AI Is a Bubble — How the Smart Money Is Quietly Getting Rich Underneath It

LEARN THE CAPITAL RAISING STRATEGIES AND FRAMEWORKS used by alternative asset professionals: https://go.fundraisecapital.co/apply

This episode is brought to you by Reef Pass | Serial Acquisition Investors: Reef Pass Investors has spent the last 10 years focused on partnering with founders to launch and build long-term holding companies, and has a proven track record doing exactly that.

To reach out to Reef Pass Investors, email holdcofounders@reefpassinvestors.com

DOWNLOAD The AI Infrastructure Capital Scorecard: https://go.fundraisecapital.co/100m-blueprint-860536-648168-166304-426463-122119-538372-417197

Download Ryan Miller's AI Infrastructure Capital Scorecard to evaluate AI investments, identify infrastructure opportunities, and avoid speculative risk.

 Is AI a bubble?

AI infrastructure investing is the question defining private capital in 2026 and most fund managers are answering it wrong.

In this episode of Making Billions, Ryan Miller asks the sharper question: where in the AI capital stack do contracted, asset-backed cash flows actually live? 

This episode delivers the framework every serious allocator needs. Not hype. A five-rung capital stack separating infrastructure money from bubble money and a scoring tool making it impossible to confuse the two in a live deal.

The pattern has played out before. In 1999, dot-com equity was vaporized, trillions gone. 

The fiber laid during that bubble still carries traffic 25 years later. The bubble lived in the equity. The durable money lived in the infrastructure. 

AI is following the exact same pattern in real time. Allocators who see both things at once come out of this era wealthy.

[THE HOST]: Ryan Miller is a fund manager, capital strategist, and former CFO turned angel investor in technology and energy. He is the founder of Fund Raise Capital and Aequor Capital Partners, and has mentored over 1,000 fund managers across private equity, private credit, venture capital, real estate, and alternative assets globally.

Subscribe on YouTube:
https://www.youtube.com/channel/UCTOe79EXLDsROQ0z3YLnu1QQ

Connect with Ryan Miller:
Linkedin: https://www.linkedin.com/in/rcmiller1/
Instagram: https://www.instagram.com/ryanmilleroffical/
X: https://x.com/_MakingBillions
Website: https://making-billions.com/

Transcript

Ryan Miller

Everyone's screaming that AI is a bubble. And you know what? They're probably right. See, valuations are stretched, hype is deafening, and a lot of people are going to lose a lot of money. But here's what nobody's telling you. While the entire internet argues whether the bubble pops, the smartest capital on Earth, it is quietly making a fortune underneath it all. Not in the names you're seeing on CNBC. Underneath it. And by the end of this episode, you're gonna know exactly where that money is, why it's safer than the hype, and how to underwrite it like an allocator instead of gambling like everybody else. 

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 investment advice or tax advice. Everything here is for entertainment purposes only, and you should always check with accredited professionals before making any decision from this show or otherwise. Now, with that said, let's dive in.

Ryan Miller

 So let me tell you why this matters right now. Today, we are living through what may be the single largest infrastructure build out in modern history. The amount of capital being committed to AI, to the data centers, to the chips, to the power required to run them is absolutely staggering. We're talking hundreds of billions of dollars a year, heading towards trillions over the next decade. Numbers that rival the build out of the railroads, the rollout of the electrical grid, the construction of telecom networks. And whenever you get a capital wave that big, two things happen at exactly the same time. A bubble forms at the top in the equity, in hype, and the names everybody is talking about. And a once-in-a-generation opportunity forms in the bottom, in the infrastructure, in the boring, contracted, asset backed cash flows that nobody's talking about. See, most people only ever see the bubble. The fund managers who also see the second thing, the ones who can tell the difference between the two. Those are the people who come out of this era wealthy and well respected. That is what I want to give you in the next part of this episode. 

Ryan Miller

So let's start with the most important reframe of this entire episode. The herd is asking one question. Is AI a bubble? And I'm telling you, that is the wrong question. It's a spectator's question. It's a question you ask when you're watching the game from the stands. A capital allocator asks something far sharper. The allocator asks, where in the AI capital stack do durable, contracted, asset backed cash flows actually live? And where is the equity like hype just wearing an infrastructure costume? Because here's the truth. That took me years to really internalize. Bubbles pop at the equity layer. They almost always do. But the infrastructure underneath those bubbles very often keep right on playing. So let me prove that to you with a little bit of history. Because this exact movie has played out before more than once, and the ending rhymes every single time. 

