Episode 7 After the AI Bubble Part 2

The conversation connects infrastructure-led booms to modern capital cycles, moving from railroads and the Great Financial Crisis mechanics to crypto’s speculative dynamics and real-world payment alternatives. From there, attention shifts to AI business models, the semiconductor supply chain, and how valuation assumptions collide with depreciation, competition, and energy constraints. The episode closes with a wider lens on power generation, nuclear perceptions, and the difference between workable technology and workable stories.

Episode Show Notes

• Context recap and why this topic continues into Part 2.
• Railroads as a “production” bubble example tied to promises, land, and financing dynamics.
• Mortgage-backed securities explained: liabilities vs assets, pooling mortgages, and why defaults matter.
• Banking constraints: insolvency rules, regulator intervention (FDIC), and why payment systems drive bailouts.
• Crypto framed as speculative and policy-supported; Bitcoin ownership distribution discussed (including a cited figure from Reddit).
• Practical payment systems comparison: M-Pesa and the role of private ownership vs decentralization.
• AI “build-out” mechanics, cashflow focus, and who benefits most from the current spend (Nvidia; plus TSMC and ASML).
• Semiconductor supply-chain concentration and geopolitical framing (ASML in the Netherlands).
• Oracle / OpenAI / Nvidia circular flow example and a “clearing” explanation using a hotel-debt story.
• Michael Burry referenced in relation to the AI bubble; Substack mentioned by name approximately (“Cassandra Unleashed”).
• Depreciation vs useful life of AI hardware compared to “car loan” dynamics on balance sheets.
• Nvidia valuation/PE ratio discussed with numbers mentioned in the conversation.
• Creative destruction referenced via Joseph Schumpeter; competition and lower-cost alternatives as a possible bubble catalyst.
• AI tooling in day-to-day work: ChatGPT, Perplexity, and Claude mentioned as paid tools/subscriptions.
• Energy constraints and nuclear power as a limiting factor for AI expansion; nuclear safety perception discussed.
• Rory Sutherland mentioned in the context of reframing nuclear as “elemental energy.”
• “Enron Egg” viral hoax discussed; Project Pele referenced as a US military project.

Episode Timestamps

00:00:00 Opening and Part 2 setup
00:02:05 Railroads, land grants, and “production” bubble framing
00:10:43 Mortgage-backed securities: how they work and why pooling changes risk
00:20:42 How financial impacts spread globally through investors and pensions
00:22:28 Michael Burry reference and Substack mention
00:23:02 Banking rules, capital requirements, and why insolvency is different for banks
00:25:54 Payment systems, bailouts, and the FDIC as regulator example
00:26:16 Crypto defined as speculative; government/policy support referenced
00:33:32 M-Pesa as a practical payment comparison point
00:39:09 AI business model pressure and who captures cashflow (Nvidia/TSMC/ASML)
00:41:11 Oracle/OpenAI/Nvidia money flow example and “clearing” explanation
00:43:38 Michael Burry framing and depreciation vs useful life discussion
00:46:32 Nvidia PE ratio discussion (numbers referenced)
00:47:21 Schumpeter and creative destruction; competition as a possible turning point
00:50:24 Summary: why it’s viewed as a bubble; AI adoption and tools mentioned
00:53:41 Energy constraints and nuclear power as part of the continuation of the AI boom
00:55:06 Rory Sutherland reference; “Enron Egg” and Project Pele

Entities

People

Emanuel
Kevin
Sam Bankman-Fried
Michael Burry
Christian Bale
Rory Sutherland
Tyrone Keynes

Companies, products, platforms, and organizations

OpenAI / “Open AI”
Google
Amazon
Meta
Nvidia
TSMC
ASML
Oracle
FDIC
Substack (referenced as Michael Burry’s Substack)
M-Pesa
Western Union
Vodafone / “Vodafones”
Rogers
Reddit
Enron / “Enron Egg”
Project Pele
US military / US Department of Defense
ChatGPT
Perplexity
Claude

Places and events (named)

United States / US
Canada
England
Netherlands
Ukraine
Africa
China
Chernobyl
Fukushima
Idaho

Crypto assets (named)

Bitcoin
Ethereum
Solana

About the Podcast

Hosted by Kevin Carney and Emanuel Petrescu, two curious minds exploring ideas, culture, and everything in between. Curious Pundits is a conversational podcast where each episode starts with a topic that caught our attention and unfolds into thoughtful, unscripted discussion. We follow curiosity wherever it leads, across disciplines, opinions, and perspectives, without pretending to have all the answers.

Listen to the Curious Pundits Podcast on Your Favorite App

Website: https://curiouspundits.com/
YouTube: https://www.youtube.com/@curiouspundits
Podbean: https://curiouspundits.podbean.com
Apple: https://podcasts.apple.com/us/podcast/curious-pundits/id1874614249
Spotify: https://open.spotify.com/show/4nvSoNRrgPPBkdZiLGanYh
Amazon: https://music.amazon.com/podcasts/26ab96d9-ab37-4369-bccc-fe2cf937f950
iHeartRadio: https://iheart.com/podcast/321051634
Soundcloud: https://soundcloud.com/curiouspundits

Follow the Curious Pundits on Social Media

Facebook: https://www.facebook.com/curiouspundits/
Instagram: https://www.instagram.com/curiouspundits/
Threads: https://www.threads.com/@curiouspundits
YouTube: https://www.youtube.com/@curiouspundits
LinkedIn: https://linkedin.com/company/curiouspundits
TikTok: https://www.tiktok.com/@curiouspundits
X: https://x.com/CuriousPundits
Pinterest: https://www.pinterest.com/curiouspundits/
Reddit: https://www.reddit.com/user/curiouspundits/
Medium: https://medium.com/@curiouspundits
Tumblr: https://www.tumblr.com/blog/curiouspundits

Transcript

[00:00:00] Speaker: This is the Curious Pundits Podcast hosted by Kevin and Emmanuel. They explore bits of everything through thoughtful conversations, prioritizing curiosity over conclusions, and they thank you for joining them now onto the podcast.

[00:00:18] Emanuel: Hi everyone, and welcome to part two of a very interesting episode that required it as it turned out, two parts because it was basically too much. It’s called after the AI bubble, but we didn’t even get to the 1900s yet.

