Is The AI Debt Bubble Finally Popping? | Weekly Roundup
Jason YanowitzSantiago Roel SantosRob
Public AI has become more volatile than crypto because markets are repricing the financing chain beneath frontier-model growth. AI-linked equities have fallen 20–30%, with 10% daily moves becoming routine, while Bitcoin held near $82,000. Anthropic’s apparent slowdown, an ARR leak around $16.3 billion, and market-share losses to open-weight models are forcing investors to ask whether today’s roughly $1 trillion of AI capex can credibly become the roughly $4 trillion that Jensen Huang appears to envision by 2030.
CoreWeave’s credit is the clearest warning that even correct long-term AI demand forecasts can produce painful near-term failures. Bonds that comparable borrowers placed near 4.5% a year ago now require low-double-digit yields and warrant coverage; CoreWeave debt traded around 15–16%, while its CDS at one point implied roughly 60% default probability over the next few years. Rob does not expect bankruptcy, but he sees “a very painful restructuring” if neoclouds cannot fund growth and service existing obligations without continually raising capital.
The unresolved Anthropic question is whether attractive enterprise API margins outweigh the enormous cost of staying at the frontier. Rob cited leaked figures suggesting $160 of compute expense per $1 of revenue and questioned whether reported third-quarter free-cash-flow positivity survives numerous adjustments; Character AI was allegedly still spending $9 per revenue dollar. Santiago’s pushback was that enterprise APIs may carry 80–90% gross margins, with consumer subscriptions functioning like Slack’s land-and-expand funnel—but open-model substitution and weak customer stickiness remain the threat.
The frontier-model race resembles the Black Queen paradox: labs must spend furiously to preserve an advantage measured in months, even when incremental capability barely matters to 98% of businesses. That shifts the more attractive economics toward trusted incumbents that can embed cheaper intelligence into existing distribution, cutting costs and lifting productivity. Open weights do not eliminate monetization because inference “still has to be run somewhere,” although the timing mismatch between eventual demand and present debt can still break providers.
The panel thinks the era of indiscriminate AI capital is ending, with survival determined by business quality rather than exposure to the narrative. Jason expects tier-four companies to disappear almost immediately, tier three within one or two years, and tier-two SaaS businesses that cannot adapt to be vulnerable; Anthropic and OpenAI may be tier one without being attractive at trillion-dollar valuations. Santiago compared the frenzy to OpenSea’s former $200 million in monthly fees: hypergrowth makes durability easy to ignore while “we’re evaporating money right now.”
Pearl emerged from Token2049 as the speculative bridge between AI inference and Bitcoin’s long-term security-budget problem. Jason said Rob had a position in the project; its miners perform matrix multiplication useful for inference, receive dollars from service providers plus Pearl issuance, and simultaneously secure the blockchain. Rob framed the thesis as Bitcoin losing hashrate to AI while Pearl gains it. Jason challenged whether miners would manufacture easy, fake workloads, but Rob argued that a Together AI inference endpoint creates demand for genuine work rather than token farming alone.
Crypto adoption is globalizing faster than regulation can unify liquidity or defend local currencies. US rules may force Hyperliquid and similar venues into isolated, KYC-gated pools, even as 65% of tokenized-stock volume on Solana occurs outside US market hours and wallet holders have risen from 170,000 to 4.7 million since January 1, 2025. Nubank’s David Vélez predicted today’s roughly 190 currencies could shrink to ten within 10–20 years as dollar stablecoins “save” the dollar, setting up a direct conflict between user choice and sovereign control.
1. AI equities are repricing growth before crypto does
Jason opened the market discussion with a striking cross-asset divergence: AI-linked stocks had fallen 20–30%, and 10% down days were becoming normal, while Bitcoin and broader crypto were holding up unexpectedly well. Santiago’s practical conclusion was simple: “Don’t use 100% leverage” when public AI can now move more violently than crypto.
Anthropic had reportedly begun pre-marketing ahead of a roadshow expected within three or four weeks, potentially above a $1 trillion valuation. An after-hours leak placed ARR around $16.3 billion and pressured all AI stocks and unprofitable companies; the market read it as evidence that growth was slowing.
