Wall Street Analysts Are Comparing This To Enron — We Had To React

Key Takeaway

Treat concentrated bets—whether in AI stocks, a sector fund, or your own portfolio—as a risk-management problem, not a prediction contest. Today, list your largest holdings, identify what they are indirectly exposed to (AI spending, private credit, a few major customers), and set a pre-committed threshold for trimming and reallocating gains. The episode’s practical message is simple: assume your timing may be wrong, diversify before stress hits, and understand what you own rather than relying on optimistic headlines.

Episode Overview

Tom Bilyeu reacts to Ed Zitron’s concerns that AI infrastructure spending and revenue are highly concentrated in OpenAI and Anthropic, while financing risks may be dispersed through private credit, pension funds, and structured products. The discussion compares the current moment with prior technology booms and the 2008 financial crisis, while stressing that the AI technology may be real even if its financing and timing prove fragile. The central recommendation is caution, due diligence, and diversification rather than panic.

Full Transcript

Transcript of Wall Street Analysts Are Comparing This To Enron — We Had To React from Impact Theory. Auto-generated from episode audio; may contain minor errors.

AI now, if you have a 401k retirement plan, this is what you need to understand about debt obligations, ways to hide them, whether it's legal or illegal , we'll talk about that . Fasten your seat belts. We'll look at Ed Zitron. This guy, on my part, is a researcher, he kind of came out of nowhere, a real AI fan, and let's hear what he has to say about where this debt is hiding . In your opinion, where are we in terms of the AI ​​narrative, and what is the reality?

Well, I think investors should be asking questions now. What am I getting into when I invest in Microsoft, Google, and Amazon? So, UBS estimates that 27% of Google Cloud revenue this year will come from OpenAI and Anthropic, and next year that figure will rise to more than 48%. This is an impressive amount of money. Next year, it will be over $124 billion. Everyone is buying these stocks because they think all this capital spending is going to diversified and distributed demand for AI, when in reality it's helping to create the infrastructure for two unprofitable, unsustainable companies.

It's entirely possible that he's right about them, even if he's definitely right about them being unprofitable. He might even be right about, um, they're essentially doomed and these won't be the players who stay. We have seen this repeated throughout history. Um, but it gets more interesting when we get to what he just hinted at, which is that these companies have all this cyclical financing. Um, later he'll talk about how it's like Enron, and, um, there are other people who say it's like Enron. He talks very softly about it, but one of the things I want to talk about today as we go through this is that the things that are being done are not illegal, but they are extremely risky.

So keep that in mind. Not "no," but wow, is it risky? One of them would be Anthropic. Yes, as you claim. So give us more data, because you have a micro company that you're referring to, Microsoft, but what about AWS? Well, that's what I meant . Barclays actually says that 13% of AWS revenue this year will come from both OpenAI and Anthropic. And next year it will be 18%. AWS is a much bigger business than Google Cloud. Now, to be clear, when I said 27% this year and 48% next year for Google Cloud, I was referring to both Anthropic and OpenAI.

Good. Most people don't know that OpenAI is a big Google Cloud customer. It's not, well, it's not a well-known fact , but it was actually mentioned by Steven Ju of UBS. So where would these companies be now without Anthropic and without OpenAI? Well, I have some serious questions about this. So, in the 2025 calendar year, according to my own OpenAI metrics report, 69% of the annual growth in Microsoft's intelligent cloud technology segment was actually driven by OpenAI. Without this, it would have grown by only 8% annually , barely exceeding inflation.

And so they are promoting to everyone what I consider to be a kind of lie. This is, frankly, a kind of scandal. This is where the interesting begins. So calling it a scandal is, um, one of those things that we as a society will have to decide: is it a scandal or not? Because that's how companies work. If you think about how even a used car salesman moves units, he says, “Oh, you can’t afford a car? Don't worry. I will provide you with seller financing.” So you are effectively borrowing money from me to buy an asset.

And that's what Nvidia did like crazy. A lot of these companies do something like, "I'm going to invest in your company, but I know you'll use the investment I just made to buy my product." So whether it's your, um, investment in a data center that they're building, knowing that they're going to do something on your behalf , or whether the data center is actually going to be built on your cloud service or something like that. Such things are very typical. So this is where people look at the map of all the cyclical financing and think, “Wait a second.

Is this really legal or not ? And the question completely turns into whether it is disclosed? And the reality is that when you push these things, it all comes out. It may be on page 88 of their reporting, but they disclose it. I haven't seen anyone say these guys aren't revealing it. So the real question is not whether this is a scandal, the real question is who does this put at risk? And this is where this whole, um, thing starts to really escalate. So we'll look at this in more detail as we go along , but for now let's just focus on this.

