cleaned_19582640.html Sign in 100%
Ryan Miller
Everyone's asking the same question right now. Is the AI a bubble? Yes or no? Pick a side. And I'm telling you, that's the wrong question. That question keeps you a spectator. It keeps you sitting in the stands arguing with strangers while people who actually move capital are already three moves ahead of you. So here's the right question. The one that makes fund managers rich. When this thing wobbles, and it will wobble, are you going to be a forced seller or are you going to be the buyer of the wreckage? And by the end of this episode, you're going to know exactly which one you are, and you're going to have the tool to prove it in your own portfolio.
Ryan Miller
Before we dive in, just a word from our sponsor. When doing deals, we all know that raising capital is the one thing that unlocks everything. That's why I've partnered with Reef Pass Investors that are actively funding deals right now. So if you're a deal syndicator or founder thinking about launching an M&A focused buy and build platform, reach out to Reef Pass Investors at reefpassInvestors.com. They are one of the best investors in the game that are helping you launch a new long-term holding company. So here's what I want you to do: click the description in the notes and contact them for a discovery call and potentially get an invite to pitch your next M&A deal. Now, let's get back to the show.
Ryan Miller
Before we dive in, just a quick disclaimer. Nothing in this episode is legal, financial, or tax or investment advice. This is just for entertainment purposes, and you should always check with accredited professionals before making any decision from this show or otherwise. Now, that's that. Let's dive in.
Ryan Miller
So on the show, you know we don't chase headlines, we position capital around them. So let me say the thing that's going to get me in trouble with both sides of this argument. The AI bulls are going to hate me for this. The AI doomers are going to hate me too. Good. That's usually a sign that you're standing somewhere true. So here it is. AI is real. The technology is a genuine revolution. Maybe the biggest productivity shift since electricity. And the financing structure has been built on top of the last 24 months, is one of the most fragile capital architectures I've ever seen in my career. Both of those statements are true at the same time. And if you can hold both of them in your head without flinching, you're already ahead of 95% of the market. And why is that?
Ryan Miller
Well, I think it's because there's something most people can't process. They think the technology is real, and this is a bubble, are opposites, where I would argue they're not. They have never been opposites. So go back to 1999. The internet was real. The internet changed everything, and the dot com and telecom bus still erased $5 trillion in market value. It wiped out pets.com, it wiped out WorldCom, it wiped out a generation of investors who confuse the technology will win with my position will survive. Does that make sense? So the technology won. Their capital did not. That is the entire lesson. And almost nobody learns from history.
Ryan Miller
So let me define what a bubble actually is because people throw that word around like it just means prices went up a lot. That's not exactly what it means. A bubble is what happens when price detaches from the cash flow that justifies it. And the gap that gets financed, it's done by the belief of some event instead of actual earnings. That's the technical definition that matters to an allocator. It's not about whether something is overpriced. Lots of great things are overpriced. A bubble is specifically when the funding of that thing depends on everyone continuing to believe. When the capital structure itself can only stand up as long as confidence holds, then pull the confidence, and it's not that price drifts down. It's that the financing mechanism actually stops functioning. The buyers of last resort vanish, the leverage calls itself in, and it is a fundamentally different animal than an expensive stock. And it's why, but the technology is real, is no protection at all. The technology was real in 1999 too. The financing, it still failed. So I'm not here to tell you AI is fake. I'm here to tell you that being right about the technology and being right about your capital are two completely different skills. And the second one, the capital one, is the only one that will get LPs paid.
Ryan Miller
So let me show you the machine underneath. Three mechanisms, then I'll show you how to position. See, everything on this show runs through one formula. It's the Fundraise Formula. It is trust times the transaction equals capital raised. And capital never is just about the deal. It's about the trust that lets you do the deal, multiplied by the mechanics of the transaction itself. Does that make sense? And here's why that formula matters for a bubble specifically. So in a manic episode in the markets, transactions go vertical. Everyone is transactioning. Money is flying. Deals are closing in days. But trust gets thin, trust gets stretched. And the whole structure starts to depend on everybody believing everybody else is good for the money. And the moment the trust cracks, not the technology, the trust. The transactions seize up instantly. That's not a slow leak. That is a snap. So when I look at the AI economy, I'm not asking, is the tech good? I'm asking, where is the trust thin? And what happens when the transactions, when it breaks, what's going to happen then? And you want to watch where three places are, where it begins to get the thinnest.
