SpaceX, AI Bubble Fears, and The Age of the Trillion-Dollar, Zero-Profit Company

Today, we're entering a new age for investors. The age of the trillion dollar company without profits. On June 12th, this year, SpaceX went public in the largest IPO in history, pricing initially at roughly $1.8 trillion and quickly trading above $2 trillion. Over night, it joined the tiny club of companies valued alongside Apple, Microsoft, Nvidia, and Amazon. But those companies have something in common that SpaceX does not. Profits. Apple and Microsoft each generate hundreds of billions of dollars in annual revenue and enormous earnings. SpaceX, on the other hand, reported less than $19 billion in 2025 revenue and a $5 billion net loss. Open AI and Anthropic are both moving toward public listings, each with valuations approaching $1 trillion. Anthropic is not even trying to produce a sustained profit this year or next, and Open AI reportedly does not expect to be profitable until the end of the decade. The striking fact is not just that investors are paying up for growth. They've always done that. The striking thing is that public markets are soon being asked to absorb several trillion dollar companies whose valuations are built less on profits than on the expectation that artificial intelligence will define the next era of capitalism. There's a couple of reactions I think that you could plausibly have to this news. One is, holy shit, this makes absolutely no sense. Investors are insane. We're in a bubble. It's all going to pop. Stuff, you're mattress. And I'm not going to tell you that that take is obviously wrong. That's not obviously wrong. But I'm also interested in what these trillion dollar IPOs represent, which I see as a new age in American technology. The tech giants of the 2010s, Facebook and Uber, Airbnb, they were asset light. AI is asset heavy. SpaceX's AI unit burned through $8 billion in the first three months of this year alone. When AI reportedly lost tens of billions of dollars last year, overall, the so called hyperscalers, the big tech companies like Microsoft and alphabet that are spending hundreds of billions of dollars a year, they are devouring chips and pouring concrete and building data centers with electricity bills that could power small towns. And this turn within technology from light tech to heavy tech, I think might be one of the most important transitions in American economics, markets, and investing. Today's return guests are Michael Batneck and Ben Carlson of Ritholtz Wealth and the hosts of the podcast Animal Spirits. We talk about the new age of asset heavy tech and the dawn of the trillion dollar zero profit company. I'm Derek Thompson. This is Plane English. Ben Carlson, Michael Batneck, welcome back to the show, happy to be here. So as I said in the open, I'm really interested in a couple trends in tech and markets right now. And they really all revolve around this one phenomenon, which is that SpaceX, Anthropic, and OpenAI, all IPO in one year means that we're going to get three trillion dollar IPOs with a combined profit well below zero. And I think this is just a sign that we're in a brave new world that requires some explaining. So let's start with SpaceX. The company IPO is at just over $2 trillion valuation already down about 25% from the peak. This is a firm with no profits, enormous costs, great satellite business attached to a money burning AI business. Michael, what are investors buying when they buy SpaceX? I'm going to answer a different question and excited to be back here. All right. Forget about what are they buying. I think everybody understands the Starlink system is phenomenal. It's a game changer. Forget about the $32 trillion tamer, whatever they spoke about with like the AI opportunity. I think for the point of view of the investor base, what's really important is the $2 trillion market cap. Holy cow, I miss my opportunity to invest in SpaceX at a $300 billion market cap, right? Whatever Facebook came public, like we miss all of that growth because companies are staying private for way longer. And now I'm being as an index investor, you're jamming $2 trillion of a company with a very with a lightning rod of a leader that I potentially do not care for. And I am now his exit liquidity, okay? So the story is is simple, complex and nuanced. You, the index fund investor, are not swallowing $2 trillion worth of Elon Musk. The Nasdaq 100 is fast tracking inclusion into the index. But what they are doing, and what I say fast track, usually there is a seasoning period where a company will have to show something like four quarters of something crazy called a profit before they can get included into the index. The Nasdaq is saying, listen, we are the innovation exchange of the world. How can we not fast track and include SpaceX? The question becomes how much? And there is this thing called the free float adjustment market cap where they are not just put $2 trillion and they will say, okay, how much stock actually exists for the public to buy? In the case of SpaceX, Elon Musk owns 40-something percent of the company. That will never be included in the market cap. So it's not to downplay and say that the $75 billion worth of proceeds that are being raised is not a lot, because it is a lot. But you are not, it's not like, so SpaceX can be the fifth largest company in the world by market cap. And it could be, if and when it's included in the S&P 500, whenever that may be, it could be the 180th biggest stock. And I think the part that people got upset about this is they came to market. They issued 3%, so they raised 3% of the company was what they issued. Normally, when a company is IPO, it's 30%, right? And so the thing that is, I think, grinding people's gears, forget like past the $2 trillion number, it's, hey, wait a minute, they know that they were such a gigantic amount of demand for the stock. And they are artificially limiting the supply because they want the demand to so far outstrip supply, because in the lockup schedule, which how many shares can insiders sell over time, there are hurdles such that if the share prices say 20% above or 30% above for ex consecutive days, then they can sell more stock to the public. They can get unlocked faster. And I know I just said a lot that's probably confusing to a lot of people, but it's