are incredible knowledge workers, why wouldn't we give them superhuman abilities?

16:42John Coogan:That's exactly right. Where does that go? If we extrapolate out two or three years from now, what is the efficiency of that all-star at NVIDIA and what they're able to accomplish? What do they look like? Well, first of all, things that, wow, this is too hard. That thought is gone. This is going to take a long time. That thought is gone. We're going to need a lot of people. That thought is gone. This is no different than in the last industrial revolution. Somebody goes, boy, that building really looks heavy. Nobody says that. Wow, that mountain looks too big. Nobody says that. Everything that's too big, too heavy, takes too long, those ideas are all gone.

17:22John Coogan:You're reduced to creativity. That's right. What can you come up with? Exactly. Now, the question is, how do you work with these agents? Well, it's just a new way of doing computer programming. In the past, we code. In the future, we're going to write ideas, architectures, specifications. We're going to organize teams. We're going to help them define how to evaluate the definition of good versus bad. What does it look like when something is a great outcome? How to iterate with you, how to brainstorm? You know, says guy that sells shovels says you should spend 50 % of your salary on shovels. But of course, Jensen's point is generally correct, which is that you should give your best people a lot of leverage.

18:04John Coogan:Yeah, I do wonder what the leverage ratio is in other industries. if you're a crane operator and you're making, I don't know. If you should give him the best possible crane. Yeah, like is it possible that if you're operating a$10 million crane or something, I don't know how much a crane costs, or like a cargo ship that's ferrying oil across the world, how much does that crew cost? And then how much does the ship cost? And then what's the depreciation on that ship? Because it might be like$100 million ship that will last 20 years. So you're looking at$5 million a year. And if the crew, total cost of the crew is like$1 million, well, then you're actually spending more on the capital asset than the underlying talent.

18:50John Coogan:And you're getting like leverage on top of that. And for a long time, software engineering has not been that. It's been, you know,$100 worth of internet and a couple thousand dollars on a MacBook and a lot of open source software because you're using Python and Linux and these things. And I mean, I guess you could maybe burden like your cloud budget once you actually deploy the app. Like if you think about if you're a Facebook engineer or Instagram engineer, you could push a feature that consumes a ton of compute. But this is also going back to like the AI talent war where you get back into if you're an elite software engineer and you're like, look, I'm going to be in charge.

19:26John Coogan:I'm going to be managing half a million dollars worth of token budget over the next year. Well, I want comp that's higher than that or something like that. It'll be very interesting to see how the value works there. A reasonable ballpark for the annual cost of a fully loaded cargo ship is around$4 to$18 million, including the cash operating cost plus the depreciation. So I don't know. In other news, the U.S. Department of Justice announced yesterday three charged with conspiring to unlawfully divert cutting-edge U.S. AI technology to China. The indictment unsealed today details alleged efforts to evade U.S.

20:08export laws through false documents, staged dummy servers to mislead inspectors, and convoluted transshipment schemes in order to obfuscate the true destination of restricted AI technology, China. said John Eisenberg, Assistant Attorney General for National Security. These chips are the product of American ingenuity, and NSD will continue to enforce our export control laws to protect that advantage. So the company, SMIC, Super Micro Computer Inc. Co-founder. They caught him red-handed.

20:45John Coogan:His name's Wally. He was arrested today or yesterday. He personally holds half a billion dollars of super micro stock, still risked it all, and now he's facing 30 years in federal prison. That is crazy. He was charged with smuggling billions in NVIDIA servers to China, used Southeast Asian Shell Company to funnel 2.5 billion in servers to Chinese buyers, 500 million worth shipped in just three weeks in spring of 2025. That's a lot. two and a half billion in servers feels like enough for like a frontier training run like that's a big big that's a big push built thousands of fake dummy servers to fool u.s compliance auditors caught on surveillance camera using a hair dryer to swap serial number stickers and so ox gigi says this man is a billionaire and was removing labels with a hair dryer personally you're simply not grinding hard enough there's always there's always a grind set lesson in any story let's take this over to linkedin it's also notable because there's the export uh tax or tip that you know that you're trying to bring chips out of the country that the u.s government kind of flips over the square terminal and they say tip please 25 so you're looking at at hundreds of millions of dollars if these were even if these were chips that were able to be We were debating the flipping the iPad around asking for a tip earlier today.

22:12John Coogan:And I think our joint stance was if the iPad is turned around, you got a tip. Yeah. It's just the etiquette is that you got a tip. But is there any is there anywhere where that line crosses and you say, I can't, I can't possibly push something. Funny thing is Erwan asks for tips. Really? On online orders. Online orders. Wait, but there will be a human delivering it or? No, there's two separate tips. Okay. There's the person in the store who puts the things in the bag and then there's the delivery person. Yeah. And they flip it over twice. Yeah. The game theory of this stuff is always hard because I feel like, I don't know, there's regulations on like taxes on tips, but it's very unclear if companies actually funnel, how they distribute the tips.

23:00John Coogan:Like, are you tipping the person? are the tips grouped together were your tips pooled when you were uh yeah they were pooled so even if you did a great job jobs at the hotel the valets yeah they pool yeah bell staff the people that like take you down to your room do not interesting huh and that was just the rules oh it's because i might be the one who parks the car you might be the one who gets well in a dynamic where you have a bunch of valets that are waiting for cars yeah it would be a bad experience for the guests if the valleys are like they see a nice car pull in and they're like fighting over themselves sure to deliver service right it should just be like whoever's available should deliver the best possible service whereas bell work is much more one-on-one okay it's like i'm kind of your guide on the property i'm gonna like you know you you might have the person's cell phone they might be texting you etc interesting and so it makes more sense to not pull uh when a when a broker who had bought Nvidia powered servers from the Southeast Asian company sent Sunny a text message containing a link to an announcement about Chinese nationals being arrested for smuggling AI chips into China.

24:11John Coogan:Sunny allegedly responded with sobbing emotions. Scroll up, scroll up so people can see. Morning Bruce says, hey man, you're probably going to jail. Super Mocha co-founder crying emoji. And scroll down and you'll see Nick Carter, literally this, your son has passed away. It's the Warthog explosion. I'm sorry, but the Supermicro thing is awful, but parts of it are genuinely hilarious. They literally used a hairdryer to move serial numbers from real servers to dummy servers to throw in a warehouse and got caught on camera. And the pictures have actually leaked. Literally caught red-handed. Yes, it's a red, it's a red hairdryer.