[00:00:36] Kevin: Yeah, we’re still doing the background information.

[00:00:38] Emanuel: Yeah. My name is Emanuel.

[00:00:40] Kevin: My name is Kevin.

[00:00:41] Emanuel: And together we’re the Curious Pundits, your hosts for the curiouspundits.com podcast, Bits of Everything.

We’re just two people who wants to share what we know and learn from one another and from everyone essentially. Kevin, I’ve didn’t interrupted you that much, as much as I would want.

In the first part, you mentioned the tulip bubble. The… a couple other bubbles. I didn’t want to make a note, and maybe you, it’s on your list or not, but I don’t want to make a note that roughly six or eight years ago when people were talking about Bitcoin, they were making a reference to the tulip bubble, the Dutch bubble from the 1600s as well.

That’s topic for another conversation, but that’s…

[00:01:39] Kevin: Crypto is on the list. I’m going to talk about it a little bit just before we get to AI. It is a classic bubble. I know people push back on that idea, but it is.

[00:01:52] Emanuel: Okay.

I don’t remember where were you, I’m just observing this one, but it’s very interesting because this is new information to me as well.

Where were you?

[00:02:05] Kevin: We were talking about how during the American War of Independence there was a difference of opinion between the English and the Americans as to how available Indian territory was. Many people who fought on the American side were enticed with promises of land that was in Indian territory.

So the basic idea is if we win the war we’re going to go take land from the Indians and we’re going to give some of it to you. And the whole viability of that concept rested on the idea… we were still predominantly a farming community at that time, and if farmers are going to move westward, they need a way to get their products to market.

The government and the power brokers, the bankers, the government, the captains of industry are like, don’t worry about it. We’re going to build the railroad. It will transport your goods hundreds of miles, and your market will be enormous. It’s going to be great.

And these guys are like, okay.

Now we weren’t yet in the era of government bailouts, so the government… portion of the railways is that the government, and this applies to this day, these land grant decisions that were made back in the early part of the 19th century actually determine the maps of land ownership of the American West to this day. They created a massive checkerboard pattern of 160 acre plots, and the government basically claimed ownership of all of them.

Then they sold every other one to someone.

And that money, I forget the exact details, it’s been a long time since I’ve read the details, but that money wound up in the bank accounts of the railroad companies and they used that money to buy the ties and buy the steel and pay the laborers and all that kind of stuff.

But it was the classic. So I’m going to introduce a new term and basically it’s one that I’m making up. This is a production event bubble. And what I mean is that somebody’s trying to make something. They’re trying to build something. In this case, it was basic infrastructure. They’re trying to build the railroad so that they can keep that promise to the farmers.

The railroads didn’t make enough money to cover the payments they had to make on their loans, so they kept going bankrupt. Now, this was prior to a proper government bailout, such as we’ve done at large scale since 2008.

So in the, so these companies..

[00:05:00] Emanuel: In the US.

[00:05:02] Kevin: All over the world.

Canada is an interesting exception to the big government bailouts. I know this is a segue, and I won’t go too far into it, but back when everything failed… and banking systems failed, and central banks were bailing them out, people were saying… Canada did the right thing by not bailing out the banking system.

But the Canadian banking system didn’t fail. Like it didn’t need a bailout. That’s why it wasn’t bailed out.

Canadian banking is incredibly boring and there hasn’t been a systemic failure in the Canadian banking system for 150 years.

[00:05:45] Emanuel: As it should be for a banking system. As it should be.

[00:05:49] Kevin: Yeah. Yeah That was a quick segue. Let’s get back on.

So basically the railroad companies are going bankrupt.

Bankruptcy courts are managing the unwinding of them, and that involved selling assets to whoever the highest bidder was. So other people are coming in and buying those assets at pennies on the dollar and picking up where the prior guy left off.

And they just kept doing this until eventually, after some number of decades, railroads had enough traffic to generate enough revenue in order to make the payments on the loans.

In the meantime, the government had a vested interest in keeping the railroad story going because they had made promises to these farmers. So the mechanism of the bailout was not big government injecting cash into railroad companies. It was overseeing these bankruptcy processes and arranging for new investors to come in and take over.

But it was a too big to fail, or probably more technically correct, a too strategically important to fail enterprise. The railroad just had to continue.

So it was a speculative bubble. The railroad companies are selling stock, stockholders are buying them. The stockholders are losing their money when the railroad companies go bankrupt.

And they just kept doing that for as long as it took to get it going.

So the next one of interest, I’m going to actually jump like decades into the future.

The next one of interest to me is the dot com bailout in the 1990s. And the reason the dot com bubble is important to me or important is not the right word, but interesting to me is because in hindsight, the fact that shareholders were buying shares in companies that had no viable future, seems obvious to us today.

People were like… raising venture capital money on the basis of having secured a domain name of dogs.com. Like we have a name dogs.com, we can’t fail. And then they’re giving like free shipping on 20 pound bags of dog food, right? They’re so like certain basic fundamentals that aren’t going to work here.

So to me, that one is so interesting because in hindsight it was such an obviously stupid thing to do. And I view this one as not so much a production oriented bubble. Yeah, these companies were all going to sell stuff and make stuff, but it was predominantly financial engineering.

Like amazon.com, one of the ones that worked. They’re not doing anything we haven’t done before. They’re just selling books in a new and very different and much cheaper way. And Amazon also lost a ton of money for the longest time, and it took them years before they became cashflow positive. And in the meantime, investors just kept pumping in money.

To say that something is a bubble doesn’t predestine failure necessarily. Something can have all the characteristics of a bubble, and then for some reason like, not burst at the last minute.

And I’m going to mention crypto in a little bit because crypto is in that state right now.

So the next one in my mind that is of interest is what I call the dark fiber bubble. And it’s somewhat ancillary to the dot com bubble. So all of these telecoms are raising money from shareholders and spending a fortune, laying fiber all over the country.

And this was like the railroad one because when a company would go bust and their assets would be sold at auction. Whoever showed up to buy the auctions picked up where they left off. And now you and I are actually talking right now on what used to be dark fiber that probably sat dark for years, if not a few decades before it became economically viable to light it up to expand Internet services in various areas or to provide more services.