a, it's a big story, but the headline numbers miss a lot of the new ones underneath it. No, one way that I recapitulate that is to say that look, most times when people, when companies go public, they go public by issuing 30% of the available stock. In the case of SpaceX, they didn't let out 30% of the stock, they let out 3% of the stock. And so there was a pressure created by that scarcity to drive up the value of those shares. And when the value of those shares is driven up to a certain number, I don't know what it is today, 175, 180, well, that calculates that that cumes to a $2 trillion valuation. But by only releasing a small amount of a stock that they know has enormous demand because Elon Musk is one of the most famous people in the world. That was one way that they knew they could get this $2 trillion valuation. Now over the long run, I think people, traders, hedge funds, institutional investors should probably think, right, like what is this piece of paper worth? What is this share in SpaceX worth? Do we think it is worth a share of a $2 trillion company? Or when we look at the fundamentals here, are we looking at something that makes more sense as a $1 trillion, $800 billion company? So Ben, I want your SpaceX thoughts as well. And we already talked about the famous page 11 or reference, the famous page 11 of the S1 document, which is the one that got everybody's attention that claimed that SpaceX had a quote, quantifiable total addressable market of $28 trillion that consisted of $370 billion in space-related stuff, what most people think of with SpaceX being a rocket company, $1.6 trillion in starlink, satellite communications, and 26.5 trillion in AI. So I do think it's important to slow down here and say, most people, when they think about SpaceX, think of SpaceX as a rocket company, but when you look at what SpaceX defines SpaceX to a TAM as, they're saying, you're 95, 99% of the value is the future of AI business. And is this, is this a, best understood is just an Elon meme stock or are you bullish on some aspects of the company's earnings perspective here? I think I'm more bullish on tech leader's ability to sell the future to people. And I think this kind of goes back to Jeff Bezos in a lot of ways. He was the first one who said, listen, don't worry about right now, worry about the future. We're going to grow into this valuation. And it took a while for investors to get used to that. You remember in the dot com bubble, Amazon fell 94% at the end of the dot com bubble. And people weren't quite there yet. But for years and years, people kept going, why is Amazon's valuation so high? What is it going to grow into? And finally, the cash flow is sort of coming into it. And I think Elon Musk has taken that to another level with Tesla, right? Everyone who's a fundamental analyst for Tesla for years has said, this thing is a joke. He keeps issuing equity. And when a company's issue equity, the current investors get diluted. So the value of the company should go down to you. Your earnings per share is being spread among a wider set of people. And he's been able to sell the fact that no, no, no, this equity will help me grow the company. And that's what he's trying to do with SpaceX as well is say, listen, don't worry about right now. We're worrying about the future. And so if you're a fundamental analyst and you looked at the numbers, you go, this is absolutely insane. Look at just look at the numbers. They made what 19 billion dollars and they're going to be a two trillion dollar company. It makes zero cents until you start going, well, what about data centers in space and what about trips to Mars? And that's where he sold enough people because if you look at it, you talk about the IPO, 30% of the demand from the IPO came from retail investors who obviously love what this guy does. And usually it's, I don't know, 5 to 10% for IPOs. So he has talked enough people and he's such a good showman in salesman. Obviously, you know, he's landing rockets too. It's not like there's no, nothing going on. But I'm more bullish on his ability to continue to get investors to buy into his vision of the future. Yeah. I mean, Michael, I am interested in your general theory of Musk valuation effects because, you know, as Ben just said, it's not just SpaceX. I think Tesla's price to sales ratio that I just looked up as 14, the automotive industry average is 0.7. So that means that Tesla's valuation as a company is approximately 20 times higher than you would expect from one of its competitors, just based on its sales alone. You've already indicated, I think, in your previous answer at least two different ways that Musk achieves this valuation effect. One is this ability to tell stories, this ability to inspire enough people to buy shares in his companies. And another, I think this is important is, especially in the case of SpaceX, just a few weeks ago, understanding that he has built a certain demand for owning shares of SpaceX, issuing a pittance, a small number of shares creates a kind of demand to supply ratio that drives up so much pressure to own some of this stock that it ends up cashing out in a company worth, you know, 2.x trillion dollars. But I am interested in your general theory of how Musk is able to do this. To get not just with Tesla, trillion dollar car company worth 20 times more than its competitors on a sales basis, but also SpaceX, 2 trillion other company with zero profits. A few things. I just want to say one thing about the lockup. I don't know that the current way that business is conducted is the best way. Typically, you have a 180-day lockup to which shares predictably in the, like, on average, the first of these companies fall into the lockup as supply is coming on the market, which is very well telegraphed to everybody, literally it's on the calendar, and then shares find the bottom and do whatever they go on to do. So I kind of like that they're thinking outside of the box, being creative. I don't know that I love that it's only 3%, which is very, that that was not an accident. The question that you asked about Elon's ability, unique ability, and I would say genuinely a one-of-one ability to do this, is he's a world-class entrepreneur, operator, money-raiser, deliverer of, he's made a lot of people a lot of money. And it's not just a retail