Santiago conceded that closed models are losing share to open weights, though he preserved Gavin Baker’s counterpoint: OpenAI and Anthropic may be losing relative share while still growing rapidly in aggregate. The difficult leap is from roughly $1 trillion of infrastructure spending today to Jensen Huang’s apparent $4 trillion view for 2030.
2. CoreWeave credit is signaling a financing wall
Rob’s chain of risk runs downstream from OpenAI and Anthropic: CoreWeave is underwritten partly against commitments from those buyers, so slower frontier-lab growth weakens the contracts supporting neocloud debt. The same question then reaches hyperscalers and every capital-intensive layer built around continued capacity expansion.
Credit markets have already moved dramatically. Deals that cleared around 4.5% a year earlier now require low-double-digit rates plus warrant coverage; CoreWeave bonds traded near 15–16%, and its CDS at one point suggested roughly a 60% probability of default over the next few years.
Rob stopped short of predicting CoreWeave bankruptcy, but the financing loop looked unstable: providers must choose between funding growth and servicing debt unless they can repeatedly raise more capital. With the market almost completely closed to new debt issues and definitely closed to new share issuance, he expects debt-to-equity conversions or a similarly painful restructuring.
Santiago’s pushback was that scarce computing access and harder US data-center approvals could keep capacity valuable, while customers beyond OpenAI and Anthropic—including five or six plausible substitutes such as Meta—might assume leases. He nevertheless agreed that constrained supply cannot rescue every neocloud: “It all comes down to structuring.”
3. Anthropic’s unit economics remain the decisive unknown
Rob cited leaked Anthropic materials suggesting that, for every $1 of revenue, the company had spent $160 on compute. The materials reportedly said the company would become free-cash-flow positive in the third quarter, but the “corrected this, corrected that” wording left him unsure whether that represented genuine free-cash-flow positivity or profitability after numerous adjustments while training costs remained high.
Character AI provided the harsher specimen: even after shifting toward open-source and simpler models, its bankruptcy filings allegedly showed $9 of spending for every revenue dollar. Rob’s inference was carefully scoped but bleak—many heavily funded AI applications may have similarly weak economics despite declining model prices.
Santiago’s rebuttal centered on enterprise APIs rather than subscriptions. Citing SemiAnalysis estimates of 80–90% gross margins, he described consumer plans as a Slack-like go-to-market funnel: seed usage inside organizations, then convert that adoption into much larger enterprise API accounts.
His concession was the crucial one: high margins invite customers to switch toward Llama or other open models, making stickiness and durability uncertain. Until Anthropic’s S-1 reveals what “free cash flow positive” actually means, he expects the entire public AI complex to remain unusually nervous.
4. Falling intelligence costs favor distributors over frontier labs
The frontier race fits the Black Queen framing: “You have to run as hard as you can just to stay in place.” Labs spend extraordinary sums to preserve an advantage lasting months, while Santiago argued that each incremental capability increase makes little difference to 98% of mainstream business use cases.
That asymmetry led the panel toward incumbents with trust, distribution, and durable economics. Healthcare and established enterprises can embed existing intelligence, reduce production costs, raise employee productivity, and benefit as model costs fall—without financing the competition to build the most productive frontier model.
Open weights do not make the infrastructure layer worthless because inference “still has to be run somewhere.” Santiago noted that DeepSeek and other open models can remain expensive to serve; frontier developers may plausibly argue that superior performance per chip makes their total inference cost lower despite a higher model price.
Santiago also questioned the durability of the GPU market. Venture capitalists say older GPUs retain significant residual value, but he found that some companies mainly make money brokering old GPUs; he heard of one allegedly making $2 million a week. That reinforced his concern that money is chasing activities that may not be durable.
Jason’s hierarchy captured the coming shakeout: tier four could disappear almost instantly, tier three within one or two years, and tier-two SaaS companies could fail if they cannot adapt. Anthropic and OpenAI may remain tier one, but he would not buy either at one- or two-trillion-dollar valuations today.
5. Pearl turns AI inference into a Bitcoin-security critique
Token2049 felt more institutional and infrastructure-heavy than prior editions, with large asset managers, banks, stablecoins, tokenization, Robinhood’s network, and prediction markets dominating fewer retail attendees. The genuinely new obsession was Pearl, alongside renewed discussion of Zcash and alternatives to Bitcoin as stores of value.