Ryan Miller

So go back to the gold rush, 1849. Hundreds of thousands of people flooded to California chasing gold. The vast majority of them went broke. Backbreaking work, brutal conditions, terrible odds. Most of them never struck anything worth keeping. But a small handful of people got fabulously generationally wealthy in that era. And almost none of them ever swung a single pickaxe. They sold the picks, they sold the shovels, they sold the blue jeans. That is literally how Levi Strauss built an empire that's still standing today. They owned the water rights, they owned the supply stores. They financed and built the railroads that carried the miners and the gold in and out. The miners took all the risk and got lottery type odds. It was the infrastructure players who took the boring, contracted, repeatable cash flow. They're the ones who won decisively. 

Ryan Miller

Now, fast forward 150 years to the dot com bubble, 1999 to 2000. Everybody remembers pets.com, web fan blowing up. The equity got literally vaporized. Trillions in paper wealth just gone. But ask yourself, who actually survived through all of that? The companies that laid the fiber, the data infrastructure, the picks and shovels of the internet itself. A lot of that fiber got laid during that bubble and is still carrying traffic and is still generating returns 25 years later, long, long after the hype names became trivia questions. It's the same pattern every single time. The bubble lives in the equity and the durable money lives in the infrastructure. And right now in AI, I can see that exact same pattern repeating right in front of our eyes. And you have a chance today to be on the right side instead of the wrong one. 

Ryan Miller

So let's get specific because I don't deal in hypotheticals or vague inspiration. I deal in frameworks that you can actually use. So I'm going to walk you through what I call the AI capital stack. Think of it as a ladder. And at the top of the ladder is the safest, most asset-backed money. At the bottom is the most speculative, equity-like money. Five wrongs. And understand these wrongs. And I promise you, you'll start seeing deals completely different than 99% of the investors out there. 

Ryan Miller

So there's wrong one, the top of the stack, the safest. Senior secured infrastructure debt. This is lending money against a real physical hard asset, a data center, a power plant, a substation where you hold the senior position. You're secured by the asset itself. And ideally, there's a long-term contract underpinning cash flow. So if everything goes wrong and investing, you can always plan for everything that goes wrong at least once. You're the first in line and you're holding hard collateral. This behaves like infrastructure, contracted, senior, asset backed. It is the polar opposite of a meme stock. And here's the part that should get your attention because this whole sector needs so much capital so fast, the returns on this senior protected money are far better than what senior secured lending normally pays. You're quite literally getting compensated like it's risky for taking risk that's structurally quite low. And that gap, that's the opportunity. 

Ryan Miller

Then there's rung two, private credit and mezz financing. This is the capital that actual financing happens in the build-out, funding the developers and the operators who are putting up the data centers and locking down the power. It carries more risk than a senior secured layer, but it is structured, it has covenants, it has protections, controls, and seniority over the equity. And here's the macro story you need to understand. The banks have pulled back from a lot of this kind of lending, and the build-out needs far more capital than the public markets can comfortably supply. So private credit is rushing in to fill that gap. And it is eating this entire space alive in a good way. A staggering amount of smart, patient, institutional money is flowing into private credit around AI infrastructure right now. And the average retail investor has no access to it. And quite frankly, it has no idea that it's even happening. 

Ryan Miller

That brings us to run three, the middle of the stack. And in my opinion, the single most underappreciated layer in the whole game, it's the real assets themselves, the land, the power generation, the grid interconnection rights. I'm going to come back to power in a few minutes because it is important. It's so important that it deserves its own segment. But lock this in right now. Whoever controls the scarce physical inputs of the AI build controls a moat that no amount of venture capital can replicate. You can print money. Venture capitalists can conjure a billion dollar valuation out of a pitch deck, but nobody can print a permitted gigawatt of power connected to the grid and deliverable in the next 18 months. That scarcity is real. It is physical and it is the most durable moat in this entire theme. 