There are three fundamentally risky ways in which debt is squeezed out of banks and into the broader system. So we all probably look at how these companies get financing and say, "Well, it's either cash on hand or they're raising this money from banks." Keep in mind that even Google has gone into negative cash flow for the first time. This is incredible. They have not been negative since going public. So none of these guys are doing this just for the money. The capital costs of artificial intelligence are so huge that they have to raise money from banks.

Okay, great. Banks take risks, right ? Well, not exactly. Banks are pushing this, ahem, risk out to the public. I'm not going to go into explaining exactly how yet. We'll get to that in a minute. But now, as a reminder to everyone, AI is now essentially the US stock market. The US stock market is a giant part of the global stock market. When money looks for a place where it can get a return, it disproportionately flows to the US, which means that that money disproportionately flows to AI.

So, essentially, the world is betting on AI. The world is betting on AI for growth, right ? So there's this basic idea that the real idea of constant growth ended somewhere around 2011 , 2012. And now this idea that tech companies are just going to grow forever, they've run out of ideas. It just became an advertising play on major social media platforms, and that's it . And so for a second it seemed that Web 3 would become something that had run out, gone nowhere . And so now, because they just need something that will give this growth story, because remember, real wages are not growing.

So, people looking for an answer to the question of how the hell to beat inflation all turn to the idea that there has to be some kind of asset growth story. Okay, great. So I'm going to invest all my money, and then this amazing story about AI comes along. I think the right way to put it is that AI is real. It will become the huge giant everyone thinks it is, but it will take much longer for the revenues to come in than the debt will last.

If that turns out to be true, and remember, it's been true for every major technological revolution, then I have no reason to believe that the one that's happening now, which requires much larger capital expenditures, will be any less true. And we saw the revenues coming in much slower than people expected, and that was before China started launching things like Kimik 3, where now we're seeing companies say, "Oh, well, it's a lot cheaper, so I'll switch to that." Okay, that's the gist of the things we're worried about , and the climax, and again, we'll look at the individual mechanisms in a minute, but the climax is that banks are not stupid.

And so they take on this debt, this high-risk debt , and they partially shift the risk, essentially, to your 401k. This is a simplification, but that's why people need to understand what's going on. We'll look at this mechanism in more detail in a minute. So it goes back to the fact that I think we have companies, circular finance, the circularity of all of this, and that creates demand for their products. So, when does it start, when? I asked a guest earlier, when do you think the party ends ?

And how, yes. So, with the OpenAI IPO, I think this could be one of the hot spots. Remember, this company was supposed to go public this year. They applied about a month or two ago. And now the New York Times has reported that they are considering a postponement until 2027. This is deadly dangerous for many people, but OpenAI and Anthropic need a steady flow of capital. They don't pay their bills from existing cash flow. So when something happens to this money, I think the first thing that can happen is something like a domino that falls.

But then again , there's also the general problem that data centers just are n't being built very quickly. This takes between 12 and 36 months, depending on how small or large the data center is actually being built. And the problem is that everyone thinks AI comes from cash flow. That AI only comes from this diversified revenue base, even though in reality it doesn't. It is extremely narrow. Information released a few months ago suggests that 89% of the largest AI companies, well, their revenue comes from OpenAI and Anthropic alone.

It is highly centralized. I'm going to take a little dive and you guys can educate me on whether or not we should do this again in the future. Um, so you have to understand that part of the reason why this accounting trick they use works is because of the way they do the calculations. So, there is something called EBITDA. This is earnings before interest, er, depreciation. Um, damn it, I need to, uh, pull this out in front. It seems we have it here . Uh, I always forget.

I-uh, all I need to see are the letters. Er, earnings before interest, taxes, depreciation and amortization. Thank you. Uh, so when you calculate EBITDA, you mean that I'm going to do my calculations somewhat detached from the real world. And the reason I'm going to do that is, you know, once I buy something, uh, I have to take out a lot of money up front, but then I can only bear the costs gradually, over time. And that just creates complications. And it would be great if, uh, your equipment lasted forever, but it does n't.

And so you will have to replace it. And so it becomes a critical part of how you do, uh , the actual state of your business will be determined by how often you actually have to replace this. And so one of the main accusations against AI is that they are dishonest about how often they will have to replace these chips. And so if you say you're going to replace them every five to six years, and that's how you manage your EBITDA schedule, it's kind of like, "Okay, that's going to be, uh, basically nonsense that you're much further underwater than you'd like people to believe." And the reality is, I think it was Warren Buffett who called EBITDA either Warren Buffett or Charlie Munger.

They called it reverse float, uh, meaning you have a huge upfront cost, but then you can only gradually get, uh, actual benefits over time, like tax deductions or expense deductions. And so, it becomes a way for people to hide a lot of things. And then there's another type of accounting that you can also use where you actually change, um, it's called adjusted income or adjusted EBITDA, adjusted profit, sorry. Um, adjusted profit or adjusted EBITDA, and that becomes something that's actually completely made up. So now every corporation can say, “Well, the way we do things is actually going to be different.