Jason said Rob had a position in the project, and Rob framed Bitcoin’s declining block subsidy and miners’ migration toward higher-paying AI data-center work as a future security-budget problem. Pearl replaces part of conventional hashing with matrix multiplication used for inference—crypto’s “proof of useful work”—so the same machines can earn inference revenue while securing the blockchain.
Rob cited a Gemma 4 endpoint that he said involved Pearl and Together AI: service-provider dollars pay for actual inference, while Pearl issuance finances a discount and adds a second miner reward. Unlike Bitcoin miners, Pearl miners receiving cash for operating expenses may face less pressure to sell every token they earn.
Jason’s objection was that miners might fabricate the easiest AI-looking calculations merely to harvest issuance. Rob answered that genuine inference demand comes through paying providers, but he remained measured on displacement: Bitcoin is “in a league of its own,” while Pearl is a distinct way to access inference growth and improve proof-of-work economics. He also framed the security-budget critique as bullish for Zcash, while saying it did not imply Bitcoin would fall to zero.
6. Crypto’s global-liquidity prize collides with US market structure
Beyond Pearl, Hyperliquid’s Jeff discussed HIP-4, options, and their complementarity with perpetuals, reviving a category whose earlier onchain iterations never became effective. Backpack’s proposition similarly extends beyond an exchange toward licensed, globally shared liquidity infrastructure, with capital efficiency and cross-margining as the prize.
Jason argued that US plumbing prevents truly shared liquidity. Hyperliquid retail users coming through NinjaTrader would face KYC, FCM-provided margin, a separate HIP-3 pool, and clearing through the Bitnomial DCO; market makers would need separate onboarding, while firms such as Millennium, Citadel, and Tower would avoid offshore arbitrage.
Santiago’s counterargument was not that those giants would participate: smaller firms value cross-margining and common liquidity more, and the structure is designed to solve their capital-efficiency problem. Jason later emphasized Backpack’s day-one focus on licensing across Europe and the US, while Santiago remained doubtful that even innovation-friendly US regulators would permit a genuinely global liquidity pool.
The demand may primarily sit outside America anyway: 65% of tokenized-stock volume on Solana occurs after US market hours. Since January 1, 2025, volume reportedly increased 550-fold and wallet holders climbed from 170,000 to 4.7 million—evidence that tokenized equities currently solve foreign access more clearly than domestic trading.
7. Stablecoins are turning monetary sovereignty into a user-choice fight
Jason highlighted Nubank founder David Vélez’s prediction that roughly 190 currencies could shrink to ten within 10–20 years. His mechanism was fiscal decay: billions of people will seek offshore dollars, Bitcoin, and other assets, making dollar stablecoins “a great savior of the US dollar” by democratizing access.
Nubank supplies the working example. Across its roughly 140 million users in Brazil, Colombia, and Mexico, its initially modest crypto wallet evolved until stablecoins represented about 75% of volume; users were not primarily choosing Bitcoin, but moving “en masse to USDC.”
Jason called the broader contest a “currency cold war”: weaker countries may eventually align with the dollar, China, or the EU as citizens and economies respond to more stable external units. Blockchains add what the internet did not—freer movement of money—potentially breaking capital barriers much as the internet weakened barriers to information.
Santiago resisted the forecast of mass currency extinction. Sovereigns retain local money because it lets them support the financial pyramid; he reduced the policy choices to raising taxes, lowering taxes, or supporting the pyramid, while state power can enforce the system. China’s repeated Bitcoin restrictions and Argentina’s shadow-dollar rate show why free capital movement is “a direct threat to sovereignty.”
Full transcript
Nothing said on Empire is a recommendation to buy or sell any investment or product.
So, hello everyone. Welcome back to Empire. We have a summary today. Santi is with us. Rob is with us.
Santi, look at this hairstyle, my friend.
Dude, I've been on the road. I went to Singapore, and now I'm somewhere else. I didn't even have time to go to my regular hairdresser. Needless to say, thank you very much. I think it looks pretty good.
Thanks, Rob. I appreciate you. That's why Rob is on my good side. Rob is in a good mood today. This is going to be a great episode.
Santi is hiding. It looks like you had to go somewhere else to get your hair cut.