Ryan Miller
So mechanism number one is the circular financing loop. This is the big one. And it's the one almost nobody outside of a few boardrooms genuinely understands. So let me give you the numbers because on this show we deal with numbers, not vibes. So in 2026, four companies, Microsoft, Google, Amazon, and Meta, are on track to spend roughly $725 billion in CapEx. That's up about 77% from the year before. Roughly three-quarters of it is all going into AI infrastructures like chips, data, power centers. Now, follow the money because this is where it gets interesting. NVIDIA sells chips. NVIDIA also invests in OpenAI. OpenAI then commits to spend hundreds of billions on cloud capacity for companies like Oracle, CoreWeave. And what do Oracle and CoreWeave do with that money? They go buy more NVIDIA chips. The money goes round and round in a circle. NVIDIA to OpenAI, OpenAI to Oracle to NVIDIA , and it just keeps going round and round. Analysts are now estimating north of $800 billion in these circular financing arrangements. And OpenAI alone has committed something like $1.15 trillion across seven major vendors stretching out to around 2035. $300 billion to Oracle, $350 billion to Broadcom, and $100 to NVIDIA and on down the list. But here's the problem OpenAI is projected to lose around $14 billion this year alone, nearly triple its last year's losses. It's a company committing to spend more than a trillion dollars while bleeding $14 billion a year. And it's promising to pay for that with revenue that it hopes to earn by 2029.
Ryan Miller
So ask yourself the trust question. When a vendor finances its own customer and that customer uses the money to buy vendors' products, is that real demand or is it manufactured demand? Because when you fund your own buyer, the incentive stops being does the end user actually need this? The incentive becomes book the deal. And the book deals look exactly like real revenue right up until the moment that they do not. And if you think that this is a new trick, let me take you back to this exact same movie last time. The telecom bottle, 1999-2000, there was a company called Lucent Technologies. Lucent Technologies had a brilliant idea to keep its growth numbers going. It would lend money to its own customers so they could buy Lucent equipment, vendor financing. Nortel did it too. The customers booked the gear, Lucent booked the revenue. Wall Street cheered on all that growth. And the stock went absolutely vertical. And then the customers, a lot of them flimsy startups, couldn't pay the bills when they came in. The receivables went into bad debt expense. The revenue turned out to be loans that were never coming back. Lucent went from market darling worth more than $200 billion to a penny stock that got absorbed for scraps. Same mechanism. A vendor financing its own demand. It is one of the oldest tells in the book, and it is running right now at that scale that makes Lucent look like a lemonade stand.
Ryan Miller
Now, look at how tangled ownership is this time. NVIDIA owns roughly 5% of CoreWeave and sells core weave to chips. OpenAI uses CoreWeave Cloud and owns a stake inCoreWeave too. Everyone owns a piece of everyone. Everyone is selling to everyone. And everyone is continuing the same dollar as it goes round and round and round in this circle. That's not a market, my friends. That is a hall of mirrors. And halls of mirrors look enormous right up until someone turns the lights on. That's the fragility I'm talking about. The loop manufactures the illusion of demand. And the day real end demand disappoints, the day enterprises don't renew, don't expand, don't even pay, the loop doesn't just slow down gracefully. It reverses violently because every link in that chain was counting on the next link to pay. And we already have a giant flashing warning light on real end demand. Hold that thought because it's mechanism twos wind up.
Ryan Miller
So mechanism two. There's two mirages here stacked on top of each other. And once you see them, you'll never look at an AI earnings report the same way again. The first mirage is on the demand side. MIT ran a study, the Gen AI Divide Report. And the headline finding should have stopped you cold. 95% of enterprise generative AI plots delivered essentially zero measurable return. Think about that. Zero measurable return. Companies poured $30 to $40 billion into these projects, and 95% of them got nothing on their PL. Not a little, nothing. Now the bulls will tell you that's an integration problem, not a technology problem. And they're probably partly right. The 5% that did it well, buying from specialized vendors, building real partnerships, succeeded about two-thirds of the time. So the value is real for the people who do it right. But here's what matters for capital. The spending is justified by a wave of enterprise adoption that, by MIT's own data, mostly isn't paying off yet. You're building $700 billion of capacity for that demand, and it is still charitably unproven. That's the demand mirage. Now stack the second one on top, the depreciation mirage. This one is sneaky and it is pure accounting. And it's where Michael Burry, the guy who called the 2008 crash, has planted his flag. The hyperscalers are depreciating their NVIDIA chips over five, sometimes six years. But NVIDIA is launching a new major chip generation every year now. Hopper, then Blackwell, then Ruben, then Ruben Ultra, year after year, which means the real economic life of these chips, the window where they're actually competitive for frontier work, is closer to two or three years, not five or six. So why does that matter? Because depreciation is an expense. If you stretch the useful life of an asset, you spread the expense over more years, which means less expense each year, which means, and say it with me, higher reported profit. Now, Bury estimates roughly $176 billion of understated depreciation and overstated profit across the industry between 2026 and 2028. He thinks reported operating income at companies like Oracle and Meta could be more than 20% above their economic reality. Think about what that means. A meaningful slice of the profit that's justifying these valuations may be an accounting choice, not an economic fact. And accounting choices that get corrected. When useful lives of assets get reset to reality, reported margins compress and they compress right when everyone is already nervous. You see how this works? So that's demand mirage and then the depreciation mirage. Two illusions stacked, both inflating the exact same story.
