It’s Already Happening: The AI Bubble No One’s Ready For

In just 18 months, over 3.3 trillion dollars has flooded into AI-linked companies. That's more than the entire market cap of Germany. Today, five AI-adjacent companies alone represent over 30% of the entire S&P 500. It is the highest concentration of market power in modern financial history. Since 2023, Nvidia alone has added more market cap than the combined GDP of South Korea, Sweden, and Switzerland. And yet, on November 20th of 2025, Nvidia posted revenue, beating expectation by billions. And yet, the stock still fell, dropping the entire S&P 500 with it. Last quarter, AI-linked stocks wiped out 400 billion dollars in gains in a single trading session, the fastest reversal since the dot-com crash. And while AI investment has exploded by 800%, US productivity has barely budged, up just about 1.3% total in 2 years. The truth is, AI has given us the biggest technological hype wave since the dot-com boom, and the narrative is that it's going to last forever. But, is it? Even when Nvidia reports flawless numbers, the kind that most companies could only dream of, the stock still falls. At times, wiping out hundreds of billions of dollars in value in just a matter of hours. Now, this video is not about predicting a crash, but it is about something incredibly important. How to make sense of a world where 10 stocks that are totally detached from fundamentals account for all the value in the market and have a tendency to whipsaw up and down. Listen, the AI boom is real, but so is the AI bubble forming around it. We're going to cover exactly what's driving the market right now, because it sure as hell isn't fundamentals, and how to navigate a market with such extreme and costly mood swings. I've got to go forward plan, as always, for you in part four, but it's not going to make sense unless you understand the detailed warning in part three. So, welcome to part one, the reflexive loop. How AI hype creates its own gravity. Between 1998 and 2000, the Nasdaq jumped 278%. Not because of earnings, but because rising prices convinced investors that rising prices were the new fundamentals. The average dot-com stock tripled in the 6 months before the crash, even though actual revenue growth across the sector was less than 10%. In 1999, 80% of all IPOs were issued by companies with zero profits, and the system was poised for a fall. By March of 2000, despite being the most valuable company in the world, Cisco dropped by 86% when belief in a forever-up market finally collapsed. If you want to understand what's happening in the market right now, the wild volatility and sense that the surface-level story just does not match the underlying reality of what is really moving the markets, you have to understand one of the most important ideas ever introduced in finance, George Soros's theory of reflexivity. It's not a technical theory, but it is highly predictive. Once you see it, you're not going to be able to unsee it. Reflexivity says that markets are not passive observers of reality. They do not simply measure what's happening out in the world and then adjust accordingly. >> [music] >> Markets, instead, shape the very reality they appear to be responding to. Most people imagine the stock market as a kind of barometer, [music] something that reacts to pressure systems, earnings, GDP, interest rates, consumer spending, et cetera. But, in reality, that's not how it actually works in real life. Markets behave much more like a thermostat. They don't merely respond to the environment, they influence it. >> [music] >> And this creates a feedback loop between belief and behavior. And it's inside that loop that bubbles form. Liquidity rises, money sloshes around in the system chasing returns, prices rise in response, people get really excited, they convince themselves and others that this time it's different, it's never different, and then prices completely detach from business fundamentals and start responding more to supply, demand, and belief. And as such, bubbles inflate. Here's how this plays out in AI and what we're living through right now. Given the vast deficit spending and resulting need to print money, combined with relatively low interest rates, and easy money has flooded the system. That has driven prices up across essentially all asset classes. As prices rise, so do expectations. Someone posts an LLM breakthrough online, a new investing meta explodes online in response, and gets endlessly covered on social media. A CEO hits a podcast and says the world is about to be rewritten and overnight, the belief that AI is inevitable translates into more money flooding into a small number of stocks. That belief pulls in fresh capital because nobody wants to miss the next epic-defining technology. That new capital drives valuations even higher, and those higher valuations act as proof that AI bullishness is entirely justified. If the prices are screaming upward, surely the underlying technology must warrant it. So, people invest even more aggressively, further pushing prices even higher, which further reinforces the narrative, and on and on it goes in a self-reinforcing loop. And there's another element. There's actual money to be made. Even if the stocks are totally detached from business fundamentals, you don't make money in the stock market because of fundamentals. You make money in stock markets because you bet correctly on the direction and timeline of a stock or segments of the market, whether up or down. It is a truly self-reinforcing cycle of belief bringing