I'll tell you where I am. They wanted to give me a line here, and I said, “No, no, trust me.” I was so stressed when I got my hair cut because I couldn't be distracted by my phone. This guy would have left me with one of those lines right across my eyebrows.
Santi said not to tell anyone where he was, and behind him was a picture of the place where he was.
Oh my God. Let me go. This is impossible to understand. Apparently, AI could determine where I am.
I'm sure if we took a screenshot, it would say where you are.
Oh my God, how are you guys? How's the week?
1. AI Sells Off, Crypto Holds Up
Great. The market does its thing. You would think the crypto market would be more volatile.
Actually, let's start with that. I would say the stock markets, especially the AI sector, are in a complete slump— incredibly volatile. Everything fell by 20% to 30%. Days with a 10% drop have become the norm. Surprisingly, Bitcoin is holding up quite well, as is the rest of the crypto market. I found that unexpectedly positive.
Okay, so crypto is holding up pretty well. AI is exhibiting insane volatility. I think the most interesting prediction market out there right now is whether Anthropic will go public. So come on, let's discuss it—not to turn this into a podcast about AI.
Obviously, I think the whole market is very focused on Anthropic's profitability metrics. They're supposedly doing a roadshow and getting ready to go public, probably at a valuation of over $1 trillion.
I don't think the roadshow has started yet. I think it will start in a few weeks.
They had discussions, and the figure was leaked yesterday. They did some premarketing, but the roadshow to assess the valuation will probably happen within the next 3 or 4 weeks. They're communicating with buyers.
What the Financial Times leaked last night was Anthropic's ARR, somewhere around $16.3 billion. That made the market—after the close of trading—send all the AI stocks and all the unprofitable companies significantly lower.
The markets are also falling this morning.
I think there are still concerns about open source. Closed source is losing market share, similar to what happens with revenue. There is definitely a loss of market share, no doubt about it. So listen to Gavin Baker and the podcast with Jensen Huang if you want to calm your portfolio a little.
The point is that Anthropic appears to be slowing down in growth. It's hard to believe they'll reach $100 billion in annual revenue by the end of the year. Gavin countered by saying that they lost market share, but collectively, OpenAI and Anthropic are still growing pretty quickly.
It's the same story again. The market is trying to make sense of this technology and of this $1 trillion in capital expenditure. Jensen seems to think that by the end of the decade, in 2030, that amount will grow to $4 trillion, which is a huge number.
The biggest concerns, I would say, are what happens if OpenAI and Anthropic slow down; if advanced models capture less market share and more money is spent on open-source models, what does that mean for their commitments? CoreWeave is the focus here. Rob will probably talk a little more about that.
There are natural worries and concerns about monetizing open-source and closed-source models. If there is such a huge amount of spending, will there be an asset-liability imbalance when Meta, Amazon, and Google run out of free cash flow and have to turn to the debt capital markets? People have concerns about this.
2. Ads (Fidelity Crypto, DAS London)
We already talked about this in the previous episode, Rob. I don't know if you have any updates from your traditional financiers and what they're saying about their willingness to fund this, but the money has to come from somewhere. I think there is a growing concern about deployment, monetization, and return on investment, at least in the medium and short term.
As you know, I've been talking for a long time about how concerned I am about the unit economics of these businesses and this whole chain. There is this huge, complex downstream of Anthropic and OpenAI, and what happens to the market if they don't continue to grow?
When you look at a company like CoreWeave, a lot of what people are underwriting is the commitments they have from big buyers. The big buyers are obviously Anthropic and OpenAI, and if they slow down, what does that mean for CoreWeave? What does this mean for the rest of the hyperscalers—the ecosystem on the hyperscaler side?
We've seen it. Bonds began to fall sharply last week and early this week. I think they are at yields of 15% to 16%. When yields rise, it means the market is very worried.
You need to pay a higher yield to hold the bonds. That means there are people selling bonds, so the market is reassessing their value. If you look at what happened with a lot of these new issuances that came out of hyperscalers a year ago, those deals were closing successfully. You could price a lot of them at 4.5%, which means that the buyers of these bonds considered them essentially very sound credits.
Now new issues are being made at low double-digit interest rates with warrant coverage, which means people are much more concerned about credit risk. We see publicly traded bonds, such as CoreWeave's, trading at 15% to 16%.