And quite literally the dominance of Netflix and Amazon Prime streaming and all kind stuff like without the dark fiber infrastructure bubble busts, like none of that would’ve happened.

We needed that dark fiber to be there so that we could light it up later when it became economically viable.

[00:10:41] Emanuel: Set the ground.

[00:10:43] Kevin: Yeah.

So the next one I’m going to mention, I’m not going to go into an extreme detail because this was… most… with the exception of the dot com, and the dot com had a quasi production value to it… there weren’t so much companies who were like producing new widgets. People were looking at innovative way to sell existing widgets, but maybe it’s a hybrid.

It’s not a pure production bubble, but it’s also not a pure financial engineering bubble.

Mortgage-backed securities… pure financial engineering, like we weren’t actually producing anything. So I’m going to…

[00:11:28] Emanuel: Explain in detail here. You said you wouldn’t go into much detail, but I would actually ask you to go into details about mortgage, what it means for, again, many people around the world might actually don’t know.

So…

[00:11:45] Kevin: this could take a few minutes if… so fundamentally, you take out a mortgage and to you that mortgage is a liability because you’ve got to make payments every month.

So the person you’re making payments to… that mortgage is an asset because they receive the payments. And this framing of whose liability is it and whose asset is it, matters, right?

I don’t know how common this is in Canada, but here in the United States, when you go to a bank and apply for a mortgage or a mortgage broker who front ends multiple banks, and you take out a mortgage, that mortgage is almost immediately sold to some other mortgage processor. So you take out a mortgage with Bank A, you make your first payment to bank A, and then you immediately get a letter from Bank B saying, we are now your mortgage provider, from now on, send your payments to us.

So who’s buying these mortgages and why? Investors are buying mortgages because people are on the hook to make monthly payments. So the mortgages are purchased at a discount. Let’s say the, and I’m going to use round numbers. Let’s say the value of the mortgage is a hundred thousand dollars.

Bank A, who originated the mortgage, might sell that mortgage to bank B for… I don’t know what the exact numbers are, but $98,000, $95,000.

So the company who buys the mortgage, you as the payee, you’re still on the hook for the full a hundred thousand dollars. They bought it at a discount, but that discount becomes a part of their yield, right?

So as an investor, if you buy a mortgage and then that person defaults on the mortgage, the value of your asset immediately drops to zero.

So a mortgage backed security is where somebody acquires a thousand mortgages, lumps them together into a pool, and then chops them up into shares or securities, and then they sell those securities to investors.

So the idea is, as an investor, you’re protected.

Because we know in aggregate, out of every thousand mortgages, two of them are going to default. So now your risk of default or the devaluation of your investment due to defaults is guaranteed to be 0.2%, but it’s guaranteed not to be any more than 0.2% because you’ve aggregated and bought into a pool of mortgages.

Whereas if you buy individual mortgages, there’s this losing the lottery thing that can happen where you bought a mortgage from someone who can’t make the payments and that’s a big problem for you.

So that’s the whole concept behind mortgage-backed securities. And essentially what happened in the early two thousands is… Everyone knows that people pay their mortgages, right? It’s not just a speculative financial asset, it’s where you live. So if you’re short of money, you might skimp on gym memberships, you might skimp on…

[00:14:58] Emanuel: A meal or two.

[00:14:59] Kevin: A meal or two… but you’ve got to have some place to lay your head every night. And if you’re not paying your mortgage, you’re going to have to rent someplace and you’re paying anyway, so people pay their mortgages.

So these were considered to be very safe investments for the people buying the mortgage backed securities, and a result of being safe…

[00:15:16] Emanuel: Very boring. They were also very boring, and somebody said…

[00:15:20] Kevin: Oh no, they got exciting pretty quickly.

[00:15:22] Emanuel: Somebody said, let’s animate them a little bit.

[00:15:27] Kevin: What happened was, because they were so safe, people wanted more of them, and literally the mortgage industry ran out of mortgages.

It’s like we don’t have any more mortgages to like bundle together and chop up into tranches as they call them, because we’re like out of mortgages.

But the people who wanted to buy them still wanted to buy them.

So then we started making bad mortgages. And by we, I don’t mean you and me… banks are now making… offering mortgages to people who… they’re going to default.

You just look at the paperwork, you look at their income and expenses and their cash flow profile and you know they’re going to default, but you don’t care because as soon as that mortgage closes, you’re going to sell it.

I actually didn’t mean to go into this much detail, but this is a, classic example of what I call… losing the peg, and this is a phrase I made up and I’m sure there’s a proper technical term, but there are things in financial markets whose value is in theory dependent upon the value of something else.

So the value of a mortgage-backed security is dependent upon are the underlying mortgages being paid, right? And I say losing the peg because… the companies who are receiving the mortgage payments, they know that a mortgage payment did come in, a partial mortgage payment came in, or no mortgage payment came in.

They know what the balance on the mortgage is, what payments are and are not late. They have all that detail, but when the mortgages got bundled and chopped up, that level of detail never flowed through.

So the credit rating agencies are saying… hey, mortgages are safe, we’re going to give these AAA rating.

Now I met a woman who… other people consider to be like the queen of credit rating and she’s pushing what I think is a brilliant idea, and I can’t believe the financial industry isn’t going for it. So I’m going to use… I’m going to, reason by analogy here. Sometimes I buy something on an online website that gets delivered by Federal Express. When that happens on the day that the item is being delivered, in a certain time window, I can pull up a map and I can see where the truck is, and it blows me away that this level of detail is considered something that customers want.

Like I don’t really care if my thing arrives at noon or at two o’clock, but I do this just because I’m like impressed with the fact that I can. So I pull up the window, I watch the little truck drive around town, and I get a sense that it might be here in an hour.

The technology, the computer and computer networking technology that allows that to happen means that allowing the companies who are or are not receiving the mortgage payments, they’re updating that in their computer systems, allowing the companies who are pricing and selling mortgage-backed securities to know what mortgages are and are not being paid, which in theory… this mortgage-backed security no longer has a letter grade, it has a number grade, and that number is what percentage of underlying mortgages got paid in full, for example.

So she has a phrase for this, she called it continuous monitoring. And I do not know why we don’t do this, but we don’t.

So anyway, we lost the peg. And as if you’ve seen the movie The Big Short…

[00:19:13] Emanuel: I Was About to…

[00:19:14] Kevin: Very few people noticed and the very few people who did notice found ways to take advantage of this and made an enormous amount of money.

