The Riskiest Moment of the AI Bubble

Added: 2026-06-10

[00:00:00]Last week, Google announced that it wanted to raise almost $85 billion by selling stock. In this day and age, it's hard to understand how much money $85 billion is. But the largest IPO in history of the American stock exchanges was Alibaba. It was $22 billion. So this is like almost four times as much as that, which is not on its face a problem. I mean, Google is very valuable. Google does a lot of buying its own stock back to raise the price of its stock. And now it's doing the opposite of that, making it available for sale. People are going to be able to buy it and people will. And that's really interesting for a few different reasons. One is that this is Google. The Alphabet is one of the richest companies in the world. They make a lot of money, tens of billions of dollars in profit per year. Some of it getting thrown off of this very YouTube video right now.

[00:00:48]You know, we're doing it, everybody. Why would Google, one of the most profitable, largest companies of all time, need to sell stock to raise money?

[00:00:55]And the answer is, of course, AI. All of the AI stuff is expensive. The chips and the data centers and the power and the lobbyists. But that's the official simple answer. There's also a less official answer that is speculated by a lot of people, which makes this all very interesting and outlines the title of the video. SpaceX is about to go public.

[00:01:16]It's about to make shares of itself available for the public to buy in an initial public offering or IPO. It's going to sell $75 billion of shares is the idea. Anthropic also has filed to go public. It's going to sell another something like that. OpenAI is probably also going to have its IPO this year.

[00:01:32]They probably feel like they're coming in a little late. Depending on how you count the upcoming offerings and private rounds, investors may soon be asked to put like hundreds of billions of dollars of new money into AI companies. And these are all AI companies. We're talking about SpaceX, but SpaceX is now an AI company. It's merged with XAI. I don't know how any of this was allowed, but it was. Here we are. But the important thing here that like people who think about investments think about, but nobody else does, is that this is new money. And that's very different from how the stock market works dayto-day. On a normal day, when a stock goes up or down, the company is not raising money. The existing shares are just being traded between investors and one person sells and another one buys.

[00:02:16]And based on how much someone wants to sell and how much someone wants to buy, the price that they will agree on will become the new market price. And everybody's shares without anybody else making a trade will become worth that new market price. So you buy $100 of Google stock. Tomorrow somebody is like, "Actually, I' I'd pay 5% more for that."

[00:02:35]Now you have $105 of Google stock. You didn't do anything. Nobody bought anything from you, but now you have more money. And that goes up and up and down and down. And that's how the stock market works. If a billion people all owned Google stock for $100 a share and nobody wanted to sell any for less than 200 and somebody was willing to pay 200, then all those people would suddenly have twice as much net worth. Of course, actually lots of people are always willing to buy and sell at various prices. So, you know, one person buying doesn't actually inflate the price of a stock, but theoretically it could. If nobody wanted to sell, then that would make it go up, which is a little bit what happens in the private market with companies like Anthropic and OpenAI.

[00:03:14]People can't really sell. They can only buy. And perhaps this is useful to keep in your head, but if you take the price that the the highest bidder is willing to pay and you multiply that by every share that a company has, that will give you the company's market cap. That's how much it's worth. If Google has 1 billion shares and one share trades for $100, we say that the company is worth $100 billion. But a hundred billion dollars did not enter the company. Hundred billion dollars did not even necessarily enter the market. the whole company was repriced based on that one marginal trade. This is why stock market valuations can move super violently. A small number of shares trading at higher prices or lower prices that can revalue the entire company. The issuance of new stock is a totally different thing from this. If a company sells $75 billion of new stock, somebody has to buy all 75 bill not one person but people institutions and people have to buy $75 billion of new stock. All of that money has to be traded. It can't just be like one share affecting the price. All of that has to be new money entering the system. And so finally, like this is kind of the moment, the most dangerous moment so far in which AI has to prove that it is actually worth itself. For the last few years, the AI trade has mostly been about belief. Like investors, I think do believe that AI will be huge. And so existing AI linked stocks have gone up in the public market. You know, Nvidia, Microsoft, Google, all going up. Private companies at the same time have raised money at enormous valuations. So, OpenAI, Anthropic, SpaceX, Coreweee, XAI, they're all taking money and they're actually, this is new money coming into them. When they raise money from the private market, people are giving them money and these companies have become a giant story about the future. And even if you're not invested, I'm like, I'm invested, but only through index funds.

[00:05:04]Even if you're not invested, oh god, a lot of the economy seems to be based on this giant story about the future. And now that story is going to ask for a lot of real cash. So Google wants 85 billion. SpaceX wants 75 billion.

[00:05:19]Anthropic has already raised privately 65 billion is preparing for offering to the public market. That's going to be tens of billions. Open AAI also tens of billions. And these are not normal IPO numbers. Like the entire public offering market last year raised $44 billion. The biggest year in the US of all times of IPOs was about 142 billion. That was 2021 when we were going crazy with the spaxs. So when people start talking about 300 billion, 350 billion, 400 billion of AI related equity supply, like need for new dollars, that is a huge amount of money that needs to come from somewhere. And that's not a question of whether or not investors like AI. I really do think they like AI.

[00:06:06]The question is whether they have enough cash to buy all of this stock at these prices. And if they don't have enough cash, maybe they still want to buy, but in order to do that, they have to sell something. Maybe they sell Nvidia, maybe they sell Tesla, maybe they sell Microsoft or Apple or Google itself.