capital, which increases valuations, which leads to a stronger belief in the thesis, which then brings in even more capital. This is reflexivity in motion. And AI is the most recent asset class to benefit from the volatility that it creates. Remember, volatility is desired by active traders. They can make money on moves up and moves down. The only thing they can't make money on is stability. Once you is the right framework through which to understand the markets. In all honesty, right now, AI isn't really a product, it's a story about the future that people can gamble on. It is a promise that everything is about to change, and AI companies will be the beneficiaries of this moment. It's a technological vision so sweeping, so totalizing, so intoxicating that it bypasses normal financial skepticism in the same way that the internet did back in the lead-up to the 2000 dot-com bubble bursting. And when you have a technology that carries that kind of mythic weight, the belief alone is enough to move trillions of dollars up and >> [music] >> down, sometimes on the same day. Nowhere is that more obvious than with Nvidia. Nvidia is no longer being priced like a traditional company. It's stepped into the reflexive center of the AI narrative, the same way Cisco did during the dot-com boom. Investors aren't valuing Nvidia based on last quarter's earnings or next quarter's guidance. >> [music] >> They're valuing it based on the future the public is willing to bet will eventually exist. In many ways, Nvidia has become the physical embodiment of AI's inevitability and the wisdom of investing in picks and shovels rather than gold itself. This means its stock price is no longer a measure of what the company is today or even necessarily what it's going to be tomorrow. And instead, it's become a measure of what investors imagine the next decade of human progress will look like. And that's why even perfection isn't enough anymore to keep prices moving up. Investing is a player versus player versus environment game. It is combative and winner take all when people are betting on specific stocks and regional moves. So, Nvidia can deliver a quarter with 57 billion dollars in revenue. It can even beat expectations by billions, as it just did. It can even post numbers that no mega-cap in history has come remotely close to matching, and the stock can still fall, violently, instantly, [music] and in falling, it can drag the entire market down with it. It's not a commentary on Nvidia, it's a commentary on the nature of a PvPvE game with reflexivity at its core. When an asset becomes the gravitational center of a belief system and a way to profit off of people's much-discussed belief in the future, it will inevitably stop responding to business fundamentals and start responding to narrative tension, because it's pulling so much of the perceived value of the future into the present. There's only so much future you can believe in. Whenever the story wobbles, even slightly, the price is going to react dramatically. It doesn't matter what the numbers say. It matters how people feel about what the numbers imply. [music] Every piece of news, no matter how objectively good, can still potentially be interpreted through the lens of expectation oversaturation. This is exactly what we're seeing right now and exactly what late-stage reflexive cycles look like. Good news that doesn't lift prices at all, bad news that punishes disproportionately, and or a whipsaw of up and then right back down with a creeping sense that belief has climbed higher than reality can justify. It is critical to understand this feedback loop because reflexive bubbles don't collapse when the fundamentals break, obviously. They collapse when people's belief in the only up phenomenon collapses. Reading the book 1929 about the stock market crash that led to the Great Depression has been extremely eye-opening. It just really drives home the point that the crash itself is not the key to understand the Great Depression. You have to understand what led to the crash, the insane borrowing, the euphoria, the absolute detachment from business fundamentals, the market manipulation on behalf of the hyper-sophisticated traders that prey on others' ignorance. It's wild, and there are just too many similarities to every massive market run-up in modern history to ignore. We all want to believe that the stock market is about business fundamentals, but honestly, fundamentals are for the slow times. Once a market heats up, it's a totally different animal. It's reflexivity. Once you understand what reflexivity is, you see the danger. You see that there's a gap between what's real and what's driving asset prices, and how fragile that makes the system as a whole. We'll get back to the show in just a second, but first, let's talk about the gap between knowing what you should do and actually doing it. >> [music] >> You know nutrition matters. You know skipping meals tanks your energy and focus. But, [music] when you've got back-to-back meetings, deadlines stacking up, and a million decisions to make, execution fails. [music] That's where Huel comes in, and they've just launched into Target stores nationwide with two products designed to close that [music] gap. First, Huel Black Edition ready-to-drink. This is a complete meal, 35 g of protein, 27 essential vitamins and minerals. Second, their Daily Greens ready-to-drink, developed by registered nutritionists and dietitians. If you're serious about your goals, your nutrition cannot be an afterthought. Huel makes [music] execution simple. They're in