Ryan Miller
Now, I want to do something that most channels won't do because they're too busy selling fear. I want to give you the strongest version of the bull case, the real one. Because if you can't argue the other side better than another side can, you don't actually understand your own position. And as an allocator, you have to understand both. So here's the bull case, and it's not stupid. One, the technology genuinely works, and it is improving on a curve that we have rarely seen in our lifetimes or in many lifetimes. The 5% of enterprises that are getting real return on investment from AI are getting enormous returns on investment. We're talking step changes in productivity, not rounding errors. The value is real and it's just evenly captured so far. And adoption curves are always lumpy at the start. Then there's point number two. Being short of compute really is more dangerous than overspending. If you believe that technology wins, if you're Microsoft or Google and you genuinely think that this is the platform shift of the century, then underbuilding capacity is an existential mistake, and overbuilding is a survivable one. That is a rational bet, not a crazy one. Then there's point number three. And this is the big difference from 1999. The companies doing most of the spending are the most profitable enterprises in human history. Microsoft, Google, Meta, Amazon. These aren't pets.com burning venture capital cash. They are cash machines funding this large out of operating cash flow. That balance sheet strength is real. And it means this can run a lot longer and absorb a lot more pain than the dot-coms ever could. So I'm not dismissing the bull case, but I am respecting it. But here's where it doesn't save you. Even if the biggest spenders survive, and they probably do, that tells you nothing about the fragile middle, the circular counterparties, the levered cloud builders, the model companies losing $14 billion a year. The bull case protects the four giants. It does not protect the $800 billion of circular papers strung between them. And it does not protect your portfolio from the concentration repricing. A bubble doesn't need everyone to fail. It just needs the marginal financing to stop. The Giants Surviving and the structure cracking are completely compatible. Both happened in 2001. Amazon Survived .com. $5 trillion still died.
Ryan Miller
Hey, if you're finding value out of this discussion, could you do me a huge favor? Could you just take a second and hit that like or subscribe button? It costs you nothing, but it tells the algorithm that this is valuable information and it helps us to get it to more people. Thank you. You're incredible. Now let's get back to the show.
Ryan Miller
Then there's mechanism three. And this is the one that reaches into the pocket of every single person listening, whether you own a single AI stock or not. Concentration. The Magnificent Seven. The seven biggest tech names hit a combined market cap around $22 trillion. At their peak, they were about 36% of the entire SP 500. More than a third of America's flagship index sitting in seven companies. Think about that scale. And NVIDIA alone crossed a $4.5 trillion in market cap, making it about 22% of that magnificent seven just by itself. Now, here's why that is not strength. Everyone treats concentration like it's a sign of dominance. It's not. It's a single point of failure because of passive investing, because of index funds. Every 401k in America, every pension, every endowment that holds SP is now running a massive, unhinged leverage bet on AI capital expenditure, whether they know it or not. Most of them have no idea. And they think they own the market. But what they actually own is a concentrated AI infrastructure trade cosplaying as diversification. And here's the kicker the thing that Doomers got wrong, AI CapEx doesn't have to crash in order to hurt you. It just has to slow down. If the growth rate merely decelerates, if $725 billion doesn't become a trillion next year, the multiples compress. And when the multiples compress on 36% of that index, that move will cascade through every passive fund into every retirement account, into every portfolio on the planet. You don't need a crash. And pauses are a lot more common than crashes. So now you see the machine. A circular financing loop manufacturing demand, an unproven end demand, and inflated accounting for profit. And a concentration structure that turns a tech slowdown into an everybody problem. And that's the bubble. Not AI is fake. The technology is real. The capital structure, however, is quite fragile.