But to, and this is, I didn’t really mean to go into this much detail, but to… to give you a feeling for what I’m going to describe as the fragility of the financial system, I’m going to give you a pop quiz.

In aggregate… in the American housing market… just take a wild guess. What percent… at the peak of mortgage defaults, what percentage of mortgages were defaulted on?

[00:19:53] Emanuel: Oh, I think very little. Very, just a few. I know. 2% even less.

[00:19:58] Kevin: No, actually you’re on the low side. Most people say God, it must have been like half or something. The answer is 12%.

[00:20:07] Emanuel: 12%. Okay.

[00:20:08] Kevin: Yeah. And a 12% default rate of mortgages in the United States brought the entire global financial markets almost to a grinding halt.

[00:20:20] Emanuel: Almost it’s because you’ve experienced that period here in North America, you’ve probably seen the apocalypse and survived the apocalypse, but I promise you there’s many places around the world that didn’t feel… weren’t affected at all by the Great Depression or the great financial crisis that would all log into other sectors as well.

[00:20:42] Kevin: These were being purchased by investors all over the world. So that’s how the impacts of the issue spread. Pension funds in England… pension funds are a big buyer of these kinds of securities, so all over the world French teachers lost money in their pension fund. And it was it was a huge problem.

And the primary reason… so the rules of banking are, in many ways, they’re weird. I don’t want to go too deep into this, but banking is a backwards business.

So ordinarily a business covers their liabilities with their assets. So you have some assets and if there’s a liability you need to cover, you can sell an asset, raise some cash and pay that bill.

To a bank their loan portfolio, the people paying them money, that’s their assets and their depositor base to them are liabilities because any depositor can come in and take their money out at any time. And yet in the language of banking, we talk about deposits funding loans.

In a weird way, linguistically, we talk about the need for liabilities to cover assets. And that’s why I say banking… when we talk about banking, like it’s like going through the looking glass and everything’s backwards. But anyway, banks are not required. You have a question?

[00:22:20] Emanuel: Another question about the pullback to the main topic. One of, because it’s relevant to, to what we were initially discussing. Oh, yeah.

[00:22:26] Kevin: I’m going to get right back to it. I was actually just…

[00:22:28] Emanuel: One of the people that took advantage of the… this opportunity in 2008 was Michael Burry.

[00:22:38] Kevin: Yeah.

[00:22:38] Emanuel: Portrayed by Christian Bale.

[00:22:40] Kevin: Yeah.

[00:22:40] Emanuel: And he also did something recently regarding to the AI bubble as well.

[00:22:45] Kevin: Yes he did. I now subscribe to his Substack. It’s called something like Cassandra Unleashed or something like that.

[00:22:53] Emanuel: I saw it. Yeah. I didn’t.

[00:22:57] Kevin: So anyway…

[00:22:59] Emanuel: 2008, the…

[00:23:02] Kevin: Yeah, so one of the interesting rules of banking, so a regular business, you’re a coffee shop, a bookstore, whatever, you’re allowed to be in a situation where you’re cashflow positive, you have enough cash coming in to cover your bills, but your assets in aggregate are less than your liabilities.

So you have negative equity, but you’re paying your bills and that’s okay. Maybe you can dig yourself out of it.

Banks are not allowed to do that. Banks have to maintain a capital requirement, which is where reserves come in. They’re a part of that, right? And a bank is not allowed to have negative equity.

In fact, a bank is not allowed to have too little equity when banks are at risk of becoming… So if your liabilities exceed your assets, you are insolvent.

And banks are not allowed, definitely not allowed to be insolvent, but they’re also not allowed to be close to insolvent. And if they get there, a banking regulator like the FDIC, literally takes them over and unwinds that bank.

So what happened during the great financial crisis is all of these loan assets. Sitting out there in the world, which are mortgage-backed securities are loan assets. You own it, it’s an asset. You’re going to get a stream of income. And banks had to basically write them down. The value of this one is not a thousand dollars, it’s 50 bucks or whatever.

And those are assets of the bank. And as they’re like writing down the value of these assets, they’re getting closer and closer to the insolvency threshold. And if they hit that threshold. A banking regulator is going to step in and do something. And it actually got to the point where there’s a couple of… I wasn’t in these phone calls, but rumor has it, there’s a couple of like senior guys at banks talking to people at Central Banks about, we are desperate, we need help, we’re headed towards insolvency.

And the central banker’s how much time do you have? And the guy’s about two and a half hours. And this is a particular nuance of the rules of banking, you’re not allowed to be insolvent even for 10 seconds. It’s just not allowed. Exactly. So the central banks I mentioned I think in the first episode that the banks do two primary things.

They are in the business of making loans. That’s one of the primary ways that they make money, but they also handle payments. If banks are being shut down, that’s going to impair our ability to like… buy gas, buy groceries.

Like maybe there’s a two day period where the gas station doesn’t have a bank because they got shut down and they haven’t been like, set up with somebody else yet or something…

[00:25:54] Emanuel: Happens…

[00:25:54] Kevin: These sound like farfetched scenarios, but they’re not impossible. And when a bank is at risk of becoming insolvent, a lot of activity happens on the back end to make sure that they don’t happen. And the reason banks have got to be bailed out is because we can’t have the payment system not working.

Right.

[00:26:16] Emanuel: I do have a question.

Is it an AI bubble?

[00:26:20] Kevin: Oh, totally. Totally. So…

[00:26:23] Emanuel: That’s me pulling you back.

[00:26:25] Kevin: Yeah, I know I have this habit. I go on and on. But before we get to AI, let’s talk very briefly about crypto.

The characteristic of a bubble, the reason you buy these things is because you expect these assets to either generate some income or appreciate in value.

What makes crypto interesting to me is that it doesn’t generate income, like there’s no income. You don’t rent them out, there’s no dividends, there’s no interest. It’s purely you buy it so that it will appreciate in price and you’ll have more money later. Like that basically is the crypto markets. So crypto to me is a very interesting, and I’m going to say it’s a speculative bubble because it fits all the criteria, but crypto is heavily supported by governments… in the US, the Genius Act.