Target stores now, or you can get 15% off your first purchase for new customers [music] with code impact at huel.com/impact. Now, let's get back to the show. So, welcome to part two, the AI reality gap. Productivity just isn't matching prices. According to a 2025 Startup Trends Report, AI has captured 64% of VC funding in the US, but a full 70% of these companies have yet to earn a dime in revenue. Even companies like OpenAI lost $5 in 2024 despite billions of dollars in revenue. MIT's 2025 report notes 95% of GenAI pilots fail to positively impact companies' P&Ls due to escalating costs and poor integration, indicating that costs currently dwarf benefits. Media mentions of AI in a financial context went from approximately 500 in Q1 of 2022 to over 30,000 by Q3 of 2023. That's an increase of roughly 6,000% Trading multiples on revenue, typically stated as enterprise value to revenue ratio, is way higher than normal, like way higher. Traditional SaaS companies ring in at about 6x. AI as a sector is clocking in at roughly 30x with outliers like xAI ringing in at a staggering 150x. That's the equivalent of banking on 150 years' worth of revenue at today's rates. That is wild. If you're not used to these numbers, it may not register just how insane that is, but it's insane. It's like paying 30 years' salary for an engagement ring instead of 3 months. Your wife would have to be Elon's daughter and guaranteed to be in the will for it to be worth that. AI investment is now growing faster than any technology in human history, and yet, the one metric that actually tells us whether society is getting more efficient, productivity, is growing modestly at best. >> [music] >> If you strip away the hype, the demos, the runaway valuations, and the endless parade of breathless headlines, the only thing that's left is economic transformation. And that just hasn't happened yet. But, we're seeing sky-high valuations nonetheless as if the transformation had already taken place. Asset markets are future prediction machines. They are the polymarket of stocks. That frame of reference makes it clear why markets are racing so far ahead of the reality. Now, listen, I am not saying that AI isn't real. I'm like the biggest AI booster ever. I believe that on a long enough timeline, it will be just as transformational as a 150x valuation would have you believe. However, when you look at hype cycles, the hype tends to wear off [music] faster than the reality arrives. And if the gap is too large and too speculative, early promise in a sector becomes a bunch of toxic bets that weaken the system as a whole. The dot-com bubble obviously did not mean that the internet wasn't going to be utterly transformational. It clearly was. But, I'll let you guess if that's comforting to the people who went broke betting on pets.com. "What if I'm wrong?" is always in the back of people's minds, and when that uncertainty reaches a breaking point, confidence falters and prices correct violently, regardless of where the industry ends up going on a longer timeline. And that's even if there are fundamentals. What happens when there's not even that? Whoa, Nelly. Hold on to your hat. You can wake up and realize not only are you not in Kansas anymore, you're not even on planet Earth. Consider this. There is massive societal turbulence between our current 150x valuations on xAI and AI actually delivering on its promises. If AI does what everyone expects it to do, what happens to society itself? If AI delivers on a tenth of its promises, it's going to disrupt the labor force more than the pandemic, and it's not going to just be blue-collar jobs that get affected. Having a job where people can work from home is not going to save people this time. We are in this weird position where if AI actually ends up having the capabilities that people have already priced in, it's going to cost millions of people their jobs just in the US. The global displacement could be measured in the tens of millions or even more. Will people be able to afford shares? Will shares still even matter? And with all that looming, we've pulled forward 150 years' worth of today's revenue? And even if you set that terrifying question aside for a second, a startling percentage of AI startups aren't even really tech companies. They're thin wrappers around the same four foundational models. They don't have moats. They don't have margins. They don't even have proprietary data. Their business model, if you can call it that, relies on reselling someone else's compute-intensive model with a prettier skin on top. Despite that, many of them are getting outsized valuations in the market. Then, you have the cost side, which is the part most people wildly underestimate. AI is incredibly expensive to train, deploy, and run, and because of that, much of the productivity gains they deliver are being [music] swallowed by compute costs, inference costs, retraining costs, and the engineering talent required [music] to keep everything from falling apart. It is no exaggeration to say that AI is the single most expensive productivity tool ever attempted, which is exactly why OpenAI and others are losing money hand over fist, and why Sam Altman went to the government hoping for a backstop. But, the market has already priced in all of these problems getting solved. That means there's a lot of ways that things could go wrong, but only one for things to go right. AI has to outperform everyone's already sky-high expectations. The catch is, the gap between the world people are betting on and the [music] world we actually