investor, because Palantir has done something similar. Alex Carpaz has garnered a lot of attention of perhaps the next Elon. Guess what? Shares of Palantir are down 50% right now, and it's not so that Tesla hasn't experienced its fresh air of crashes. It certainly has, but he has the unique ability to continue to issue equity. There was times in 2020, 2021, where Tesla was raising money, and the stock would go up 10% because it was like, see, look how much demand there is for Elon shares. That has to be bullish. So he is a one-of-one. I think that there are some read-throughs into the rest of the market and the narratives that we're telling ourselves, but I genuinely think he is a one-of-one. There's nobody else in the world that could do what he does. I want to move the conversation from SpaceX to the other two IPOs that are expected later this year, which is open AI and anthropic, both of them racing toward what are expected to be trillion dollar public offerings, instead of staying private and dodging quarterly earnings, which is what most valuable private companies have been doing for the last 10, 15, 25 years. I see this personally as a pretty major inflection point, and I don't necessarily know what the meaning of that inflection point is, and that's why I wanted to talk to you about it. So a couple of thoughts here. One thought is that tech used to be asset light. The companies that defined what we thought of as technology in the 2010s, Meta, Uber, these were asset light companies, but the companies that we think of as the frontier of tech today in AI and certainly in space, they are desperate for cash because they are capex heavy businesses that have to spend a lot on chips or infrastructure or steel and rocket technology itself, and because they're desperate for liquidity, they are going public faster than their slightly older tech brethren would have. And that's an interesting inflection point to watch tech going from being this asset light software forward business to being more of a asset heavy business. I think that's interesting. The other thought that I have is like, this is freaking insane. This is crazy. $4 trillion of collective valuation between three companies that will in 2026 in all likelihood be probably, I don't know, negative 50 billion, negative 60 billion in terms of net earnings. That's a crazy thing to happen in one year. So then without allowing the prosecutor to lead the witness too much, I just wonder where your head is at here. But understanding that as there are market historians, I think, that are going to look back at 2026, whether there's a crash or not crash and say, wow, that was the year of the three IPOs worth $4 trillion, which for the first time created trillion other businesses, three of them with zero profits. Like, what does it mean to you? I mean, think about how fast these companies have grown up too. This has all happened in the span of four years, essentially. These are the anthropocas, potentially the fastest growing company in history in terms of revenue. You're right. It hasn't translated into earnings yet. I think the ability for them to do a perfect baton handoff from all this cat-packs and you're right, going from high margin businesses very efficient to, all right, we're going from intangible to tangible, like we've done a full reversal. That's why all the companies in the 70s and 80s, the valuations, they had to be so much lower in them because the margins were so much lower, right? So it makes sense that companies are more highly valued today, but we're seeing this trade off from all these companies. And there was a chart that went around this week from Nomoro that was totally viral about taking the big hyperscale, it was Amazon, Oracle, Microsoft, Google, and Meta, and showing that their free cash flow has gone from $750 billion to essentially projected to be zero almost by the end of this year, because they're spending so much money. Because we're going to get that a second. Can you slow down and just explain to people, like what that actually means, that free cash flow among the hyperscalers has basically gone from hundreds of billions of dollars toward projected zero in late 26, early 27. Well, these companies have just been cash flow producing machines, they need fewer employees, they have really high margins, right? And they've had this cash flow to just continue to, that's how they provided such high value to their shareholders. And they've said, wait a minute, for this AI build out, we're going to use all this cash flow and we're going to build out all these data centers, right? Because there's so much demand for compute, there's so much demand for these LLMs and businesses want to use it. And so they've said, we're just going to take this cash flow and reinvest it into our business, which is honestly something that people for years have been crying about, like stop, tell the CEOs stop buying back their own stock and actually invest it in the economy. Well, these companies have done this to a degree we've never seen before in history. They're actually reinvesting in their business at a scale that's unprecedented, which is very risky, but also kind of interesting when you think about them, what they're trying to do. This is such a bouncing act, they're trying to thread the needle here. And what we've seen is that the ones who are benefiting from it are not the companies who are doing this, it's other companies in the memory, semiconductor space, right? Like Microsoft and all these companies, their shares have been crashing. And so it's interesting to see that them spending this money hasn't been rewarded yet. They're not getting an ROI on it and their stock share price is hurting to the benefit of other stocks. It's not just that they're not being rewarded. I think they're being actively punished. I think Michael, if you look at the S&P 500 without the Magnificent 7, so the S&P, you know, 493, it's up 678% this year. But if you just look at the Mag 7, which had been driving the majority of earnings and stock market gains in the previous 18 months, they're down six or seven percent year to date. What do you make of this inflection point with regard to the Mag 7, which clearly has something to do with what Ben just mentioned, which is the free cash flow of these extraordinary businesses of Alphabet and