When the government is putting money into something that makes more money available to buy it, and that keeps the price up. So I firmly believe crypto is in a bubble and it’s a bubble that’s being supported by… mostly central banks, but government policies, shall we say.

[00:27:42] Emanuel: Interesting. Your opinion. I happen to know a thing or two about this in industry and technology and ecosystem altogether, right? Because we have the blockchain technology that kind of like where the crypto, the currency itself has spanned from the ecosystem has evolved and has matured enough to…

[00:28:06] Kevin: Oh yeah.

[00:28:07] Emanuel: Create all those things that you mentioned earlier. Maybe in 2018, 2016, we didn’t, couldn’t have this conversation, but it has matured.

There’s things that you can actually do with it all the… technology people that I’ve talked to, mostly people in IT understand, and they mentioned that the future belongs to the blockchain technology, the internet of things, and tokens, which essentially are the currency with those platforms, it belongs to that inevitably.

So everybody agrees with that. What I might argue is that, obviously government see it differently. They try to take control, they try to take ownership of some things. What you call crypto right now, crypto, there’s many cryptos out there right now. We’re talking simply about Bitcoin. That’s still sovereign when it comes to crypto.

It’s still… has its own that it’s not controlled by anyone. It’s been going up and down speculative for the past years or so. Right now it’s at year low. If I remember correctly, we’re recording this at the end of 2025, but AI bubble because of what you said. Yes, for most tokens. But for Bitcoin and perhaps one of one or two others.

One could be Ethereum, could be Solana, could be a few others, and perhaps even some state issued ones, I don’t think they’re in necessarily fit the description that you just made about the crypto. They could theoretically fit in the description that you have about the bubble.

[00:30:04] Kevin: Okay, now I’m going to take an opportunity to explain why it doesn’t exactly fit my description of a bubble.

[00:30:12] Emanuel: Please.

[00:30:12] Kevin: In short, government support. There’s a guy named Tyrone Keynes who’s really good at modeling complex dynamic systems in terms of money flows, and I watched him display a model once where he showed that the price of Bitcoin goes up and down with the amount of bank created credit money in the economy.

Now, this is a correlation. It’s not a causation. But still the correlation is really tight and it looks like what is supporting the purchases of Bitcoin and therefore the price of Bitcoin, is the fact that people are taking out loans to buy Bitcoin.

[00:30:51] Emanuel: Again, the US government support is what you’re referring to?

[00:30:55] Kevin: Yes.

[00:30:56] Emanuel: Okay. US represents 400 million, 300 something million. While the rest of the world consists of a few billions more. What many people don’t understand, and I think I said this in a previous episode, if not I’m going to reiterate, is many things that exist… US, Canada, they probably don’t need. Bitcoin don’t, they don’t need crypto.

But other places around the world actually need. They’re unbanked. And also think of, and I know people from Venezuela for example, you know the situation right there, right now. And for the past couple years. Wasn’t exactly the best, one, right? People lost almost everything and they had to go to Columbia.

Let’s say… your next door neighbor, or people from Ukraine right now who need to flee the conflict to somewhere in Europe. Usually you either don’t have any kind of assets, don’t have a bank account, or they, get frozen. You can’t access your fund. I cannot tell you how… it’s a life and or death situation of being, having even something like $2,000 and being able to carry with you in your new location.

Know from Venezuela to Colombia, if you carry that in cash, in gold or in anything else, they’ll probably kill you and get them out of you. But if you carry the value in crypto, which is essentially, you can memorize essentially the wallet addresses. When you go to that destination and actually have access to those funds, it can be life and death.

Somebody said in a podcast, I think was a fighter, a professional UFC fighter said that having millions of dollars doesn’t make the difference. Having two, $3,000 at the right time makes a difference and that’s what crypto has allowed people to do.

[00:32:57] Kevin: I agree with what you’ve said, but Bitcoin is not currently being used for that.

It has the potential to…

[00:33:04] Emanuel: No it has been used and it’s used and has been used. And even in Canada…

[00:33:08] Kevin: Not in large scale.

If you… own something that is a highly volatile asset and it’s liable to be worth 20% more in two months or twice as much in two years. You’re not going to spend it on something because the thing you’re buying is now more expensive.

Your Bitcoin…

[00:33:29] Emanuel: If you’re in the US, you don’t spend it. Yes.

[00:33:32] Kevin: I want to tell you about something. In my other monitor, I Googled for something to get the name right, but there’s a payment system that’s being used in various countries in Africa called M Pesa. Have you ever heard of it?

[00:33:43] Emanuel: M Pesa, yeah.

[00:33:44] Kevin: Yeah. So that’s not Bitcoin, but it serves the mechanism and the needs of what you were describing.

[00:33:53] Emanuel: The difference between Bitcoin and the blockchain technology is that it’s decentralized. M Pesa is owned by somebody, if I’m not mistaken. Could be… not Rogers could be one of Vodafones or one of the tele companies.

[00:34:05] Kevin: Oh yeah. It very well could be… I don’t know.

[00:34:07] Emanuel: Yeah, so that’s the difference essentially because they own it. It’s essentially… it’s Western Union… It’s a Western Union. They took over, actually they bought some places on Western Union. So it’s owned, by a private company.

Whereas crypto and the blockchain technology is not owned by somebody.

It’s the network in itself. That’s probably a topic for another episode as well. We, I’m happy that we come up with so many ideas, at least five or six just from this session.

[00:34:36] Kevin: But just very briefly and I do want to get back to the AI bubble.

I’m going to actually Google for how many, because the distribution of Bitcoin ownership is highly centralized.

[00:34:49] Emanuel: Yeah, probably.

I know for sure Asia has, China for sure, has a significant chunk.

[00:34:56] Kevin: No in terms of the number of people. In fact, I’m going to ask a more specific questions. How many… what percentage of humans own Bitcoin?

[00:35:10] Emanuel: That’s a good question.

[00:35:13] Kevin: And it was surprisingly small. 1.3% of the world’s population owns Bitcoin that’s 106 million people according to somebody on Reddit. He could be wrong, but it’s, probably, that’s probably not extremely far off.

[00:35:29] Emanuel: Within the parameters. Yeah.

[00:35:31] Kevin: And if we go back to M Pesa, M Pesa has three times as many users just in Africa.