live in is massive. And when expectations grow faster than reality can deliver, the tension becomes dangerous as all it has to do is knock people's confidence hard enough to trigger a psychological contagion that causes buyers to evaporate. Nothing bad actually has to happen to a business. People just have to lose faith, and we've seen this movie before, more than once. Housing in 2008 and the already-mentioned parallel of the dot-com bubble are two recent examples. Perfect companies became casualties when the narrative failed to overcome the math. The AI boom is real, but how much higher can valuations go before people lose faith in a future that's just too far away? Every bubble starts with the same early signs, a concentration of belief in a small handful of companies, expectations that blow through the ceiling, a market that stops pricing risk and starts acting like success is inevitable, and we can see those same specifics playing out again crystal clear with AI. And then, you always get a signal, a moment where something objectively good happens, but the market still reacts negatively. That's the tell. It's the crack in the ice before you fall through and get a cold, expensive dose of reality. Nvidia's recent perfect earnings equal quick rise and fall right back down is that signal. It doesn't mean Nvidia is doomed, but it does mean the market [music] is starting to get shaky. With valuations this detached from fundamentals, people are starting to get tempted to remember that saying one day AI is going to be huge is not the same [music] as getting the timing and the specific picks right. And getting the specifics wrong or the timing wrong is the same as being wrong. Now, the point of all of this is not to panic. The point is to understand what really drives valuations if it's not business success. Reflexivity is not the only factor that explains why trillion-dollar companies are now swinging like small caps. And that brings us to part three. What's the real cause of the AI bubble? In 2024 alone, global liquidity increased by over six trillion dollars, and historically, every major bubble from Japan in the '80s to dot-com in 2000 formed in liquidity waves just like this one. In 2023, the average borrowing rate to trade on margin at major brokerages was about 5 to 6%. But the S&P 500 was returning 26.3%, and that is exactly why people have been shoveling money into the market and why margin debt has hit 1.1 trillion dollars for the first time in history. Asset prices now have two inflationary pressures on them, first by the Fed printing money, and second by margin purchases creating more demand as well. It's an insane double whammy as people chase a return from AI stocks that are already priced decades into the future. In the last year alone, daily trading volumes in the AI mega caps have eclipsed the combined trading volume of the entire stock market of most G20 nations. This is speculation gone mad. If reflexivity goes a long way towards explaining the psychology driving the AI bubble, fiscal dominance is needed to understand why the adults are never coming home and why the bubble cannot be slowly deflated. This is the part people really don't want to look at, least of all me, because this is straight black pill. But if we don't contend with the structural reality of what's happening, we will never be able to protect ourselves from the potential fallout. Right now, there is simply too much liquidity sloshing around the system for how hot the economy already is, and that liquidity has to go somewhere. It will not just sit still. It will always aggressively seek a return. That's what people do. And when the dollar is losing international credibility at the same time that it's being inflated through debt, deficit spending, and ever-increasing amounts of money printing, it certainly can't park in cash. It also can't flow into bonds when yields are barely keeping pace with inflation. This part is so important. In an environment like this one, capital becomes a heat-seeking missile targeting the biggest return it can find. And as discussed earlier, that return is far more likely to be predicated on narrative rather than business fundamentals. So, whatever narrative promises the highest future return is going to get that money. And right now, all eyes are on the belle of the ball, AI. In a sane environment without all of the debts, deficits, and margin dollars, we wouldn't be in this position. We would have long ago raised rates. Dear Fed, I'm talking to you. Yes, raised rates. That would have slowed everything down. More people would opt for the risk-free rate of return, and money would be too expensive for so much margin trading, which would deflate the prices. We might still be in a bubble, but I very much doubt we'd be in a 150x bubble like we are now. Interest rates should be acting like brakes on the market. As borrowing money gets expensive, people do less of it, and that decreases liquidity, slowing the economy and cooling speculation because money wouldn't have to go so far out on the risk curve to get a reasonable return. But because of fiscal dominance, raising rates is structurally impossible at the moment. So, the question is, what exactly is fiscal dominance? Fiscal dominance is when the government's debt burden has grown so large that the Federal Reserve can no longer raise interest rates without making it impossible for the government to make its interest payments without turning the money printer on full blast, which would run the risk of eventually hyper-inflating the currency. The