And so we're going to cut a lot of things. So, um, maybe we won't even consider the fact that we compensate our employees with stock options. Okay, this is an important question. So if you want to know why people get angry about corporations buying back their stock, part of the reason is that they buy back their stock so they can pay their employees in stock; but since these are corporate shares that they are then going to give away as stock, they don't actually count them as financial expenses.

And so people say: " What the hell are you talking about?" You just bought those shares back." “So now the thing is, they don’t even report it as an actual cash outlay. And in this way they make their business better and better. If you're a financial analyst, you know where all these things are hiding. And so you can do your due diligence. You can actually find out what's going on in these companies. You can see exactly where they are getting rid of debt , where there are expenses they are not talking about.

And that's what Michael Berry was trying to show everyone. He says, “ Their, um, amortization schedule is unfair .” And because he's dishonest, they're covering up, I think they've already suffered losses of about $72 or $76 billion, but they were able to cover it up by saying, "No, no, no, what do you mean by 3 years?" These last 5 or 6 years, chips." So Michael Berry says, "Hey, their numbers don't match what they say." This is important because remember that the entire AI game , the entire bet that the world is making right now, is based on whether the revenue comes in before the debt matures.

If they say, "Oh, [ __ ], we expect this to last 5 or 6 years, but we're actually going to have to refinance again to replace these chips in 3 years," then you just keep increasing that yield, increasing it, increasing and increasing , and increasing, and increasing, and eventually they'll succumb to the burden of their debt. So, Ed Zitron here points out, “Hey, these guys are essentially selling a big lie. I don't think this would stand the legal test of lying. But are they trying to engage in propaganda to give free rein to the human imagination, to give you a simple number at the beginning, a headline, which is really just an adjusted profit, which I call fake?

And I think it was EBIT EBITDA, which is better than adjusted profit, Charlie Munger called nonsense. Literally just outright, he says it's just nonsense. So we'll get back to the show in a second, but first I want to talk about the customers you're losing without even realizing it, and why today's episode brings you Quo, a business phone system built to ensure you never miss an opportunity. Most companies track incoming leads , which is very important, but very few track those that slip through the cracks after a lost follow-up or a hidden voicemail.

Qual fixes this. All your calls, texts, and Voicemails are now stored in one place, so your team has a complete history and nothing slips through the cracks. Qual is the number one business telephony system on G2, trusted by over 90,000 companies. Qual's built-in AI agent handles after-hours calls, answers questions, and makes appointments, so you never miss a lead. When money is at stake , always say hello to Qual. Try Qual for free and get 20% off your first 6 months at qual.com/impact. This is q u o dot com slash impact.

We'll get back to the show in a second, but for now let's talk about what keeps me up at night. People are impersonating me. My team is constantly being sent fake emails and fake text messages pretending to be me. Things like, “This is urgent. Click this link. I need gift certificates. Send them right now." “And in that split second of panic, someone almost always loses their guard. The scam works precisely because personal information can be bought and sold, and that's where Incogni comes in. You authorize them once, they track your personal data across hundreds of brokerage sites and delete it .

When they reappear, which happens, they remove them again. Additionally, with individual deletion, you can send them any link that displays your information and their team will take care of the rest. They can't chase you with data they can't find. So go to incogni.com/impact and use the code impact to get 60% off the annual plan. Try it risk-free for 30 days. This is i n c o g n i.com/impact and use the code impact. Now let's get back to business. It becomes very important for people to know what the amortization schedule actually is, when the material will be profitable, when they will probably have to raise money, when that revenue growth trajectory will actually intersect with the opportunity…” it’s a self-sustaining economic engine.

And without this, you find yourself in a very dangerous situation . Okay, back to Ed. Is it expensive to do or not? Regarding the construction of data centers, etc. And what will make, um, generative AI, the ability for it to be really, really good, is to have access to a lot of information. So, shouldn't it be somewhat, Ed, concentrated? Well, when I say concentration, I mean the concentration of revenue in these two companies. No, I understand, but to make it good. So, doesn't it make sense that those who are most exposed would be concentrated to some extent?

Well, I mean, when we talk about ...So, Sideline Climate said back in February they saw about 190 gigawatts of data center capacity being built over the next few years. It has either been built or is being planned. Now, if you calculate this with PUE, which is simply an efficiency rating of 1.3, you get... $12 million per megawatt over the $1.6 trillion in annual revenue needed to power these data centers. Having two clients won't do it. Even if they are very spent. And they could probably invent artificial intelligence, but they can't afford it.

They need venture capital, but they're only going to spend 400 billion a year, and that's if they go that far, which I don't believe. And then you have 1.6 trillion. Dude, these numbers are so wild. They've been pushing income for God knows how many years to say, " Hey, our stock is worth this." "By the way, which we didn't even talk about." I don't remember if he touched on this, but, uh, Space X AI is still, please confirm, but I'm sure it's still trading lower than the opening when it did its IPO.