Microsoft, just 40% profit margins, all the cash in the world. And they are spending through, through, through all the way down towards zero to build out the AI infrastructure that they hope will be their next big business. What do we make of this? Microsoft, which is the biggest? software company in the world. It's a lot of other things that is a gigantic cloud business, but it is the biggest software company in the world. It isn't a full-on crash, Derek. The stock is down 35%, but what's happening that is so interesting is it is happening while the stock market is basically at an all-time high. So if you compare it apples to apples, and you say, okay, you divide the price of Microsoft by the price of S&P 500, and you look at it on a relative basis, Microsoft, the ratio of Microsoft to the S&P 500 is back where it was in 2019. Their performance has been the same since 2019, and Microsoft kicked the crap out of the S&P, so all of the gains have been ripped away because of the transition that you're mentioning. So what we're seeing now is the 27-day spread, I know it's a random stat, but over the last 27 days, the S&P 493 have outperformed the Mag 7 by 18%. That's a gigantic number, an 18% spread. And again, the S&P 500 is basically at an all-time high. When you've seen similar levels of dispersion since 2018, the S&P 500 was in a full-on bear market, which makes sense because we were so reliant on the Mag 7 to carry the gains. So there's one more thing that I want to say about the IPO market, and this ties into, it's a bull market story. That's what we're talking about here. Because Elon getting all of the attention for being the first trillionaire, there's a lot of people that are very upset, and I understand where the sentiment is coming from. I am a capitalist, and I think Elon's wealth creation, even though I'm not an Elon Stan, I think wealth creation is incredible. I am also like a social safety, that person is a good thing too, so I'm not saying that people that don't have means are deserved, whatever. But I think that capitalism works really well, okay? I think wealth creation is a good thing. In terms of the current IPO landscape, this is very, very important. You have to normalize. There has to be a denominator. So we're talking about comparing today versus the dot-com bubble. When you have all these IPOs that would go up 300% in a day, that's not what's happening today. Yes, the scale is a lot larger. You're talking about $3 trillion plus companies, but if you look at the aggregate IPO proceeds, so actually how much money was raised as a percentage of the starting market cap, this looks nothing like the dot-com bubble. Now, it'll change a little bit as Anthropic and OpenAI come out, but it's like a tenth the size. If you just compare the aggregate IPO proceeds, so in SpaceX's case 75 billion, as a percent of the starting market cap, it looks nothing like 99. Obviously, in other ways, it does, but I think you have to talk about the denominator. You referenced previous bubble crashes. We were absolutely getting the possibility that AI is a bubble and about one and a half to two questions. Ben, I think there's a way in which we told the story that if someone hasn't been following the stock market in the last six to 18 months, they'd think, if you're telling me that the S&P 500 is up, but the big hyperscalers that are building out AI, a lot of them are down like meta and Microsoft, more than 20%. Sounds to me, like AI is getting its butt kicked right now in the markets, and that's not entirely true, because through 2025, the story in AI from an investor standpoint was really compute compute compute. You had NVIDIA going crazy, the hyperscalers going crazy, cloud capex, all of that was flying. That's when Microsoft and the rest of the hyperscalers were leading all the S&P 500 gains. But now what you're seeing is those trades are moving into other sectors within the AI ecosystem, into memory and energy and cooling, the chip maker, microstock, tripled between March and June. And so I wonder if you can help us tell a story here, which keeps the following balls in the air. On the one hand, a lot of the hyperscalers are getting kicked in the teeth. On the other hand, there's a lot of other stocks that you could think of as AI stocks that value is flowing toward, which suggests that what we're seeing is not so clearly the demolition of the entire AI story. How do we make sense of this? I think today is a perfect example of what's going on. And so there was a story today on Bloomberg where Apple announced that they are raising the prices of everything, iPads, Mac, desktops, right, laptops. And it said the reason is because there's a short of memory chips and storage, right, there's not enough compute to go around. So Apple fell 5%. Pretty big fall for one of the biggest companies in the world. Micron today, a company that does that, they produce the memory chips, they're up 16%. So what you're seeing this trade off between these big, huge hyperscalers, and in the last year, Micron is up almost 1,000%. In one year, Sandisk, a company that is in the same business is up more than 4,000%. Let me repeat that. 4,000% stock price gain in one year, where as Michael said, Microsoft is down by a third. Oracle is down by a third. Meta is down by 25%. So you're seeing a trade off from the, you're seeing a divergence in the winners and losers now. And it's like these companies are sowing the seeds of their own demise. They've almost, they've almost gone in and disrupted their own businesses. Now the hope is, hey, listen, of course, this CapEx is falling now. It's going to come back. You know, this is a 2028, 2029, 2030 story. They've all said we, the bigger risk for us was not going hard enough on this. We weren't just going to sit back and let someone else take corner of this market. If one of us is going to jump off the bridge, we're going to hold hands and I'll do it at the same time. But you're seeing this change in leadership that we haven't seen in many, many years because people have been so worried about, hey, it's only these Mag 7 companies that are powering the stock market higher. And that's just not the case now. The stock market is in a totally different place in terms of its leadership now. I understand the story that there is an era where the hyperscalers were