money stays cheap, the market runs riot, bubbles form, and there's no way to slowly deflate them. The Fed wants to raise rates. It knows it should raise rates. And yet, it lowers them as slowly as possible, but it lowers them nonetheless. More money floods the system, pushing asset prices even further while making the dollar worth less and less with every dollar printed. If you own assets, life is good until the music stops and the bubble bursts. And if you don't own assets, you're having a harder and harder time making ends meet with each passing day. And the Fed is stuck holding rates somewhere between too high for politicians to be happy, which is why you're seeing the pressure, and too low to force the market to be disciplined. This is a classic debt spiral, and we're already trapped in it. But don't worry, no one seems to care. And if we ignore the problem, it's likely to go away, right? Right? Well, actually, sort of. When the system has this much cheap liquidity chasing this much future price optimism, and there are no adults in the room with the ability to cool things off, the problem will eventually fix itself. When? When the bubble bursts and prices slam back to Earth. The only way to avoid that result is for AI to deliver on its promise without somehow gutting the labor force. And that is a contradiction. For AI to deliver on its promise, it has to drive energy costs and labor costs to near zero. That would truly be a spectacular world of abundance, but there would be an immense amount of disruption on the way to that future. Lord knows, the easiest way to look stupid is to try and predict the future. But when I look at how the market is riding on the back of roughly 10 stocks, and even spectacularly good news fails to rally a stock for any meaningful period of time, and it all feels very precarious, especially through the lens of history. Under normal conditions, markets rise on fundamental growth and fall on fundamental weakness. But that's not what's happening right now. And here's the truly dangerous part. This isn't happening because people think AI is a scam. It's happening because people believe AI is real, but they don't believe these valuations are sustainable. Most traders, however, believe that they're smarter than the average bear and that they're going to get out just in the right time, despite the fact that famous investors like Michael Burry of The Big Short fame just liquidated his entire hedge fund and gave the money back to his investors because he no longer believes he can understand this market. It's so out of whack. And how could he understand it? How could anyone? It's not mathing right now. It's not even just the hype. It's the macroeconomic trap of fiscal dominance creating sustained cheap money plus an addiction to debt, deficits, and gambling on margin all mixed together with a once-in-a-generation narrative that has some believing that this time really is different. And all of that means that navigating this moment requires a very different strategy than simply buy the winners and hold. Burry's out, Buffett's sitting on cash, and Dalio sees the tale of two economies racing away from each other towards open conflict, all while anxieties climb over what artificial general intelligence could mean for society, let alone the dangers of superintelligence. Put it all together, and you've got a balloon balancing on the head of a pin. The slightest downward pressure, and bang, it pops. And that's why part four is so important. The people who made it through the dot-com bubble did not just believe in the internet's future. They understood the structure of the market they were doing battle with, and they adjusted accordingly. We'll get back to the show in a moment, but first, here is the brutal truth about scaling. Most entrepreneurs don't outright fail, they plateau. And if you're stuck right now, you know how true that is. It could be that your revenue flatlines every time you step away, or maybe you're trapped in a commodity market that's racing to the bottom, or maybe you're one of the lucky people who is navigating a very complex partner dynamic that turns every decision into a battle. These problems and a whole lot more can seem impossible until you break them all down into first principles. My partners and I used this thinking to grow Quest Nutrition by 57,000% in our first 3 years alone and scale to a billion-dollar exit. And now I'm teaching this framework to a select group of entrepreneurs >> [music] >> who are ready to scale. Now, I want to be clear, this is not for everybody because I'm looking to work with serious entrepreneurs that already have an established business and a proven track record of execution. If that's you and you want to learn how to break through your biggest business bottlenecks using first principles thinking, be sure to apply now. Just go to impacttheory.com/scale or click the link in the show notes. Again, that's impacttheory.com/scale. Now, back to the show. So, welcome to part four. What the dot-com bubble tells us is the path forward. At the peak of the dot-com bubble, investors poured money into over 4,700 internet companies. In the year 2000, however, the Nasdaq didn't just pull back, it plummeted by nearly 80% wiping out over 5 trillion dollars of market value and taking thousands of can't-lose tech stories with it. Pets.com [music] went from IPO to liquidation in 268 days. That's how fast a beloved brand with a Super Bowl commercial can go to zero when there's a ton of narrative but no real business. At one point, even Amazon had