So the argument I’d like to make here is one of practicality. Like I’m a big believer in the future value of Bitcoin technologies. And I know that not in one of the Curious Pundits podcasts, but somebody else, I used the analogy of the invention of electricity. Yes. Like electricity was invented in the early to mid 19th century, but it didn’t change the world until the 20th century.

And I have a feeling that blockchain is going to be something similar. It’s an invention which is ahead of its time, and there will be a time in the future when it will be critical for all kinds of things. We’re just not there yet. And currently we’re using blockchain for Bitcoin and cryptocurrencies, whereas… whoever these M Pesa of people are, they’re using cell phone minutes as currency and it’s working, right.

But anyway, back to… so in my mind, crypto is a classic bubble, but it’s a bubble like the railroads with strong government support. The difference being that when a… when somebody… I’m looking for the right analogy.

So far they’re… no, that’s not true… some crypto companies have gone bust. Sam Bankman-Fried is in prison for fraud, right? So when one of those crypto companies…

[00:37:02] Emanuel: He managed digital assets, so he managed digital assets. So people gave him their Bitcoins. Essentially, that’s what he did.

[00:37:12] Kevin: Yeah yeah. And most of them like cease to exist in various ways, right? Yeah. I don’t have all the details of how it happened, but when they ceased to exist and when the investors who own them lost all their money, there was no infrastructure left to come in and buy at pennies on the dollar.

That’s what makes cryptocurrency different from like dark fiber or railroads. There is no infrastructure that somebody can buy cheap and then pick up where they left.

[00:37:40] Emanuel: No there is… there is an argument to make to that as well, but let’s cover that in a different different episode because I know for sure many people don’t know, and I’m not aware of this, but there are ways, there are mechanism to.

Get that because… he controls their digital assets, but their digital assets can be, let’s cover that in a different episode. I’m really adamant on coming back to the AI bubble, because I’m curious to know.

[00:38:09] Kevin: Okay, so now we’ll segue straight into AI. So AI in my opinion, is like railroads. Is like canals, right?

We are spending enormous amounts of money on infrastructure and it seems pretty clear to me that without significant government support, OpenAI isn’t going to make it.

Like they’re burning $700,000 a day. They’re announcing these billions and billions of dollars of future expense commitments, and they don’t have enough money coming in to cover it.

However, they’re building real infrastructure in terms of data centers. So in the event that OpenAI ever finds itself in a default situation where they’re insolvent and they’re being wound down in bankruptcy court, we will see that other investors come in to bid on the assets.

That technology won’t go away, but it’ll have a different company’s logo on it later.

[00:39:09] Emanuel: Google.

[00:39:11] Kevin: It could be Google. It could be Amazon. Who knows?

[00:39:13] Emanuel: Could be a brand new one.

[00:39:14] Kevin: That could be, it’s got to be somebody who’s rich.

Now google, Amazon, Meta, they don’t have a problem because these companies are so damn rich they can afford to lose big on their AI ventures because their other ventures are just so incredibly profitable.

But in terms of the pure mechanics of AI… the business model isn’t there, like the only AI company that is actually cashflow positive and making money is Nvidia, and I guess you can spill that over a little onto TSMC and ASML.

[00:39:51] Emanuel: Do you consider them a hardware company, a production company?

[00:39:56] Kevin: Yeah, they’re definitely a production company. They don’t make the chips themselves, but they design the chips, they license the chips. The chips are built at a TSMC FAB using the machines at ASML. So…

[00:40:09] Emanuel: Important to know that there’s just one company that does for everyone essentially.

[00:40:16] Kevin: Are you talking about ASML?

[00:40:18] Emanuel: Yes.

[00:40:18] Kevin: Yeah. I don’t get that. And I know this is slightly off topic, but there was…. at some time in the past, I don’t know the name of the company, I just heard about this in the last month, but there was an American ASML competitor and at some point they decided it like didn’t make sense.

So now the entire world’s ability to produce semiconductors is dependent on this one company who fortunately is on our side, they’re headquartered in the Netherlands. So if they were headquartered in Ukraine, that might feel risky but they’re on our side of the quote unquote what’s left of the iron curtain.

So we feel safe about that. But still, it’s like just one company.

[00:41:05] Emanuel: Put your money in a single basket, or all your eggs in a single basket.

[00:41:10] Kevin: Yeah.

[00:41:11] Emanuel: I might have also, because I’m the devil’s advocate right now, I might have a different different opinion. I’m not saying it’s not a bubble, but, and they’re not doing financial tricks right now because I was just reading the other day about Oracle committing to build for Open AI, these warehouses with money that they don’t have, but Open AI will pay them with the money that they’ll get from Nvidia…

[00:41:37] Kevin: Right, yeah.

[00:41:38] Emanuel: They’ll offer you some of that money to buy chips from Nvidia. Oracle will buy the chips from Nvidia to put in those data centers for Open AI and so forth. They’re changing the money from one to another. And you know how, that joke goes, right?

When was about the bankruptcy, and for those that don’t know… it was this guy at the hotel and he has a visitor, and the visitor said, can I check the rooms to see if I like them, if I could say here. Sure, but you’ll need to give me a hundred bucks deposit. Okay? Here’s the a hundred bucks.

He goes in, checks his room. In the meantime, he calls the cook. The chef here’s that a hundred bucks that I owe you. Okay? He takes it away. The chef calls the housekeeper, the cleaning lady. Here’s the a hundred bucks that I owe you. And the cleaning lady goes back to the hotel owner.

Here’s that a hundred bucks that I owe you. And you have the end of the day, the prospect, the customer comes and says, I don’t like this hotel. Take his hundred bucks with him and goes away.

This is what’s happening right now with the AI… with these AI companies or AI dependent companies, essentially.

Because they’re not even AI to the true sense of AI but we’ll get to that later on.

[00:42:55] Kevin: It’s slightly different. So what you’ve described in the world of banking is called clearing. So debts are simply being cleared. And clearing…

[00:43:03] Emanuel: I know the Romanian term.

[00:43:05] Kevin: A thing that we just do independent of what the debts were incurred for.

But in the case of the… I’m going to call it the AI build out, like Nvidia gives Open AI a hundred billion dollars and then Open AI gives Nvidia a hundred billion dollars back in exchange for GPUs. The hundred billion dollars went back and forth and if you like, okay they cancel each other out, take them out of the equation.