up. And now hyperscalers are hurt. And because memory is scarce, memory stocks are ripping. But at some point, who's buying all those chips? Who's buying the memory? It's the hyperscalers. It's the big boys. At some point, can the memory stocks keep surging if the hyperscalers keep dropping? Well, this is a great question. Another difference between the.com bubble and today is that a lot of the build out, a lot of the fun that gets not just retail speculation. It is the hyperscalers. The absolute most profitable companies that the planet has ever seen are funding this build out. And there is no sign, despite Microsoft's 35% pullback, there is absolutely zero sign that they are pulling back anytime soon. So Ben mentioned micron. They're up 16% today, not just because of apples raising prices, but because they reported earnings last night. And they reported $41.5 billion in revenue. That is up 74% quarter over quarter. It is up 346% year over year. And guess what? It is a very, very profitable business. So in April, all the rewind the clock back to April of 2026. I asked one of my guys, hey, show me what microns 12 month earnings per share has done over the last 12 months. And it went from $9 a share in March of 2025 to $85 a share in April of 2026 when I asked it for it. Today, it is $132 a share. So investors like growth, okay? And we've never seen growth like this. So since the beginning of January 2025, microns earnings are up 1,440%. And investors are slapping a discount on that because the stock is only, and I'm obviously teasing, the stock is up 1,380%. So in general, finance 101, a stock follows or leads, I should say, generally, the earnings. It's a business. That's what we're investing in. It's not just numbers on a screen. This is a business. And the business of micron, the performance of the underlying fundamentals has matched the share price. Now it's not to say that people aren't speculating their assets off they are, but it's rational. And one of the things that micron said was we now expect supply demand conditions for both DRAM and NAND to remain tight beyond calendar 2027. So this is the question, Derek. You can drive a truck between these two things being true. Hey, this is an obvious bubble. Hey, wait a minute. How much of the world is actually on the AI train right now? Is it 2%, is it 3%, so it's very hard to square those two circles of, wait a minute, it's so early, but micron is up 1400%. And that is where investors are going to twist themselves into a pretzel, trying to figure out where this thing goes. Here's what I want to do now. We've thrown around the B word a few times. I want to make the best AI bubble case as I can see it. And then I want you to evaluate it, maybe tear it down. And Michael, because you're championing it the bit, well, we'll let you get the get the first swing at the pinata and then you can follow. These companies, the big boys, the hyperscalers, they're spending 600, 700 billion dollars a year with no end in sight. And the valuations of Microsoft and alphabet and meta are predicated on a certain operating margin that assumes that eventually what we'll need is $1 trillion a year in de duplicated external AI revenue quite soon. We are not there yet. The latest bottom up estimate from a Zim Zarr's exponential view put real duplicated AI spending at about 175 billion annualized this year. So we need to roughly quintuple the AI industry to get to a point where the hyperscalers can expect to make from their investment. Income, which maintains the cooperating margins that currently justify their business and their valuations. So the bar is enormous. The bar is enormous for growth. As fast as anthropic is grown, the bar is enormous still for growth. And there are at least three things that I can see pushing in the wrong direction. Number one, free cash flow falling at the hyperscalers means they're going to have to dip into debt. They make a lot of money. They've got a lot of incoming earnings, but they're going to have to dip into debt. I think somewhat soon. Number two, you just said it. Memory and chip costs are climbing. So it's not as if this mountain is going to suddenly get a little bit shallower. It's going to get steeper. We might see continued increases in necessary cap expending of the next few years, $700, $800, $900 billion annualized. And then finally, costs are going up. What's happening to revenue? Well, look, you've got these cheap Chinese open models that are just a few quarters. Maybe a few years behind the frontier models in terms of capabilities. And they give customers a place to go. The costs a lot less than what Claude and OpenAI are offering. And so when you put all of that together, less cash, higher costs, slower revenue. Why isn't this thing going to crash? When you say this thing, let's be specific, are you talking about the entire stock market? No, let's just talk about the AI infrastructure project. Why isn't there going to be a major correction as these companies recognize that the thing that they're building is impossible to stomach given the operating margins that they are required to hold on to to justify their valuations? So the hyperscalers, which is maybe the list that you're talking about or maybe not, and if you collect them, they are trading at a forward PE, so a multiple of what their forward earnings are projected to be, is that a multi, multi, multi year low. All of the risks that you are describing are very, very well known and understood, which is why if you want meta, the second best advertising platform that the history has of the globe is ever seen, you can buy it for 17 times next year's earnings. You have never had the opportunity to do such a thing. I don't know, I have been commenting on the market for a long time, and over the last 15 years, what I've seen consistently is people trying to scare investors because they want the glory of a Dr. Michael Burry, and I would much rather, somebody be mad at me for not quote, getting them out, then to have been the person that caused people a lifetime of earning a return of their money in the stock market. So I'm not going to say that's the stock market may not pull back, that is always a risk. Of course, it will be pullbacks along the way, but we've been having this bubble talk, Derek literally for 15 years. And so quickly, we