lost roughly 95% of its value. If you'd bought the Nasdaq at the peak in March of 2000, you had to wait 15 years just to get back to even. The investors who bet everything on the hot names in 1999 mostly disappeared, but the investors who focused on real revenue, real infrastructure, and real diversification are the ones who today own a huge chunk of the modern world and have the wealth to prove it. Amazon, for instance, had a real business, so it was able to weather the storm. And then it went on to become one of the biggest companies in the world, and its stock price went up by roughly, drumroll, please, 100,000%. A full two decades after the boom, a handful of internet companies like Amazon, Google, and eBay, they still remain relevant, proving that if you look beyond the hype, there are often real businesses available to invest in. But assuming you know who's who early on is a very dangerous game. There's much we can learn about today from the dot-com bubble if we zoom out. It is the only other tech mania in human history that really rhymes with what we're living through right now in AI. Same belief that we are stepping into a new age. Same sense that if you're not all in on this, you're going to be left behind. The difference between the people who got vaporized and the people who came out the other side with life-changing wealth wasn't who believed in the internet. Almost everyone believed. The difference was how they invested through the mania. That's what we're going to walk through right now, the five pillars that would have helped during the dot-com bubble and that will [music] almost certainly help us now. Pillar one, be humble when placing your bets. In 1999, the smartest people in the room were absolutely certain they knew who the winners were going to be. AOL was going to own the internet forever. [music] Yahoo was the operating system for the web. Pets.com, eToys, Webvan, those were the can't-miss category killers. Amazon was just a joke to most traditional investors, just a money-losing online bookstore. Cisco, however, was so dominant, people were seriously saying it would be the world's first trillion-dollar company at a time where that was wild. They did not think they were guessing. They were convinced they were [music] right. But here's how it actually played out. AOL collapsed 98% after the Time Warner merger and became a punchline. Yahoo lost 96% of its value and never recovered its former relevance. [music] Pets.com, Webvan, Cosmo, eToys, all bankrupt and gone within roughly 18 months of the peak. Cisco still hasn't hit a trillion-dollar valuation even now, and that's 24 years later. And Amazon? Well, you know how that turned out. If you thought you knew who the winners were in 1999, you were almost guaranteed to be wrong. And if you were right, you had to survive all of your picks dropping massively in value. That's why humility in the face of so much uncertainty is an investing superpower. Ray Dalio has many times said that he built the largest hedge fund in the world by realizing just how often he's wrong. Pillar two, own the picks and shovels. Infrastructure requires less hype. During the gold rush, the people selling picks, shovels, and denim got rich. The guys who were actually panning for gold did not. The dot-com era played out exactly the same way. There's a reason that Nvidia is out to an early lead in AI. Underneath all of the crazy websites and doom portals, you had the picks and shovel companies. Semiconductors, networking gear, data infrastructure, servers, the boring stuff that actually made the internet work. Companies like Intel, Cisco, Qualcomm, Oracle, they got smashed in the crash. Nobody was spared. But here's where it matters. They survived. Intel recovered and kept compounding. Oracle eventually traded to multiples of its dot-com highs. Qualcomm not only recovered, it went on to 10x from the ashes. Cisco never made new highs, but it remained a backbone of the modern internet. Now, compare the app layer darlings of our time to their dot-com equivalents. Infospace, Lycos, Excite, Pets.com, Webvan, Cosmo, eToys, one after another, they either went bankrupt or lost 98 to 99% of their value and never recovered. If you'd owned a basket of semiconductors, networking, and core infrastructure on the other hand, you would have taken a brutal hit during the crash, sure, but [music] you would have been alive to participate in the next 20 years of the internet's growth. If you'd concentrated on the shiny front end, however, you were done. Translated back to AI, the picks and shovels are compute, chips, data centers, networking, infrastructure software, energy, security, etc. The stuff every AI application needs, whether it hits big or strikes out. That's a very different bet than going all in on a consumer-facing chatbot that could be obsolete the second the underlying model changes. Pillar three, bet on real revenue, not narrative. The dot-com survivors had one boring thing in common. They had actual paying customers for a product that actually worked. In 1999, Amazon did about 1.6 billion in revenue, not exactly lighting the world on fire. In 2000, 2.7 billion. In 2001, 3.1 billion. Going in the right direction. While their stock fell 95%, their revenue almost tripled. They were building real logistics, real customer relationships, real infrastructure. The share price was hallucinating, not the company. Pets.com was the flip side of that coin. In '99, the revenue was about $619,000. In 2000, roughly 5.8 million, but the net profit was zilch. They lost around 147 million