Nvidia just gave them a bunch of GPUs.

[00:43:38] Emanuel: That’s where Michael Burry sees… saw the opportunity. Both the technology and based on what my understanding is… you let me know if I’m right or wrong, but you said that… the following, that all these companies are dependent on Nvidia to produce the chips that they need in order to be relevant, but also the fact that the… their… the lifespan of those chips to become profitable, to become sustainable for 7, 8, 10 years, you’ll need to reinvest again… in technology in these chips made by Nvidia to the newest version as well.

And these companies, as my understanding goes, is that they put them in their books, as how you call amortization within 5 to 10 years, where in fact after 2, 3 years at most they become irrelevant.

So that’s where Michael Burry saw the, opportunity. And that’s where he thinks the…

[00:44:41] Kevin: And I think fundamentally he’s right, they’re depreciating these assets…

[00:44:46] Emanuel: Depreciation…

[00:44:47] Kevin: …on a longer term then they’re going to keep them for.

It would be like taking out a 10 year car loan on a car that you know is not going to last 10 years.

I’m… and maybe you know it because you’re going to sell it in three years no matter what. Maybe the car is just fine and it might last 10 years, but in three years you know there’s going to be a better car that you’re going to want. But when you buy the better car, you still got seven years on that old loan.

So this is going to cause debt levels to increase and people are going to be carrying… companies are going to be carrying these increasing liabilities on their balance sheets and that can’t go on forever.

[00:45:25] Emanuel: Yeah. And it’s not a car that you would want. It’s a car that you actually need to become relevant because nobody will… this won’t be good for… where you are going essentially.

[00:45:40] Kevin: From a financial perspective, the reason that you did this is irrelevant, but the bottom line is you’re no longer using that car after three years and you still owe seven years on the loan. That’s what gets organization’s in trouble, and…

[00:45:56] Emanuel: That’s what’s happening right now.

[00:45:57] Kevin: The AI world is doing that. Michael Burry is a hundred percent correct on that, and that’s not necessarily part of it being a bubble.

The bubble is just these outrageous expectations.

I don’t have the numbers in front of me, but I did a back of the envelope analysis of Nvidia. They’re currently trading at like 25 times earnings.

And then I was looking…

[00:46:20] Emanuel: Maybe more.

[00:46:21] Kevin: They were… I forget… in fact, let me Google for that.

[00:46:25] Emanuel: Okay.

What I see you still turn to Google, not

[00:46:32] Kevin: PE ratio for Nvidia stock.

Oh no. It says if I understand it correctly, it says it’s like 23.

[00:46:49] Emanuel: Yeah. Let’s see.

Let me go, by the way, we shouldn’t state somewhere that…

[00:46:58] Kevin: Yeah, we should just focused and wrap this up.

I don’t actually see the number here someplace, but I remember doing like a back of the envelope calculation and it’s like how much revenue does Nvidia need to generate in order to grow into their stock price?

And it was like an astronomical number. And I’m like, I don’t…

[00:47:17] Emanuel: 12 months is 43. 99.

[00:47:21] Kevin: That’s the PE ratio.

[00:47:22] Emanuel: Yes.

[00:47:25] Kevin: Yeah, so I don’t…

[00:47:26] Emanuel: Double than what you said.

[00:47:27] Kevin: …Clearly remember what I was looking at previously, so I probably shouldn’t have brought it up, but I did this back of the envelope calculation of what kind of revenue increase do they need to realize in order to… what I was calling grow into their stock price and it was just such a big number I remember thinking, yeah, what are the odds of that?

It’s probably not great.

And Nvidia does have a clear headstart, so I know… how far away God, we’re 47 minutes in already.

So I’m currently reading The Theory of Economic Development by Joseph Schumpeter, and he’s the creative destruction guy.

Capitalism progresses through creative destruction. So in his classical definition of creative destruction, that’s what’s happening now with Nvidia and in his models you have a monopoly advantage for a limited amount of time. Could be a year, it could be 10 years, could be 20 years, but it’s going to come to an end because other people are going to produce what you produce that’s going to have a downward pressure on prices, therefore a downward pressure on profits.

And it’s extremely likely that in terms of the… I don’t know what the right word is, unfolding progression. What’s likely to burst the AI bubble is some fab in China is going to start making ships that are like… definitely good enough and they’re going to layer on top of them software similar to… so Nvidia has software called Cuda that handles the parallel processing.

I don’t know a whole lot about it, but every chip they make is Cuda compliant for a variety of reasons.

Some Chinese company is going to come up with GPUs that aren’t as good, and Cuda software that’s not… Cuda equivalent software that’s not quite as good, but it’s going to be a fifth the price right?

[00:49:23] Emanuel: Enough.

[00:49:24] Kevin: And if and when that happens, it’s going to be a problem.

Anyway I rambled on way more than I intended to, and I apologize if it got too weird.

[00:49:37] Emanuel: You got exciting and interesting.

[00:49:38] Kevin: I would like to say that I have a little bit of knowledge on these topics, and if I said anything that was not right, please let me know.

I am constantly learning more about this and if you know more than I do on a topic, please let me know.

[00:49:56] Emanuel: By the way, this is not financial advice. We should put this disclaimer out. We don’t… we haven’t given any kind of advise say purchase something or anything like that. We expressed our opinion, but we’re just two guys that…

[00:50:09] Kevin: Yeah don’t bid up tulip bulb future contracts.

[00:50:13] Emanuel: Yeah. That’s, yeah. Know what you’re doing. And the first rule of money, I think Warren Buffet said, make sure you don’t lose money. That’s fairly important.

[00:50:24] Kevin: Yeah.

[00:50:25] Emanuel: So is it a an AI bubble or not?

[00:50:27] Kevin: Oh, definitely. The question is in which way is it going to… how’s it going to unfold?

[00:50:34] Emanuel: And why is it …

[00:50:35] Kevin: and we don’t. So really to tell it depends on the rules.

[00:50:37] Emanuel: I was trying to summarize again our conversation. So is it an AI bubble? Yes. Why?

[00:50:42] Kevin: Yes, because the valuations of the AI businesses… so it would be an AI business within Meta or Open AI as an AI business… the valuations of those businesses are just not in line with the fact that they’re not making any money.