It is back on a green candle, but still lower than it was at the time of the IPO. So, phew. Yes, a red candle. This is a red candle. Today it fell by 9%. Oh, God. Okay, so now, Jesus. Uh, imagine the situation getting even worse while we're sitting here watching this. So, imagine you have this big bet that everyone is making. Uh, you have these income requirements that are way above the income you actually make, and you have this potential question of whether your, um, chips will last as long as you say, which could mean you're even further underwater than people think.

And then, on top of that , you have all these risk mechanisms that push it down. So what I want to do now is to roughly know what minute we are at now. Good. So, I want to look at three risky ways in which debt is actually eliminated. So it's critical to understand, as he said, these companies can't make cash from what they do. So they have to find a way to raise money. And considering that even Google now has a negative cash flow for the first time since going public, everyone has to rely on banks.

But banks can't really touch it because it's so speculative. Artificial intelligence is a huge question mark. Artificial intelligence is incredible . No one believes in anything more than I do. But if you're a regulated bank and you start taking huge positions in these very speculative companies, the regulators will say, " Brother, come on. "Don't be ridiculous." Good. So, the banks are saying, "We have to find ways to get rid of some of this risk ." Here are three key ways that this risk gets pushed to a position where you have to worry about it , where it becomes a systemic risk for, ahem, the average investor.

So, one of them is called wholesale financing. These are loans to creditors. In many cases, banks still lend money to these banks, to companies, but they do so through an intermediary, a shadow bank. What is a shadow bank? It is essentially a private bank that is not subject to the same rules as those companies that make loans. So it could be a venture capitalist going to a shadow bank, a private equity fund, whatever that goes to a shadow bank or receives that money. This could be a venture capitalist or the private equity firm itself raising capital.

They can look to pension funds as where these guys are getting a bunch of this money from. But they need to get that money somewhere. So these guys say, "We're going to take money out of the bank." The bank now has a plausible deniability, because they did n't lend to the AI ​​guys, they lent to us. And then we go and lend to artificial intelligence, okay, companies. So, this is the first way. The second is what is called credit risk transfer. Sometimes banks make loans to corporate borrowers, but they use financial engineering to remove risk from their books, and they do this through a mechanism known as credit risk transfer.

By doing this, you are essentially trying to remove the risk, you are packaging things. Okay, you say, "Okay, we just took on this debt, we're going to create a financial instrument, and then we're going to sell this debt to people." So they're going to buy this debt, and if it reminds you of 2008 with mortgage-backed securities, that's absolutely true, but for AI... So you're trying to remove credit risk from your books so that again, if regulators look at it, you're not left with the bag. So this is sold to places like pension funds.

Again, why do people need to worry about this. "Uh , so that when it becomes known, you can make a loan, but you don't have to hold it." in your books. And the third is from origin to distribution. You can think of it as a conveyor belt. So, uh, in this method, the bank acts solely as an intermediary. They're collecting a bunch of corporate, ahem, like software loans , artificial intelligence, etc. They combine them into a complex financial, uh, very specific, sorry, complex financial product. It's called a CLO, a collateralized loan obligation, and then they immediately sell pieces of it to people and institutions who are really trying to generate income.

These will be life insurance companies, pension funds, but this creates a kind of trap. Thus, the bank can appropriate large commissions for arranging the deal, but then they can walk away with virtually zero risk on their balance sheets. Now the reason other guys take this is because they won't be able to find different loans and credit products and so on, so they want someone to package them, but they also just want to take a piece of the deal. So they say, " Okay, this is getting super boring," and it turns into something called mezzanine debt, where you look at it like I'm lending you money, I have three ways to look at it.

The bottom layer is equity. I'm just taking a share in the equity. I make the most profit, but if I get destroyed , I'll be the first to go bankrupt. So, if your company has problems, we will burn everything I claim first, it will just disappear. Then we move to the middle layer, this mezzanine, and that's where I kind of defend myself. I am protected from people who will be the first to be eaten by equity, but I will be eaten before the senior debt holder.

So, I figure that, um, I get a little protection, but I get a smaller profit, but I get wiped out even more sooner than the senior debt holder. So if I can't be the oldest, I want to be in the middle. Steeply. So they're going to package it up, sell it to people who think, " I need something that's medium risk." It will be something like: " This is definitely not going to be a triple A." It could be something like a BBB rating.” So I'll be somewhere in the middle, but I have potentially more growth potential because I have a class of people who will be eaten before I am eaten.