forget that guess what? There was a bubble in 2020, and guess what happened? It burst. Apple fell 30 percent, Google fell 44 percent, Amazon fell 56 percent, Nvidia fell 66 percent, Netflix and Metal lost 75 percent, Derek, and we forget that it happened. That was a bubble, and we just talk about it, it popped, and we just move on to what's the next bubble. Now, that is not to suggest that you can't credibly make a bubble case for where we are today. And if somebody would say, hey, Michael, you dumb bald asshole, how do you not see a bubble? Here's what I would say to you, this is the data point. If I were to look back on five years and say, oh, shit, okay, this should be the one. Morgan Stanley projects that SpaceX's total revenue could reach $3.4 trillion in 2040. $3.4 trillion in 2040. That would be the data point that I looked back at and said, whoo, I wish I sobered up. That's 10 percent of the current economy essentially. But I see so many people with these legitimate concerns. I don't see, and it's very easy to find excesses pockets of speculation, so I'm uplined. But in general, I don't see people that are just thrown caution to the wind. I see the opposite. I see people looking for reasons to say, hey, why shouldn't I sell? Tell me why this doesn't end in tears. I don't know what time it is on the clock, but to me, that makes me feel like it's not yet midnight. I feel that there's two issues here that I want to make sure I disentangle. One issue is like, I mean, even three. One question is, should we understand AI to be an infrastructural or financial bubble? Number two is, can the typical investor stomach a typical bubble? And number three is like our bubble's common, which is related to number two. I understand that like, bubbles are common, and I understand that bubbles can be stomached because we have a resilient economy and we can bounce back. But neither of those things, I think, goes directly to the question of, is this a bubble? And so Ben, why don't you take a crack at it? I am interested in just what you think. When putting together, just even the well understood ingredients, you know, as Michael said, the less cash, higher costs, possibility of slowing revenue growth as more people switch from the really expensive frontier models to the cheaper models. I think one way that Brian Armstrong at Coinbase put it is that he said, the lower 80% of AI use is going to rely on models that are 99% cheaper, right? That was one way that he predicted than your future. Why should putting all of this together not make us a little bit bearish about the future of AI? So I have cognitive dissonance is this idea that when you have two competing theories in your brain at one time, your mind seeks to reduce that discomfort by picking one of the theories and kicking the other one out, right? I'm just going to latch onto this one. And I have extreme cognitive dissonance going on because you've talked about the railroad bubble in the past. Derek, right? I've written about that. This, the dot com bubble, the telecom build out every innovation in history that had capex relative to the economy of this size. This one checks all the boxes, right? All the spending. Anytime there's an innovation, there's always extrapolation. Michael talked about these insane SpaceX numbers. There's always extrapolation. There's too much excitement. We got everything we wanted and more of the dot com bubble or the dot com boom, but we had to go through the bubble had to burst first to get there, right? If you to told someone in 1997, you're going to have YouTube someday streaming, streaming movies, everything we have on the internet, right? This little thing on your, on your, on your pocket that has wireless internet, like people to go, yes, everything where we want and more is going to happen, but you had to go through the bust to get there first. And so I think that if you check all those boxes, you got to go, oh my gosh, this, you know, this has to be that, but I can, I can only get 75% of the way there because there is so much demand for this and it feels like we're only scratching the surface on how much people in businesses are using it. Now, the problem is, Dario Perkins is this macro analyst from TS Lombard, he, he calculated that 80% of the current revenues right now is just circular. It's spending from the hyperscalers. So this thing hasn't completely jumped into the mainstream yet, right? 20% or so is from businesses actually spending. Could we actually see this magical threat of the needle, Patan Handoff, where this turns into ROI before we have a crash? And I think there's a difference between a bear market, which is, hey, these stocks fell 30 or 40% because this happens in volatile industries like this versus the Nasdaq fell 80% in the.com bubble, right? That's a crash. You know, I, I think anything 50% or 60% of nowhere is a crash. That's what I can, I'm having a hard time figuring out, is this going to be just a bear market when it gives up a little bit or is it going to be a full-fledged crash? And I'm kind of leaning towards just bear market where no one's really satisfied, right? We didn't get a huge bubble popping, but we still had some give back because the hyperscalers pulled back for a time or whatever. That's kind of where I'm, where I'm right down the middle of a fairway on this. Michael, I feel like Ben's hedging there just a little bit. I appreciate the explication of cognitive dissonance and I, I basically, I think agree with Ben, which is that I am incredibly divided on this issue, but I mean, give us a hard take, man. Like, is this, is this a bubble or ain't it? No, this is not a bubble. If you had to pick one stock that represents everything that's happening with the AI build out, it would have to be Nvidia. Would you, would you gentlemen agree? Yes. Okay. Nvidia is trading at 24 times its forward earnings. That is not a bubble. That is not close to a bubble. Now, perhaps the earnings are overinflated or overestimated and it turns out that actually, hey, it wasn't 24 times buddy. It was 50 times. All right. Fine. I suppose that's possible. But I will say, I am happy to have the guy with the egg in his face who says, no, this is not a bubble. I know the signs are there. I know it's very, very, very easy to say, bubble. But we haven't even spoken. We don't, this is