The CDS, or credit default swaps, for CoreWeave at some point this morning indicated that the market believed there was a 60% chance that CoreWeave would default on its debt in the next few years.
The reason for this is that these hyperscalers have essentially become—I don't want to use the word “Ponzi” in a completely derogatory sense—companies that are, in most cases, unable to fund the deals that are supposed to provide growth and also pay off their current debt and interest without raising additional capital.
They have to either finance growth, pay down debt, or raise more capital to be able to do both. This is the situation they're in now. The market is almost completely closed to these new debt issues, and it's definitely closed in terms of issuing new shares.
If they can't continue to raise capital, they suddenly find themselves in an extremely difficult situation, and they're going to have to restructure in some way. They'll have to figure out how to convert debt into equity or something like that.
That's where I expect we'll end up, although I don't think CoreWeave will go bankrupt. I think many of these hyperscalers are in for a very painful restructuring.
I also think it's true—and this is kind of a digression—that if you look at some of the leaked financial statements from Anthropic last year, the unit economics showed that for every dollar of revenue, they were spending $160 on computing. They were supposed to become free-cash-flow positive in the third quarter, but the wording was very strange. It sounded like, “We adjusted this, we adjusted that.”
I would be very interested to see what this actually means. If they're free-cash-flow positive, we're in a much better position than if it was something like, “They're profitable after 16 different adjustments,” but they're still spending more on training and new cutting-edge models than they're making in revenue.
If this is true, then there is a kind of race going on right now. OpenAI and Anthropic continue to insist that they should be at the forefront. They have to train new models and make all this capital expenditure, but they're losing market share to open-weight models, and Anthropic is also losing market share to OpenAI.
They can't stop spending money. They have to keep raising funds, and then you have the hyperscalers down the chain.
I digress a bit, but this week there was a bankruptcy involving Character.AI. If you look at its bankruptcy filings, even after trying to essentially transition its business to open-source and simpler models, it was still spending $9 for every dollar of revenue.
I think it's probably true that the unit economics of a lot of the big AI startups are about as bad—or really bad—because even though prices are coming down...
3. Can Frontier Models Make Money?
It's an interesting point, because I've seen—and it seems to be a common concern that I hear from a lot of people—your argument about unit economics, even in SemiAnalysis and elsewhere. How does someone like Anthropic make money on models?
It's tempting to say that most people think, “Well, I get a lot of benefit from subscribing,” but they don't make enough money from subscriptions.
My argument is that they make the most money from their APIs, from enterprise customers, and their gross margin on that is very high. As far as I can tell, for example, SemiAnalysis estimates that gross margins are 80%–90%. So, they’re using subscriptions as a go-to-market strategy to get more people into their models and eventually convert them.
It’s like Slack’s strategy, right? You provide, for example, a premium version, people start using it within the organization, and then you eventually convert that into an enterprise customer who pays you through an API. That’s a business model, and it’s very profitable for them. At least, that’s what Anthropic and OpenAI do, as far as I understand.
Of course, then you add the question: What is the stickiness of this product? Can they just take it and replace it? Can they change the direction of their spending? You can also raise concerns in the market: If the gross margin is that high, then they’re very motivated not to use, you know, token max; they’re motivated to move to Llama or some open-source model. What’s the sustainability in that case?
So, these are all extremely valid concerns. In my opinion, look, the market is going to remain extremely nervous and concerned until the S-1s come out. To your point, Anthropic is becoming cash-flow positive, and the market needs to realize that. Until that happens, expect more volatility in public markets for AI companies than in cryptocurrency.
Just a point: Don’t use 100% leverage. Expect 10% fluctuations on any given day. But look, if you analyze all of this, that’s why I like listening to Jensen, who, of course, speaks for himself, but I think he paints a very good picture. He has the best vision of what’s going on here.
And, you know, if you really believe that $1 trillion will turn into $4 trillion, people will continue to be more motivated to spend on intelligence. This is the largest capital infrastructure development in our lifetime. Support it, and understand where the dollars are going. For every dollar of capital expenditure that is invested, where does it flow? That’s what I think about.