dollars, and it wasn't just them. Many, many others were in the exact same boat. The companies that survived the dot-com collapse were the ones that could actually charge money and keep customers coming back, period. The rest were narrative plays. It's all investing theater. We're seeing the exact pattern with AI. About 70% of AI startups right now have no meaningful revenue. The vast majority are wrapping someone else's foundational model and calling it a business. The companies with positive cash flow are the ones that are going to be able to survive long enough to become part of the AI-enabled future. Pillar four, don't use leverage, and please, diversify. If you tried to [music] pick the winners in '99 by loading up exclusively on Infospace, WorldCom, Nortel, Yahoo, Global Crossing, and the other once-hyped public failures, you got absolutely wrecked. Most of those names went to zero and got dragged through bankruptcy court. And if you used excessive leverage, you almost certainly got wrecked. If, on the other hand, you did not use leverage and you owned a boring, broad basket of tech and sought out uncorrelated revenue streams, your outcomes were very different. And if you stayed in for decades after the crash, odds are you were up big time, life-changing wealth. The exact life-changing wealth that makes millennials hate the boomers. A diversified portfolio of future-leaning names would have almost certainly included Amazon, up around 6,000x since IPO, Apple, up roughly 4,000x from the early 2000s, Google, up approximately 70x since IPO, Nvidia, up 600x just since 2001, and inevitably, a whole bunch of losers. It's just the way it goes. But you didn't need to know which ones were the winners and which ones were the losers. You just needed to have enough humility to hedge your bets and own the entire sector. That's certainly my approach to AI. Pillar five, hold forever >> [music] >> whatever survives the inevitable crash. This is the part almost nobody gets right. The dot-com bubble did not make people rich in '99, it made people rich from 2002 to 2024. The real wealth came after the crash, long after the crash, for the people who survived it with the right assets in hand and the emotional resilience to hold them. Amazon fell about 95% after the bubble burst. From that bottom, it went up over 100,000%. $10,000 put into Amazon in 2001 turned into something on the order of $13 million. Apple, once left for dead, went on to split over and over and compound something like 600X from the early 2000s. Google grew from a modest IPO to one of the most dominant companies on planet Earth. Nvidia went from a sub-dollar curiosity to one of the most important firms ever. The pattern is clear. The wealth wasn't made by predicting the bubble or predicting the winners. The wealth was made by surviving the bubble and then holding the winners as they rebuilt the world slowly over time. The goal is not to perfectly time the top or the bottom of the AI cycle. The goal is to structure your portfolio and your psychology so that A, you don't get wiped out when the narrative breaks. B, you still have exposure to the real innovators. And C, you're emotionally and financially able to hold what you own for a decade or more as the real productivity gains may take years to get fully realized. If you put all of this together, reflexivity, the current productivity gap, fiscal dominance, and the lessons of the dot-com bubble, a very clear picture begins to emerge. AI is almost certainly going to be revolutionary. The world is almost certainly going to change forever. And yet, markets can remain irrational for far longer than most people can remain solvent, especially when you remember that the markets are manipulated by hype and hyper-experienced traders with very deep pockets and a profound understanding of how markets move, and some of them even like to see other people lose. Plus, humans are just irrational. They do everything, especially investing, based on emotions, not mathematics. We over-believe, we way over-leverage, we confuse narrative with reality, we price in the future long before it gets here, and when money is cheap, we shovel it into markets as fast as we can until something forces us to stop. The dot-com bubble wasn't a story about predicting winners. It was a story about surviving long enough to let the winners reveal themselves. The AI era will be no different in that respect. AI is [music] real. It's already powerful. Over the long run, it will almost certainly be far more transformational than the internet. But that doesn't mean every AI investment is wise, that every valuation is justified, or that every company with AI in the name will even exist 10 years from now. Your job is not to outguess everyone else. Your job is to stay humble, own the infrastructure, look for real economics, diversify your exposure, and hold on to the survivors when things finally reset. Do that, and the AI bubble won't be the thing that [music] wipes you out. It'll be the moment the real gains begin. If AI doesn't kill us all first. All right, you guys, if you want to see me explore topics like this in real time, be sure to join me live Wednesdays and Fridays at 6:00 a.m. Pacific on YouTube, X, Twitch, or Kick. You can join the debate or just chill in the community. I hope to see you there. Till next time, my friends. Be legendary. Take care. Peace. If you like this conversation, check out this episode to learn more. IRS enforcement staffing has fallen by 34% even as the tax code ballooned to more than 4 million words. Over 25