beyond the point of this conversation about all the supply chain bottlenecks and the natural governor and I don't mean the governor of 24 times forward earnings for Nvidia. I mean, what production can actually keep up with at Taiwan Semi and Derek, I know you have talked to these experts all the time, you can't just turn it up. And so I think that this supply, this, this, this compute constraint that I'm, everybody's probably really tired of hearing about, get used to it because I don't think it's going away any time soon. When we ran your show in 2022, remember there was the Bloomberg headline like, economists see 100% chance of recession and we were on the show and we kind of did the same thing where we hedged and you said, no guys make a choice and we both said no recession. I think if you're going to do that and hold my feet to the fire, I would also say no bubble. Corporations are so much better run today than they were in the past. I think that, I think that they have the ability to see us through this and the fact that these businesses do produce so much cash flow as it is and they're so diverse in dynamic. I think we've never seen companies that are this mature in this well-run. run, try to pull off something like this, and I think if anyone can do it, these are the biggest best companies we've ever seen in history, ever, and I think if anyone can do it and sort of make this handoff, I think they can do it. So if I push to it, I'll say no bubble. A couple of comments, number one, I like that last thought, which is that we live in a society and swim in a news environment that has overcome with negativity bias. I've talked about this a lot with psychologists, sociologists, sometimes with investors. I mean, their heaven study is including published by Brookings showing that financial news is more negative today, relative to the underlying reality than it's ever been in recorded history. And so there's a way in which, despite the fact that I don't like negativity bias in the news, negativity bias is actually a kind of interesting medicine for bubbles because if you have a financial news system that is addicted to calling out potential bubbles, I would have to think that at the margin, it makes those bubbles less likely. It at least creates a set of conditions that seems quite unlike the conditions of 1927, 1928, where there was such like unbridled exuberance about the future of America never having a recession again about, you know, the people, you know, driving the cabs or the, you know, the horse-drawn carriage is telling their passengers about all the stock that they were buying on margin. That's the opposite of a new system that is sort of coded with negativity bias. And so maybe there is a world in which the negativity bias that is endemic to financial news is at the margins, sort of, helpful for taking, say, the price earnings ratio of Nvidia from 50, 60, which might have been, I don't know if this company existed in 1928 to 24, which it is today. That's point number one. Point number two, I will say this about the Nvidia PE ratio. I was at a party in San Francisco just after I had this podcast interview with Paul Kodroski where he made this really full-throwed of case for why he thought that artificial intelligence was the most obvious bubble of all time. And for the most part, Michael, he or Ben, he, he recited your case. He said, look at every time we've spent, you know, full percentage points of GDP on a new infrastructure project, it's a bubble again and again. It's the canals. It's the railroads. It's fiber optic cable in the dot com bubble. It's always a bubble and I went to this party and I don't want to say this is clearly off the record. There were a lot of folks in the general, like, extremely online AI commentary that were there. And one of them comes up to me, sort of buttonholes me and says, I heard your podcast with Paul Kodroski about how AI is definitely a bubble. Do you, sir, know the PE ratio of Nvidia? I said, I do. I know that it's 31. I just had a conversation with the Zemezar about the case for AI not being a bubble and that was the first case, first point that he made. So for, for all the listeners out there who want to sort of successfully banter with professional investors about whether or not AI is a bubble, definitely Google the PE ratio of Nvidia before you go to the party because it will come up. And the last thing that I want to say and then Michael, I want your, your reaction to this whole, you know, monologue that I'm giving in the middle of the podcast, which is not particularly good radio, but exponential view, which is the great sub stack that Zemezar oversees. They just came out with a report today on the state of artificial intelligence and it's their conclusion, again, that total external AI spending, right? So this isn't open AI, buying the chips from Nvidia and then paying Microsoft for the cloud services, which is all inside of the system. This is me, Derek, you Michael, you been giving open AI money because we say we want to pay for your service. He said that that number right now is about $175 billion annualized, which is roughly what the depreciated value of the CapEx is for a single year. So we are at a point right now where it doesn't necessarily seem like it's a headed toward the obvious, obvious bubble zone. Michael, you made a face when I was quoting some of those numbers, but in general, how do you feel about this, about this general case for AI, not being a bubble? >> I made two faces, my first reaction was, oh, that's not really that much money. It hasn't even started. Every one of your listeners, I'm sure, is playing around with the tools and stuff. But the average American, the average person around the globe, we haven't even spoken about the robots which are coming. We are so, so early. And it doesn't mean that the stock market hasn't digested all the future, it did it in 2000. I don't know what the future holds. But in terms of where we are in the AI life cycle, we are in the top of the first inning. I mean, it really hasn't begun. >> Yeah. I do think what I'm saying, we haven't really seen what's happened with Fable and Mythos and that offering from Anthropic that essentially was advertised or maybe just honestly communicated as being a top-end cybersecurity expert, but based on Silicon. I mean, we