Ryan Miller

Now that brings us to rung four, infrastructure equity. This is owning the facilities outright, owning the data center, owning the power asset, owning the physical thing. It's equity. So it carries equity risk, but it's equity backed by hard cash-flowing real-world assets. And if you watch how the big infrastructure funds, the pensions, the sovereign wealth funds, are actually playing AI, an enormous amount of money lives right here. So pay attention to that. The most sophisticated, longest horizon capital on the planet is not buying the chatbot. It's buying the building that the chatbot runs in. 

Ryan Miller

And then there's rung five at the very bottom. The most speculative. Application and model equity. This is the headline layer. The AI startups, the model companies, the names everybody's fighting over and arguing about online every single day. Now look, I want to be fair. Some of those companies will be generational winners, a few. But this is venture risk. This is bubble beta. This is the layer where the bubble actually lives. And it is a layer where the most amateur capital is crowding in the hardest, precisely because it's the only layer most people can even see from where they're standing. So here's the entire game compressed into one sentence. Most investors are crowding into rung five and calling it investing in AI. The allocators who are going to win this decade are climbing up to rungs one, two, and three, and they're getting paid. Contracted asset backed protected returns on the exact same mega trend with a fraction of the downside. Same wave, completely different risk profile. That right there, that is the edge. That's the whole thing. 

Ryan Miller

Now, I want to give you the single most important constraint in this entire build-out. The thing almost nobody outside this industry truly understands is the thing that separates a real underwriter from a tourist with a thesis is power. And here's the reality: you can't run an AI without an enormous amount of unimaginable amounts of electricity. These data centers do not draw what a normal building draws. A single large AI data center can draw what an entire small city draws. And the binding constraint of the entire AI boom right now, it's not chips and it's not even capital. It's power. Specifically, secured, permittable, deliverable power, physically connected to the grid on a timeline that actually matters to investors. In a lot of regions in this country and around the world, the interconnection queue, the literal line you wait to plug into the grid is years long. Years. So when a developer struts in and tells you they are building a massive AI data center, the first question a real allocator asks is not how big is it and not what's the return. First question is show me the power. Show me it's secured. Show me it's permitted. Show me exactly when it gets delivered. Because here's the trap that's going to wreck a lot of people in this cycle. There are deals all over this market right now that are quietly assuming power. They have not actually locked down. And the returns look absolutely gorgeous on the page. And the entire structure is sitting on a power assumption that may not materialize for years, if it ever materializes at all. That enormous risk is almost never priced into the headline number that you get shown. So when you're underwriting anything in this space, power is not a footnote. Power is a detail. Power is the deal. No secured power, no deal. Write that down and tape it to your monitor. 

Ryan Miller

Now, let me give you the second hidden risk. Because if power is the risk nobody prices in, obsolescence is the risk everybody underestimates. The hardware at the heart of AI, the GPUs, the advanced chips, they are on a brutally fast refresh cycle. A new generation comes out and the previous generation loses a meaningful chunk of its value fast. So if you're financing or owning an asset whose value is tightly tied to one specific chip generation, you have to model that refresh cycle against the useful life of the asset and against your own hold period. And here's a sophisticated move that the pros make and the tourists miss. The building, the data center shell, the power infrastructure, the cooling systems, that can have a long, durable, multi-decade life. The silicon inside it might be largely obsolete in just a handful of years. So a real underwriter separates those two cleanly. They get long duration, durable value out of the parts that last, and they refuse to overpay for the parts that go obsolete in a few years. The tourist treats the whole thing as one undifferentiated asset, pays a premium for all of it, and then gets crushed when the chip cycle inevitably turns. Okay. So you understand the stack, you understand the power, and you understand the obsolescence. Now let's get to the part that you can actually act on how a fund manager positions around all of this. And these are the moves I would be making, and these are the moves the smartest allocators I know are making right now, today, while everyone else argues over a bubble.

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

So move number one, anchor on contracted cash flow. The closer you can get to a long-term take or pay or a long-dated off-take contract with an investment grade counterparty on the other side, the more your position behaves like an infrastructure and less like a bet. A signed long-term contract with a creditworthy buyer of that compute or that power, that's your foundation. But spot exposure and speculative future demand is the opposite of a foundation. So you want to look for that. So with every single opportunity, you ask, who is contractually obligated to pay me for how long? And how strong is their credit? If the answer is funny, your returns might be two. 