Ryan Miller
So let me walk you through how this actually breaks. Because if you can see the sequence in advance, you can position in advance. I'm not predicting a date. Anyone who gives you a date is selling you something that I want nothing to do with. I'm just going to give you the order of operations as I see it, my opinion. Because mania's always unwind in roughly the same order. It starts quiet. It always starts quiet. One of the circular counterparties misses a payment or stretches one or renegotiates. Maybe it's a cloud provider that can't deliver capacity on schedule. Maybe it's a model company that can't raise its next mega round at the valuation it needed. We already got a preview of this. Oracle's stock dropped about 30% in a single quarter on exactly those worries, which was can it actually deliver on its commitment to OpenAI? And can a company losing $14 billion a year actually pay for the compute that it just ordered up? That's the first crack. A trust crack, not a technology crack, and that's important. Then the credit market notices. Spreads on most levered names, they start to widen. The high yield desks start pricing in the possibility that some of this paper may not get repaid. And remember, Burry flagged that the high yield debt issuance in this AI build-out does look a lot like the debt issuance in the 99 telecom media technology bubble. Debt is the accelerant. When the debt market sniffs trouble, refinancing gets really expensive. And a lot of this build out was counting on cheap, endless refinancing. Can you see how that may not be the most rational thing, which is another hallmark of having a bubble? And then the accounting catches up. Somebody, an auditor, a regulator, an activist short seller, forces a useful life reset on the GPOs or a write down on the stranded capacity. And suddenly, 20% overstatement of operating income that Burry talked about becomes a real number on an earnings call. Margins compress. The profitable part of the story, it wobbles. And then concentration does its ugly work because 36% of the index is these names. The repricing doesn't stay contained. It pours through every passive fund into every 401k and every LP portfolio all at once. The forced sellers appear. Funds that were levered, funds that have redemptions, funds that have to raise cash, and forced selling is indiscriminate. It sells the good with the bad, which is exactly, exactly when the real assets go on sale. That's the moment. That is the window the whole episode is about. The forced sellers are dumping real infrastructure to raise cash, and almost nobody has the cash to buy it because almost everybody was fully invested at the top. The person with dry powder in that moment isn't just getting a discount. They're the only bid in the room. They get to name the price. So the only question left is the one I opened with, forced seller or buyer of the wreckage. Let's make you the buyer.
Ryan Miller
So this is where I split from every doom scrolling, fear-mongering channel on the internet. They say, sell everything, the crash is coming. That's not analysis, my friends. That is a tweet. Fund managers don't get paid to panic. We get paid to position. So here is the framework: four moves. Educational, not advice or guarantee of anything. But this is how an allocator is likely going to think about it. So move one, separate the physical layer from the financing layer. Underneath it all, the circular paper, there are real assets. Power generation, the grid, cooling, land, transmission. Those things have genuine cash flows and genuine scarcity, and they outlive any single chip generation. The financing layer, the circular revenue names, the companies whose revenue is partly their own capital making a lap. That's where the fragility lives. So when you look at your book or your portfolio, you should be able to say which of your exposures is a physical layer and which one is a financing layer. If you can't answer that, you don't know what you own.
Ryan Miller
Then move number two, don't short the revolution. And this is critical. Fund managers who got destroyed in 99 weren't only the ones who were long the junk. They also included the ones who shorted the whole internet because they knew it was a bubble and got run over 18 months before they were actually right and already bankrupt. So being early and being wrong are actually the same. So the move is not go cash and short the Nasdaq. The move is the barbell. Keep real sized exposure to the genuine productivity winners. And while you hedge the financing structure risk around it, so you want to participate in the upside and ensure the downside.
Ryan Miller
Then there's move three. And this is the alternative assets to play. This is the one that's my world. Dry powder is the edge. In a mania, everyone's fully invested. Everyone is levered. The one thing nobody has when the music stops is liquidity. And liquidity in a force-selling environment is the single most valuable asset on the planet. The fund manager who walks into the unwind with committed dry powder gets to buy the real infrastructure, the data centers, the power assets, the physical layer from distressed sellers on pennies on the dollar. Sovereign wealth funds do this all the time. The best private equity shops, they also do it. They don't try to time the top. They just build the war chests, they stay disciplined, and then they wait to be the only buyer in the room. That's not luck. That's positioning and structure.