And then you have the top of these three layers, which is like a "triple A," you just owe me that damn money. And they will all be destroyed before I lose even one dollar of the principal . Okay, again, I'm going to find out here that people like these boring numbers are doing well. Yeah, yeah, but that's what they do. So they take that middle piece and sell it. This frees up a bunch of money on the bank's imaginary balance sheet . Remember, they're still at risk because they only sold a portion of it, but they look much better on their balance sheet.

So the regulator looks at that and says, “Oh, you just released $26 million, whatever you want, increasing that number. Uh, you just released that money. Great job, buddy. " You look good." But if things go awry, they will still be at risk from this debt. They look good on paper, but their risk still exists. So when people go through this, everyone asks questions like, “Is this legal? "Is this a scandal?" Instead, we should just talk about how much risk is being pushed out into the markets, how it's being pushed out, and the scariest thing is that it's so complicated that the average person just doesn't want to look.

They don't want to try to figure out what's hidden where, what's in my portfolio, so at the very least, never lose sight of the huge capital involved. Risk is diversified across as broad a market as possible. On paper, everything will look relatively clean, like in 2008, but in reality, all this risk seems to be floating around. Will it ever end like 2008? I don't know. Nobody knows. Since this is a private loan, it is very difficult, if not impossible, to ascertain its scale. But if you guys look at the in-depth research I did on the private lending market, the private lending markets are showing signs of decline.

Let's... Let's tie some of these things together. The whole world is betting on AI. AI is not capable of financing itself. AI is expected to attract so much capital that even companies like Google are now cash flow negative. When they take on all this debt, they can lie to themselves or just outright lie about the depreciation schedule of the asset, the chips, and that makes a huge difference in terms of how much capital expenditure they're going to need, which is directly related to how much debt they're going to have to raise.

And then this debt is largely provided by shadow banking, known as private banking, which is not regulated in the same way that regular banks are regulated, but regular banks use this as a way to do it, but hide it, huh? It's easy to hide because it's still there. It is visible. Everyone knows what's going on, but the math makes them amenable to their regulation. So, the banks that we think are protected or were protected from that systemic risk that we were exposed to in 2008, we think we have protection from it because those banks are regulated.

But in reality, these banks lend money to a largely unregulated shadow banking sector . And then the risk they carry is further into the regular markets through pension funds, life insurance, and so on. OK? So all of this looms as we question whether we can actually generate revenues fast enough to handle this, as China says, "Hey, these revenues never come in because we're going to do it cheap." As the yen fluctuates and the US Treasury Department orders the Fed to allocate more dollars to try to stabilize it, so we can stabilize our own bond market.

And it's all starting to get really unnerving. I did n't even address Iran . I didn't even reach out to China. I did n't even talk about the dollars that may never come from the Middle East, which we certainly count on. A lot of instability. When I hear people talk about this, they do n't just put it back to the risk profile. What is happening? Regarding what? So when you hear things like that, it's really important that you take it back. What is the risk profile?

How many of these things are distributed? What is my likely risk? How do I diversify so it does n't end up like 2008, when private credit collapses, and then when those dollars are sucked out of the economy, what will my life look like? If, um, AI has the same gap as the internet, and it turns out to be exactly what we imagine it to be, but with a 10- year delay, what does that look like to me? How can I make sure I survive this? Good?

This is something we'll hear more about in a bit. But, as we consider this, we want to remember about balances. Really? Well, I've spoken a lot from my own experience about open AI, because I reported on their audited financial statements for the Financial Times. Right. And this is a company that is simply burning money. They lost $20.9 billion in 2025 , and the situation is only getting worse. And what's even more amazing is that over $800 million in open AI revenue came from SoftBank for their crystalline intelligence, and yes, that's what it's called.

Their Crystalline Intelligence program, for which I can find no evidence that anything actually happens. And SoftBank, a major shareholder in open AI, has no board seats. So, Carol, it will be done. One more question. As you were talking about Google Cloud, in terms of impact, right? And you said 48% next year, if we talk about these two customers . I must say that these companies are run by smart people. And usually you would say that your influence is only on a few clients–that’s not very good.

Are you saying that these companies that don't do due diligence, whether it's Alphabet or, you know, choosing their hyperscaler ? I think they did their due diligence in the sense that they said, “We’re going to build our biggest customers and we’re going to own a significant portion of them. And on top of that, we will own all of their infrastructure.” Google has a good...They have a good business here. They buy TPU from...Oh, sorry. Broadcom sells Google's TPUs. They are then sold to Anthropic and then leased back to Anthropic via Google.

Google doubles its revenue. This sounds really good until you realize that Anthropic and OpenAI are unstable. So, who could they be? The problem is that when I say these are smart people, it immediately makes me think of Enron. The smartest guys in the room. I'm not saying that anything like that is happening. But I'm just saying you have a fiduciary responsibility. And you are right. Go back to Enron or Qualcomm. I understand. And I think the point I'm making is that with Google, they probably thought there would be more customers.