Ryan Miller

And then there's move number two climb the capital stack for protection. So in a frothy hype-driven theme, seniority is your best friend in the entire world. Seniors secured before mezzanine, mezzanine before preferred, and preferred before common equity. So when you climb the stack, yeah, you give up some of the upside. But you buy yourself protection. You buy yourself a margin of safety. And you buy yourself the ability to sleep at night when the headlines get messy. And in a cycle like this one, the headlines will get messy. In a bubble, the survivors are almost always the people who are sitting higher on the capital stack that we mentioned before than all the gamblers underneath them. So don't be too proud to take the safer senior return on this wave, because it's enormous. The wave is so large that the senior protected returns are still genuinely excellent. Let the ego chase the home run. Let your capital take the protected double over and over and over. 

Ryan Miller

Then there's move number three. Underwrite the power, not just the silicon. I've already hammered this, but it's so important of making it an explicit standalone move. So before you commit a single dollar, you verify the power. Secured, permitted, deliverable on a real contractual timeline. If a sponsor can't show you that, you do not have a deal. You have hope dressed up as a deal. And hope, my friends, is not an underwriting strategy. 

Ryan Miller

Then there's move number four. Stress the obsolescence. Build the chip refresh cycle directly into your diligence model explicitly. And then separate the long life infrastructure from the short life hardware and value them differently. Make absolutely sure your return does not secretly depend on chip generation holding its value longer than it realistically will. And run down the downside case. Not the base case, the downside case. If the deal only works when everything goes perfect, it's not a deal. It's a prayer. And we don't allocate institutional capital to prayer. 

Ryan Miller

Then there's move number five. Pre-wire the exit before you commit, not after. Ask on the way in. How do I get out of this? Is there a clear refinancing path? Can this be securitized down the line? Is there an obvious strategic buyer who would want the exact asset? You want exit visibility on the way in every time. The amateurs fall in love with the entry, ride the excitement, and never once think about the exit until the day they find themselves trapped with no buyer. The professionals know their exit before they ever wire the money. Entry is optional. Exit, that's everything. 

Ryan Miller

Now, I built you something to make every bit of this concrete and usable, and it's waiting for you in the link below this episode, and it's completely free. It's called The AI Infrastructure Capital Scorecard. It takes everything we just walked through over the last few minutes and it turns it into a simple, powerful scoring tool. So there's six dimensions: contracted cash flow durability, counterparty credit quality, your position in the capital stack, the power and interconnected moat, obsolescence resilience, and your exit path. You score the opportunity from zero to ten on each one of the six, and you apply the weightings, and you get a single composite score of 100. 

Ryan Miller

80 to 100, that's infrastructure green. That's anchor worthy. That's the stuff you build a position around. 60 to 79%, that's selective. It's fundable, but you need real structure and protection to do it. 40 to 59, be very careful because that is equity risk wearing infrastructure clothing. Below 40, that's a bubble beta. That is a bet, not an allocation. Walk away, keep your powder dry. That's what I would do, just in my opinion. So print that scorecard out, keep it right next to you, and run the very next AI deal that crosses your desk straight through it. I promise you, you will see that deal in a way that you simply did not see it before. Let me show you how this actually works live because a framework is useless until you run it on a real deal. 

Ryan Miller

So picture this: a sponsor walks into your office and brings you a data center development deal of a lifetime. And the pitch is gorgeous. 16% projected IRR, brand new facility, AI demand through the roof, charts are pointing up and to the right. Now the amateur hears 16% and says yes before the sponsor even finishes the sentence. You don't. You pull out the scorecard, contracted cash flow, you see the offtake. Turns out there's a letter of intent, but no signed long-term contract yet. That's maybe a four out of 10. Not the way they want you to assume. Counterparty Credit, the anchor tenant, is a well-funded startup, not an investment grade hyperscaler. Call it a five. Position it in the capital stick. They're offering you common equity, the riskiest run on the ladder. That's a three. Now there's power, and this is the killer. You ask to see the interconnection agreement, and what you actually get is an application sitting in a queue. A three-year wait and no guarantee at the end of it. That's a two. Obsolescence. The whole model leans on a single current generation chip holding its value, a four. And the exit. No clear refinancing or sale path articulated at all. Another four. You run the weighted math, and that gorgeous 16% deal scores somewhere in the low 40s. That's bubble beta. It is a bet wearing a tailored suit. And you just saved yourself from a loss that the 16% IRR headline would never ever have warned you about. 