In fact, they’re losing a ton of money.

[00:51:06] Emanuel: And also you and I are digital marketers are involved with digital marketing. We do SEO, so we can like inevitably talk about AI and use AI deeply. Most people don’t. That’s the reality.

We don’t use all these things that we… I become dependent. I couldn’t imagine right now doing anything without the help of ChatGPT, Perplexity, Claude, tools that I paid for.

That I pay for monthly subscriptions.

[00:51:37] Kevin: Yeah. And I’m about to join you. I’m about to pluck down $20 a month for ChatGPT, so I don’t mean to…

[00:51:42] Emanuel: Three years late.

[00:51:43] Kevin: …demean the AI products. I’m about to get on board with that. I’m going to help Open AI stave off insolvency a little bit.

[00:51:55] Emanuel: Buck by buck.

That’s how you make a living. But no, definitely there’s… or at the… I think it’s called strategic infliction point, right? It depends.

[00:52:07] Kevin: Very much.

Yeah. I think the impact of AI is almost… the full impact of AI is almost impossible to know at this point. It’s just too early.

[00:52:19] Emanuel: And we know that there’s some voices out there that say we’re many… we’re close to quantum computing.

We’re closer than they said initially with quantum computing. Make it usable for the, operations, not you and me, but obviously the big corporations. So that’s… that moment when any of these corporation, it could be some corporation in North America that has plenty… some corporation in another part of the world, private or state owned.

Once they crack that code, then we’re going to actually see the artificial intelligence, the intelligence part in the AI, because right now what we have is glorified database query is essentially in these LLMs, that’s a better description, right? But the moment that quantum computer will happen, that’s when we will see first the intelligence and part and also the changes that we are all expecting at the moment. Until then many things can happen, right? Wars, economic downturns, natural disasters that can inflect political stuff as well.

[00:53:41] Kevin: Part of the continuation of the AI bubble is going to be finding sources of energy. Now finding them is not a big deal, but bringing them online… building like gas powered electrical generating facilities… companies are now applying to build their AI data centers next to nuclear reactors so they can just… I don’t want to buy the electricity from an electricity wholesaler, I want to buy it directly from the nuclear reactor because they just need so much of it.

[00:54:15] Emanuel: Yeah. And nuclear is cheap, clean. And fairly easy to produce, and I would argue, even safe. Not just me who’s saying this, obviously we had some events and I was born in 86.

A couple of months after Chernobyl, you got the Fukushima event happening, or more than 10 years. I believe it’s 12 years, right? 2011. Time flies.

[00:54:38] Kevin: It’s getting safer. Every year we’re figuring out new ways to do it that are safer. But fundamentally, at some point, we as a civilization may face a choice of, do we want to consume less energy?

Which means do we want to lower the standard of living for everybody, or do we want to go nuclear? And my prediction is we’re going to go nuclear.

[00:55:06] Emanuel: I would say the sooner the better. And we have that. And there are some private companies as well that try to scale it at at lower level. What, so there’s this guy, Rory Sutherland, he’s a marketer.

I think he’s a vP or something for for an agency from the UK saying about nuclear. It has the worst term, and there’s some people trying to change that.

They call it elemental energy, which I think is quite good to say. Nuclear energy, elemental energy, but also the Enron egg. Not sure if you’re familiar with what happened, but there’s, it was a, one of those hoax this year called the Enron egg, where essentially they said that they’ve I’m not sure if there’s actually a company called Enron right now… that does stuff. I’m not sure, or it’s a made up one by comedians. Haven’t checked that one, but they said they invented this egg that can power up your home. And in, that’s an excellent concept, right? To have a nuclear power plant a size of an egg, something like this big to power up your…

[00:56:16] Kevin: We’ve gone over, but would you mind if we went over a few more minutes?

[00:56:24] Emanuel: Enron Egg. You associate Enron, which has a terrible name with the nuclear energy, which also has a terrible name.

[00:56:30] Kevin: No, it’s a different thing. It’s called Project Pele. It’s a project of the US military.

[00:56:38] Emanuel: The one that I’m talking about is Enron egg. I’m Googling it right now is a viral, elaborate hoax from early 2025 by…

pretending to be a resurrected Enron, selling a personal egg-shaped nuclear reactor for homes that promised 10 years of power. This is, as far as I know, this is…. this can be done.

Essentially, the technology exists, but putting the name Enron, I think didn’t benefit anyone in the long run.

[00:57:05] Kevin: Yeah, that’s bad marketing.

But if anybody’s interested in what I think has a very good chance of being the future of nuclear energy, Google for Project Pele mobile nuclear reactor. The US Department of Defense is building a prototype at a military research facility, I want to say in Idaho.

And the reason I think it has the potential to be the future of nuclear reactor is because like it’s real, we’re pouring money into it… we’re going to develop a prototype. And then from there it’s liable to leak out into the civilian market, which is a common…

[00:57:46] Emanuel: Happens with all the technology.

[00:57:47] Kevin: Over and over.

[00:57:48] Emanuel: Yeah. Satellite, GPS, so the internet and…

[00:57:51] Kevin: Yeah, yeah.

[00:57:52] Emanuel: I think we’re a little bit over.

[00:57:54] Kevin: We certainly did… for those of you who hung in here, thank you for listening to our ramblings.

[00:58:00] Emanuel: Thank you so much.

[00:58:00] Kevin: I appreciate it. If you like the way we ramble, please subscribe to the podcast. You can find on curiouspundits.com where you can find us, whereever it works best for you.

And if people who you think would enjoy hearing us ramble about stuff let them know. Thank you.

[00:58:24] Emanuel: Send us a message, go to our website and send us a message and tell us what you like or what you don’t like or what you didn’t like about it as well. We’re open.

[00:58:32] Kevin: Yeah. Kevin should talk less.

[00:58:35] Emanuel: Oh, no. That’s why I suggested having this podcast. That being said, Emanuel…

[00:58:40] Kevin: i’m Kevin.

[00:58:42] Emanuel: Signing off.

[00:58:43] Kevin: Okay. Take care guys.

[00:58:45] Speaker 3: Thank you for listening to this episode. If you like the podcast, please like, subscribe, and tell others. Visit curious pundits.com to learn more and stay tuned for the next episode.