haven't even entered the possibility of AI taking a sliver of cybersecurity spending, which itself is a $100 billion business. And so I think you're right to point out that, yes, right now, we're looking at enterprise use cases for coding. We're looking at people like you and me sort of asking questions of OpenAI and Claude on a daily basis. But there's all kinds of applications for this technology that are really hard to see if they really live like two and a half years away from where we are right now. Like two and a half years ago, it would have been entirely reckless and frankly quite stupid to say that artificial intelligence could be used by expert coders to help build artificial intelligence. It wasn't a part of the offering and now it exists. So I think this stuff is difficult to think about, but I did want to get your takes and whether or not you thought it was a bubble. I want to move on to a couple other pieces of the investment landscape. In addition to the hyperscalers being down, so is Bitcoin, down 32% this year, 44% in the last 12 months. And what is going on, you think, with Bitcoin, is to a certain extent, is it like there's a sort of universal conservation of frontier tech energy. And it just like went from crypto to AI. And so a lot of people that were just like really obsessed with crypto for four and a half years were like, nope, sorry, I'm not interested that anymore. I get out of my portfolio. I just want to be long micron like why do you think without a recession and without further interest rate heights, hikes, Bitcoin has had just such a calamitous last 12 months. I do have this theory that investors can only pay attention to one rocket ship at the moment and one risk, right? That's all we have time for is we're going to pay attention to one risk and one risk only. And then we're going to move on to another thing, right? And I think that's what's happened here too, because the one thing you could hang your hat on for crypto and Bitcoin was it's at least a risk on asset. Sure, it didn't hedge inflation, ensure it hasn't hedged against like an economic slowdown or something like that, but it's when tech stocks go up, Bitcoin also goes up. That's been at least the one thing you could hang your hat on. And then this year, it hasn't been the case. And I do think there's something to the fact that this was seen as like a startup technology, right? It's a way to invest in the startup ecosystem of this blazing new technology. And then AI comes along and it's like, oh, there's something here that actually has use cases you can see with crypto, you can't see them crypto, but with the use case. Yes, right? It's almost like, wow, how long have we been waiting for this crypto use case? The big retail boom was in 2017, and that's when people really started to talk about it over Thanksgiving dinner with their families, right? Remember that whole thing where the first Bitcoin huge boom happened. And we haven't really got that great use case like people have been predicting. And I think it is kind of fascinating that in a risk on moment, Bitcoin is not taking off. And for a while, they're people said, well, it's like a software stock, I guess, but then software stocks came back and Bitcoin kept going down. And it's really, and you had gold going up this year earlier in the year or last year and Bitcoin wasn't keeping up with gold. So I think I think it's led a lot of people to perhaps finally question, you know, Bitcoin really was more of this religious cult than anything. It was you had true believers and they could shift their narratives every time something happened. Well, we said it was this, but now it's this and we said it was that and now it's this. And it's like, they're, you're kind of running out of narratives and I think a lot of the investing public has moved on after we got this big boom from a, from an ETF. It's like, okay, now what, AI, actually, you can see it, right? It's right there. I do think a lot of it, like, it's just taken a lot of the air out of the balloon. Michael Bitcoin was basically like a mini bubble. I mean, down 44% in 12 months, like, that's pretty catastrophic. What do you think happened? It's software. I mean, if you, if you overlay a chart of Bitcoin versus the ETF IGV, which are software, it looks very, very similar. I think that a lot of the narratives, a lot of the disappointment are, that's all valid. It turns out that Bitcoin is software. It is software, literally, and it acts like software. So at least in that case, investors understand exactly what it is. It's just, you know, part of the software ETF, but it's like leverage software because Bitcoin is down 50% and the software ETF is down 30% or something. It's like software, but on steroids. Yeah. So if you think Salesforce is going to come back by Bitcoin? All right. Ben, I think we only have time for one more question. I mean, we've talked about a lot of the biggies, SpaceX, AI, crypto, and all the various components of the AI ecosystem. What do you see as the state of the market? I just-- I think I wanna take a step back and every time you've had someone on your podcast to talk about this as a bubble or this is not a bubble. Like I find myself nodding ahead. And I think it's one of the things that I wanna say just being in this market environment shows how fun and exciting this stuff can be. 'Cause I know some people really don't like the markets and they think they're boring, but it's really interesting to think about this innovative technology and what it could be and what it could lead to. It could lead to catastrophe or it could lead to like this amazing thing in these huge returns and I just, it's as a market observer. I think living through times like this that you've only usually read out in history is really interesting to go through in real time and realize how it's so easy with the benefit of hindsight to know like of course that happened. Of course that was the top or that was the bottom when you're in it like this. It's nearly impossible to tell. You, everyone knows when you're in a financial crisis. No one knows for a fact that you're in a bubble when it's happening. That's the hard part and that's what makes it so exciting to follow the markets right now. Michael Batnik, Ben Carlson, thank you very much. - Thank you, Derek. (upbeat music)