Ryan Miller

Now flip it. Same sector, completely different deal. Senior secured debt, an investment grade hyperscaler on a 15-year contract. Powers already secured and energized. Conservative chip assumptions, they're baked in, a clear securitization path on the exit, lower headline return, let's say 9, maybe 10%, but it scores in the high 80s. Infrastructure great. That is the deal you anchor a real position around. The amateur grabbed at the 16% upside and got wiped out. You took the nine and compounded it safely for a decade. And that right there, that is the entire difference between gambling and allocating. And the scorecard is the thing that makes it visible right in front of your eyes. Now, before I let you go, let me name the three biggest mistakes I see fund managers make in this exact space. Just so you can sidestep every single one of them. Mistake number one, confusing a great sector with a great deal. AI is a phenomenal sector. That tells you absolutely nothing whether the specific deal in front of you is any good. The sector is the wave. The deal is the surfboard. And a great wave on a broken board still drowns you. Then there's mistake number two, anchoring on the headline return and never once interrogating the capital stack underneath it. A 16% common equity return and a 10% senior secured return, they're not even the same species of investment. And pretending they're comparable just because the numbers look close is exactly how people blow themselves up. And there's mistake number three, the most expensive one of them all. Taking the sponsor's power claim at face value. I'm telling you this right now. This single mistake is going to cost the market billions of dollars over the next few years. Verify the power yourself, independently, every single time. The sponsor who is trying to raise money from you is the last person on earth whose power timeline you should ever trust on good faith. 

Ryan Miller

Now, let me address the pushback head on because I already know that some of you are thinking, Ryan, if all of this is obvious, why isn't everyone already doing it? Well, I think there's two reasons. The first is access. A lot of the best positions in this entire stack, the senior secure debt, the private credit, the real assets, they're simply not available on your broker job. They live in private markets, they live inside funds, inside relationships, inside rooms that most people never get into in their entire life. And I want you to understand that that is a feature, not a book. That difficulty of access is precisely why returns stay durable. If everybody could click a button and buy it, the edge would be competed away within a week. The second reason is that it's just not sexy. I bought a senior secure debt against contracted data center with locked-in power. That does not get you an applause at a dinner party the way I'm an early investor in the hottest AI startup does. But here's the thing: you're not here for an applause at a dinner party. You're here to build real, durable, generational capital for yourself, for your family, and for your investors. And the boring, contracted asset back money is exactly how the wealthiest, most sophisticated allocators on planet Earth actually play waves like this one. So let everybody else chase the lottery ticket and brag about it. You go quietly, take the picks, the shovels, and the water right, and you let the results do the talking themselves. 

Ryan Miller

So let's bring this all the way home. AI is a bubble. At the top and underneath that bubble sits one of the greatest infrastructure opportunities of our entire lifetime. The difference between getting wiped out and getting wealthy in this era comes down to four things. Knowing the capital stack, respecting the power constraint, modeling the obsolescence honestly, and positioning with cold discipline instead of hot hype. That's it. Don't buy the bubble. Underwrite the build out. That is the whole game. And now you know how to play it. So if you got value out of this, and I genuinely hope you got a ton, go grab The AI infrastructure Capital Scorecard in the link below. It's free. And it will change how you look at every single one of these deals from here on out. And if you're a fund manager who's serious about raising capital and deploying capital at the highest level, surrounded by other managers from around the world who are doing exactly the same thing, that is precisely what we have built inside of The Fund Raise Capital Community. You'll get an education on how to identify, research, and approach institutional capital sources. You'll get the capital raising frameworks and the playbooks. You'll get a global network of fund managers raising at the highest levels and weekly live education sessions with me. If that sounds like you, go apply to go.fundraisecapital.co/apply. You do these things and you too will be well on your way in your pursuit  of Making Billions.