I guess with Azure and AWS they thought it would be more of a big player. But the problem with Anthropic and OpenAI is that they raised $200-300 billion in funding, but they actually raised more because OpenAI and Anthropic got all their infrastructure built for them by Microsoft, Google, and Amazon. It seems they didn't have to pay, as in the trial of Sam Altman and Elon Musk, one of the Microsoft executives said that the infrastructure cost a hundred billion dollars, that is, approximately 70-80 billion dollars. Yes.

The problem is, no one else can become as big as them. No one else can get that many calculations. No one else could afford these calculations and be able to conduct the necessary preliminary training runs. Unless now China is catching up with them. Right. And it's unclear how anyone is actually solving any of the problems I've been listing for years, namely the instability, the unprofitability, and the lack of real ROI on AI. An important part that I don't think Ed is exactly talking about. This is not discussed in this conversation, but you have to understand the psychology of the people who are investing in this, the psychology of the companies, who, when he says, I think, thought there would be a lot more big players.

I think they have a different mentality when they look at what's happening here. So the original narrative was one thing, and now it's shifting to something else. I think the initial narrative, even before people understood the scale of the capital expenditure, was that they would have to keep scaling these brains, these data centers, to make these things smarter. Before people realized that there wouldn't just be 300 OpenAIs, the narrative was different. Now the narrative shifts to the fact that we are essentially building an electrical grid, but instead of electricity, we are getting intelligence.

So yes, we'll have to invest a lot of money up front, but it's just like laying a pipe on the internet. By doing this , we will create our own customers. So not many people are using it right now because people haven't understood what artificial intelligence will be like, just like they haven't understood what the internet will be like. But keep in mind, these people look at this and say, "Oh, I know how that ended." So last time, what happened was that everyone who made more than their money, there was this delay, and so people still won , making just incredible amounts of money.

Remember, all the mega-rich people today, except for Elon Musk, who really made his money in cars, but he was able to do it because of his success on the internet. In any case, the vast majority of hyper-rich people today got rich during the second wave of the internet. So after the collapse, people say, " Thank God this pipe exists, and now we're going to build on top of it ." So all these guys seem to forget about the pain of this gap, about the fact that there was such a big gap.

They understand that the history of hypergrowth is no longer there. If they want to make their money back, they need to find the next big thing. Artificial intelligence is turning out to be the next big thing. They deceive themselves into thinking that there will probably be a lull in the middle; that we're almost certainly going to wipe out a bunch of investors who are bankrupt with debt before we get to the generation of heirs who come along and essentially become the Jeff Bezos of the world and build, you know, the next mega-things on top of that.

So, I think they are wrong to still take on all this crazy amount of debt. I do n't think they're wrong about the fact that we need to be here, we need to be in AI, we need to build on this. I think they just need to be more mindful and remember that if you want to be one of the people delivering the intelligence product, then yes, you have to step up now, you have to be one of the people in the data centers. Okay, okay, fair.

But you better figure out who your customers are going to be, because if you can't wait 5-10 years now to get all the revenue you need to pay for this data center, and know that you're going to have to keep buying new chips and all that, you're going to be in trouble. That's what worries me. If you're thinking, "Okay, great. I'll wait. I'll let the first generation of people figure out if there's a problem or not." Because it's really about the companies that ultimately use intelligence, not the underlying intelligence itself.

And yes, I understand the idea very well that you want to be pickpockets. But remember, you want to be the successor generation of the pickaxes when it comes to incredibly expensive infrastructure. So, a typical thing, imagine you had to build gold mines so people could mine them. You can't be a pickaxe until there are gold mines. These are data centers. So if you look at this as commercialized intelligence, the application of electricity, and now we're waiting for all the things that use electricity to be built, and you want to be part of, um, you know, any shared chips or something like that, so you can be agnostic about who wins.

These are picks. Or do you want to be part of companies that will be built on the backend just like Facebook was built on the backend of the internet, Amazon was built on the backend of the internet, etc. And this is where it becomes important to understand where we are on the timeline? Good? The future is not guaranteed. So no one knows for sure whether this gap I'm talking about will actually happen. It's happened every time, but hey, there's no guarantee it'll happen now . But where are we on the timeline?

How confident am I that this will happen? And then, wouldn't it be better for me to try harder and harder, to get there now, to be on the ground floor of what I think are the picks? And am I right? Because, I think, if you think that a data center is a pickaxe, you will have problems. But am I right? Which generation of investors do I belong to? Am I the one who will be destroyed if I am right about these terms? Am I the one who will be part of the generation that will pass on the legacy?

And then obviously we have what we talked about earlier, which is systemic risk, and how can I protect myself from it if this downturn really happens and all the companies that are big right now , all the companies that make data centers worth building, go bankrupt. And then we have some downtime before we can bring in this generation that will take over the legacy to actually build it all. Good. Very, very quickly. Yes. It seems like we're getting into "too big to fail" territory, where even though we should let them be railroads, non-profit companies that fail, it seems like we're too...