[00:06:22]Maybe they sell index funds or something really boring like they pull money out of money markets. But the money does have to come from somewhere and it's a lot of money. I don't know how much money is sitting around out there, but 400 billion is a lot. And this all brings me back to Google suddenly, to me anyway, appearing out of nowhere and making what is probably an extraordinarily savvy move. Google does not want to pop the AI bubble. Google benefits massively from the AI bubble.

[00:06:48]They want the AI bubble to continue to exist, but at the same time, they want safety from the AI bubble. And they may have noticed that investor capital is finite. If investors are willing to pay enormous prices to be exposed to AI, Google can say, "Hey, you get exposure through us." And also, you get like a safe, great, obviously well-run company that has a lot of different sources of income. We're already public. We're already profitable. We already have users. We already have distribution. We have data centers. We build data centers. We are even working on custom chips. If AI disappoints and everything crashes, we've still got YouTube, baby.

[00:07:26]So, it's not Google like trying to pop the bubble, but Google is trying to raise money while the bubble still exists. And also by offering a bunch of new Google stock, it's going to soop up a bunch of money that might have otherwise gone to other AI companies. If Google absorbs $85 billion of AI investor demand, that money can't go to SpaceX or Anthropic or OpenAI or anybody else. And maybe market demand is strong and confidence is high and every offering it will totally sell 100% and then more and all the stock prices will go up. But Google obviously could have waited for SpaceX to happen, waited for Enthropic to happen, but they didn't.

[00:08:03]They wanted to be first. A bubble can survive super high valuations for a very long time. As they say, the market can stay irrational for longer than you can stay liquid. And that can make it hard to make bubbles difficult to identify in real time. And like I always remember that in 1999 everybody said it was a bubble for like two years and then as soon as people started saying I mean maybe it's not a bubble that's when but like that's just vibes. Bubbles are difficult to identify you know like tulips and bitcoin are one thing but in terms of like the big bubbles there's usually real value like railroads the internet broadband infrastructure there's definitely overinvestment but the underlying technology really did change the world. It really was a big new business opportunity. The existence of a bubble does not mean technology is fake, but it does mean that investors are paying too much too soon, which happens like basically every time there's a big new technology like railroads or like the internet. I think this is why so many people are comfortable acknowledging that AI looks really bubbly while remaining bullish on AI. They can imagine a future where AI becomes enormously valuable and they can imagine that today's spending will eventually be justified. What will happen between now and then? Will it just be like a straight line that always goes up? It would be if capital is always available and prices keep rising.

[00:09:21]But it seems to me that bubbles don't end because everyone suddenly decides that there's actually no value or the story was wrong. They end when the cost gets so high that it's hard to hold on to the confidence that everyone had for so long. So really what pops a bubble is cost is how much it costs to keep inflating it and whether confidence continues to inflate along with cost. It seems like eventually the cost gets too high and the confidence isn't like saying no AI will never happen and ever be valuable. What the confidence is saying is like that's too much, you know, too rich for my blood. You're actually asking for real money now and I'd rather sort of wait and see for 6 months. When investors believe that there will always be another buyer, valuations can totally keep climbing. A stock can double and then double again because each new buyer is just assuming that someone else will eventually pay even more. The system is just feeding on the confidence. And so far, if you look back, you're like, man, I wish I'd been more exposed. I wish I'd had more cash in that. But when I say that this is like the most dangerous moment for the AI bubble, this is when the confidence has to be converted into cash. and Google getting in first, like working to get in first has the feeling of like needing to move fast. And that need to me seems a little like just in case the actual movement of actual capital puts a dent in people's confidence. And there has been need for capital before like OpenAI and Enthropic and SpaceX have all raised tens of billions of dollars, new money from private markets. So, the money has existed, but all of that money is oriented toward the eventual goal of being able to sell it. And not all of the investors are going to be able to sell immediately. Most of them won't.

[00:10:57]They'll have to hold on to it for a certain amount of time. But if you bought $100,000 of anthropic stock 3 years ago, you got like super lucky and you did that and now it's worth $6 million and that's like 3/4 of your net worth, you're going to want to sell a big chunk of that as fast as you can just to diversify. So that's other money that's going to have to come into the market. In the private market, all these investors are trying to get in before the IPO. At the IPO, everything has to get really real. And while the revenue story at Enthropic does look better than I would have guessed that it would have at this moment, this is all very new.

[00:11:29]And I think a few cold feet could really mess this up if like people start to shift from that confidence to caution because they like it. The need for capital got real. So yeah, Google stock sale here plus the IPOs of SpaceX, Anthropic, and Open AI, all that's going to have to turn a lot of enthusiasm into real actions in the real world. And I don't know if you've ever been in that situation where you're like, "Yeah, I'm gonna go buy that car." And then you get there and you're like, "I'm not going to buy that car." I was like, "I'm going to have to live with that car for 10 years.

[00:11:58]Take a little more time, figure out if you actually want to buy that car, get a Honda Civic." This is a really risky moment for the AI bubble and thus for the entire economy. And I'm making this video in the full awareness that I can provide you no amount of certainty. But I will say I remain happy that I use the investment strategy version of a Honda Civic by buying lowcost index funds and never looking at them. My job is to do my job and then if I have extra money, a little bit of life insurance cuz I got a family and lowcost index funds, that's for me. I'm a Honda Civic boy. I just wish they'd make an electric