Ryan Miller
And let me get specific about the structure, because we have dry powder is easy to say and hard to do when you're a fund manager under pressure to deploy. Your LPs hate idle capital. Every quarter you're not deployed. Somebody asks why they're paying you a management fee to sit on cash. So building dry powder isn't a market call, it's a communication discipline. It means going to your LPs now, before the storm, and selling them on patience. It means structuring your fund with a real capital call mechanism so the money is committed, but not yet drawn. Committed dry powder, not cash drag. It means setting your investment period term so you have the runway to wait 12, 18, 24 months for the window without being forced to deploy into the mania just to put the money to work. The managers who get this right aren't smarter than everyone else. They just had the harder conversation with their LPs early. And they had it from a position of trust. There's the formula again. The trust you bank today is the patience your LPs grant for tomorrow. And I'll be transparent with you because that's how we do it here. The instinct to deploy when everyone else is deploying is one of the hardest things to resist in this business. The fear of missing out doesn't just affect retail investors on their phones. It affects professional allocators like you and I who are sitting on billions because the career risk of underperforming during mania feels more real than the portfolio risk that it could unwind. I've watched brilliant managers abandon their discipline in the last innings of a cycle because standing still felt like falling. See, discipline in a mania doesn't feel like strength. It feels like you're missing the party. And then the lights come on, and the disciplined ones are only the people still standing with a wallet.
Ryan Miller
Then move four. You want to watch for the right signals, not the headlines. And here's your dashboard. One, hyperscaler free cash flow turning negative. Amazon is already projected to go free cash flow negative this year under the weight of the spend. Two, any announcement of GPU useful life resets because that depreciation mirage is correcting in real time. And then three, credit spreads on the circular deal counterparts that we mentioned. Because the debt market always smells at first. Four, private credit exposure to AI data centers, because that's where the leverage is hiding in this cycle, I think. When those signals move together, the unwind has started, and your dry powder becomes the most important position in your fund.
Ryan Miller
And let me bring it back to the formula because it's the whole point. Trust times a transaction equals capital. In a bubble, everyone has access to transaction. But what separates the survivors is trust. The fund manager who has spent years building genuine trust with their LPs, who has been transparent, who's been disciplined, who didn't chase the mania. That manager is the one who can raise the fund that buys the wreckage. Your reputation for the discipline is not a soft asset. In a downturn, it is the hardest asset that lets you raise capital when no one else can. The trust you build today is the dry powder you deploy tomorrow.
Ryan Miller
So let's go back to the beginning. Is AI a bubble? Well, I've just shown you it's the wrong question. The technology is real. The capital structure is fragile. Both are true. And the people who understand the difference, who can hold both ideas at once, those are the ones who don't get carried out on a stretcher when this thing finally takes a breath. I've learned this discipline the hard way. Long before I ran a dollar of anyone else's money. When you've been at the bottom, when you've had nothing and had to claw your way up, you develop a very specific relationship with risk. You stop confusing motion with progress. You stop needing to be in every deal to feel like you're winning. The people who've never been wiped out are the ones who get wiped out because they don't respect it. They don't respect the downside. I respect the downside. And I always think that that respect gives me, and anyone else who has been in this game for a while, gives us an edge. Let's you sit on dry powder while everyone, they can call you a coward or maybe whatever it is they want to do, and then write the check of your career while everyone else is still disinfecting the wounds. The amateurs are arguing about timing in the pot. You're not going to do that because nobody can. The allocators are doing something entirely different. They're separating the physical from the financing. They're keeping their upside but ensuring their downside. They're building dry powder while everyone else is fully invested or all in. And they're protecting the one asset that lets them raise capital in the storm. Trust. That is the whole game, my friends. You don't fear the unwind, you fund it.
Ryan Miller
And understand what is really at stake here because that is bigger than any single cycle. The next 18 to 36 months are going to transfer an enormous amount of wealth from one group to another. From the people who confuse being right about technology with being right about capital to the people who knew the difference. From the fully invested to the patient, from the leverage to the liquid. That transfer is already loaded into the structure I just walked you through. The only thing left undecided is which side of it you're going to be on. And it's being decided right now while it's still calm. Not later when it's chaos. The positioning you do today before anything breaks is the entire ballgame. Nobody gets the position after the move. They only get to react to it. Now, if you want to go deeper than a podcast can take you, if you want to actually run the numbers on your own book, I built you a tool. It's the AI Bubble Exposure & Capital Positioning scorecard. You plug in your concentration, your circular revenue exposure, your capex coverage, your dry powder ratio, and it scores exactly how fragile or how positioned your portfolio is right now. And which of the four tiers that you're probably in spectator, exposed, position, or predator. It's free. The link is in the notes. And if you're a fund manager who wants to do this work shoulder to shoulder with people who are operating at the highest level, fund managers raising real institutional capital, building real war trusts, that's exactly what we do inside the Fund Raise Capital community. It's where the frameworks on this show become the capital in your fund. I'd love to have you in this room. So download the scorecard, run your book, find out which tier you're in before the market finds it out for you. You do these things, and you too will be well on your way in your pursuit of Making Billions.
Displaying cleaned_19582640.html.
Connect with Ryan Miller: Website | LinkedIn | Instagram