They're so entrenched in all the other MAG 7 companies that now... It's like if OpenAI and Anthropic fail, it's going to affect Google and Microsoft , it's going to affect the internet, and I'm starting to understand that they're getting into bed with all these other people, so if we fail, everyone fails. This is even scarier. Remember, my mental model of Iran is that Iran was too important for AI to fail. I have to make sure that this $2 trillion in investment comes to me. So, part of what I'm doing here is that these guys promised investment , and now I have to take care of Iran, to make sure they do n't become a problem.

Obviously a mistake, but nevertheless, I think it was part of his calculation. Banks are truly too big to fail. Now, if AI crashes, your economy will explode so much that I can't even imagine what kind of money printing you'd have to do, what that would look like, it would be barbaric, man. This will simply be a nightmare of incredible proportions. Uh, especially because you're doing this at a time when the yen trade has already created instability and problems there, and could create problems for you in your debt market.

And so if you suddenly have to issue the world's largest amount of debt, uh, to stabilize the economy, then AI is going to crash, my friend... As your largest buyer of debt, I'm trying to sell it right now. That would be wild. This would n't be the best time. Well, this is where the question arises, that now is not the time to panic. Everyone should be very thoughtful about what the risk is? Like, I'm not going to do anything in the stock market or anything like that .

Um, I think you need to be a lot more careful. I think it takes a lot more humility to say, "Okay, there are a lot of risks here." That doesn't mean it will happen. You just need to be aware of the risks . So I think the banks will be bailed out if they get into trouble. I think that's why AI companies are desperately trying to get the government to actually take a stake in them, so the government will say, "Well, we can't let this fail." They're trying to re-present this as a national security issue so that if they don't have the money, they'll say, "Brother, do you really want China to give us away?" We have to be absolutely cutting edge ." So, the fact that we don't have the dollars and cents to make this business profitable doesn't matter.

Why does this matter? We just need to make sure the government funds it because it's a weapons system. So, this artificial intelligence has become such a giant, and the scariest thing is that it's actually real. As we just saw with the latest version of Claude, something like jailbreaking or OpenAI, jailbreaking OpenAI and exiting, and then thinking, "Oh [ __ ], like, we need to suppress this." And before that, there was the Fable explosion, and then thoughts like, "Oh [ __ ], this is a hack search, etc.

We can't release this." So, this is a weapons system, and it's a national security issue. The risk is now extremely systemic. And the current generation of companies are not generating revenue fast enough to be called anything other than, as Ed said, unsustainable at this point. Income may start coming in and maybe that picture will change, but the question is whether it will change quickly enough to deal with this debt burden at a time when interest rates are rising here and in Japan, where so much of this capital is actually coming from.

Oh my God, man. Yes, the more I research these things, the more the hair on the back of my neck stands on end. Is this like a time bomb, is there a time limit here? Or they can drop the box, offer a circular financing, an IPO, double the investment, get some private equity, and keep dropping the box until someone creates a real trillion- dollar LLM company. I don't know. Yes. Nobody knows. Mhm. Um, I think we're under a lot of pressure because of the stability of the US economy, but any prediction of real numbers is not realistic.

So, uh, it's more like , uh, if you were to watch a guy doing push-ups, Mhm, and say, "How many push-ups do I think this guy can realistically do? " Like, I can believe he can do 100. Can I believe he can do 500 non- stop, by the way? Can I believe he can do 500? Yes. Do I think he can really do 1000? This is starting to be hard to believe. Do I think he can do 5,000? Absolutely not. The record for the number of push-ups is 10,000.

Or more. Maybe there will be more. Look at Ryan, could you look for him? What exactly? So, what is the world record for the most continuous push-ups? It seems like it's 10,000 or more. And this number is so fake that I would never believe it. 10,507 from Charles Servicio in 1980. Without stopping? The 80s? Yes, 1980. It was all cocaine. These were, these were, these were cocaine push-ups. No way. So perfect. I like it. So now I don't know what the price of cocaine will be in the markets to pull this off.

So I'll look at all of this and think: bro, there's no way. We can't go beyond (insert, insert) an absurd number, but if mine is (for example), there's no way it's going to be more than 1,000. That's ridiculous. And in fact, that's 10 times more cocaine than the body can absorb and still function. This is, unfortunately for all of us, one of those situations where you have to invest with the right hypothesis, but at the wrong time, considering that this is the same as making a mistake.

And it's almost impossible to take it into account correctly in time , which means we're all almost guaranteed to make mistakes. That's why I build my strategy by trying to say, "I know I'm going to be wrong." So how do I diversify enough that, given that I won't know what I'm going to be wrong about, but not everything, if I'm diversified well enough, not everything could mechanically be wrong at the same time , because everything is moving in different directions, then the question becomes, " Okay, can I do this?" And therein lies the magic of it all .

Can you actually put yourself in a situation where you're going to make a mistake, but you're willing, literally mechanically willing to make a mistake, and come out of it with a good result? Because being up 3% while everyone else is down 30% is like laughing all the way to the bank. And I remind you that Ray Dalio, for example, during the 2008 crash, still returned 9%. They rose 9% as the rest of the world collapsed. So there is a way to diversify smart enough that you don't get caught with your pants down.

So that's pretty much how I think about it. So, I had a big win on one of my stocks, it's just ridiculous. When I told my wife how funny it was, hmm , it got some praise. I will be very honest with you. And I said, "Sell it immediately." Because I don't need to be greedy. As if from here it could double again, and I'd say, "Yes, I'm fine ." Uh, I'm perfectly happy to receive my winnings now, but most people do n't. Most people say, “Well, that’s fine .

I just need to keep moving. Let's continue. Continue. Continue". You should have a number in your head that you would decide: "If I reach this, I'll diversify into something else." Otherwise you will be caught. You will be caught. The last time you were with us, we got a really incredible reaction, to be honest . And a lot of people who weren't typical viewers or listeners of our show saw what you were doing and listened to what you were doing. And it really seemed like what you were saying resonated with a lot of people.

It's like there 's some kind of... like there's some kind of anti-AI fervor somewhere. And I'm just curious why you think that. So I'm not sure it's anti-AI. Don't get me wrong, but I think these are also anti-financial scams. I think everyone sees cyclical financing. I think they see that Microsoft , Google, and Amazon get almost all of their AI revenue either through the products they push on their customers or even through the compute spent on Anthropic and OpenAI. The existence of the average person is now so expensive, so difficult, so complicated.

Getting a mortgage is so difficult for the average person . But if you theoretically build a data center in 36 months for Nvidia GPUs, banks will randomly give you money. CoreWeave just raised what, 9% bonds? I mean, you can bring in anything if you have a data center. And I think ordinary people can see that AI is not delivering on its promises. They can see the wealth of people at the top of the AI ​​industry. They can also see that they are being lied to and deliberately intimidated on top of all this blatant cyclical funding.

So, here's what worries me. We create all this systemic risk, and if this risk is really real. And if something goes wrong, you want to talk about pitchforks, people will go crazy. There is already widespread anti-AI sentiment. Uh, to add to that, you guys just destroyed the economy with your, um, high-risk, um, AI ventures because you wanted to grow at all costs, and you guys were already rich, and it was already hard for the average person to make ends meet , and now you guys wanted to get even richer.

What did you want? An even bigger bunker? Uh, people will lose their ever-loving minds. And so, this is one of those times where, boy, do I hope that common sense starts to prevail, that at least the average person starts to protect themselves from some of the systemic risks that are being created out there, so that if they hit, they don't hit you. I have long believed that you may not be able to save society, but you can always save yourself. And so think about where this is all going to lead, because boy, if this goes down, we're really going to have a problem.

And that's why it's very interesting. Ed, I think, becomes, um, as big as he is precisely because he's a bear. He's talking about...listen, guys. This isn't going as well as you think. This is something people need to pay attention to, and he comes with evidence that it was actually all painted. So, definitely a voice that we will be voicing even more in the future. But, as if you're trying to understand why I covered what I covered, it all has to do with why we have a cultural orientation in the US, and more broadly in the West, that is becoming more and more distorted every day.

And so understanding how all these pieces are connected, how a lot of what we're going through right now is really an echo of 2008, when we never got back on our feet. This is an echo of what has happened in terms of the average person's life becoming more expensive during COVID. This is an echo of the wild way people become optimistic and then deliberately blind themselves with debt. The thing is, the world is much more interconnected than we like to think: from the investments of Iran and the Gulf states in the US market to Japan and the way they have provided global liquidity, to the general malaise of a country prone to radicalization because you've spent so much time financializing the world that there aren't those good middle-class jobs that people used to count on, which created a thriving middle class.

Also all of this is happening at the same time when you have, um, Islam as a culture that has a lot more energy, a lot more drive to go out and engage people, and all of this is happening at the same time . This is a crazy time, with the world order culturally, economically, militarily changing right now in real time. And I try to make sure that I pay attention to all these different threads that reinforce each other and share energy in the same cyclical way that these companies invest.

And until next time, my friends, be legendary. Take care of yourself. Peace. If you enjoyed this conversation, check out this episode to learn more. Everyone is comparing today's AI stock boom to the dot-com bubble of '99 , but the sharpest minds on Wall Street have a different vision. One such person is Warren Buffett. Warren Buffett , uh, I know he's not ...

Go deeper

• All Impact Theory episode guides