But now there is a huge duration mismatch. And that’s a big part of the problem, isn’t it? If we need a lot of capital spending to meet demand, but the capital isn’t there, and there’s a timing mismatch, with a significant portion of that being debt, what happens?
Even if you’re right about the future, you can get a bunch of failures in the neocloud space. This does not mean that the neocloud business is successful. This does not mean that all neoclouds are created equal. It all comes down to structuring and other things.
So, I completely agree with this point. What I probably disagree with is that access to computing power will remain very limited because, without getting into politics, it looks like the Democrats will be in power. It will be a rather blocked process. For example, approval of new data centers will become more difficult, at least in the U.S.
So, access to computing will remain complex and limited. If, say, you have 1 neocloud, OpenAI and Anthropic are big customers. They are not the only customers. There are 5 or 6 other players who could step in and take over the lease, so to speak.
Of course, it’s going to hurt if these guys have a slowdown in revenue, and if they don’t have positive free cash flow and can’t monetize their advantage, then yes, there’s going to be a problem with that. But take, for example, Meta, which everyone forgot about in the AI race 6 months ago, or even 3 months ago. Now they have caught up with the others.
So, I just think that, again, I’ve said this many times: I wouldn’t want to invest. I wouldn’t buy Anthropic for $1 trillion or $2 trillion. I wouldn’t buy OpenAI at these valuations. I just think that at this level, I still can’t understand this model. I think this is extremely difficult.
The game theory here is this: You have to keep investing in training, which costs crazy money, to have any advantage. But it’s not an advantage in years; it’s an advantage in months. So, you need to convince enterprises that you have an open-source solution.
The reality now is, by the way, that each incremental increase in the capabilities of these models actually makes little difference for 98% of businesses. It’s this strange, paradoxical situation that I would have to refer to if I were more well-read. It’s not Jevons; it’s someone else. But they’re in this strange state of confrontation where they have to spend a lot of money just to stay on the front lines, while this advanced feature doesn’t really matter for 98% of mass-market use cases.
Yes. That’s all anyone wants to do.
Yes, I continue to think that companies that have distribution and can monetize intelligence today will be some obvious beneficiaries of this process if the cost of intelligence continues to fall. For example, the health care sector would probably benefit from this, right?
Traditional businesses are likely to benefit greatly because they can integrate intelligence, reduce costs, increase profits, and improve the productivity of their current employees. I think we should look deeper into this. Like you said, it probably comes down to timing, right? When will it manifest? So, prepare for volatility. That’s all I’m saying.
4. Which AI Businesses Survive?
So, Rob, I think what you were looking for—I just checked—is apparently the Black Queen paradox that you mentioned.
Yes, I just looked. "Alice in Wonderland." The Black Queen says you have to run as hard as you can just to stay in place.
Look, I’m not as deep into the subject as you both are. I think it’s worth noting, Rob, that we missed Character AI here.
Yes.
The most important thing that’s happening—I’m not going to talk about margins and capital expenditures, and I know that CoreWeave is a case in point right now, hinting at this 5-year default risk, and then there’s $40 or $50 billion in debt and liabilities. I don’t know financial metrics as well as you do.
What I will say is what is absolutely obvious to me: There are companies that are riding this wave of endless capital and growth right now, and the market is treating them all the same. But in reality, there is a first tier that will survive, and then a series of second-, third-, and fourth-tier companies that, as soon as the market turns around, will start to disappear very quickly.
I would say that tier 4 will disappear almost instantly. Let’s say Character AI is here. Then tier 3 will disappear, probably within the next 1 or 2 years. Tier 2 is perhaps SaaS companies that are actually great businesses but are unable to adapt.
So, I believe this is the beginning of the end for this wave of endless capital and endless growth. And look, there are going to be some amazing things. I won’t be buying Anthropic or OpenAI in the next year, but maybe I will in the next 10 years, because I would classify them as tier 1, right?
So, the question is, who belongs to the first tier, and what is your time horizon? This, in my opinion, is what is happening. I don’t have a deep enough understanding of CoreWeave, but I have a pretty strong opinion that we are at the beginning of the end of this infinite-capital game.
Yes, I think that’s right. I think that’s true on the venture capital side as well. We probably have a little more time because of the venture capital monoculture, but Ramp just raised money today at a $60 billion valuation.
Yes. You know what the funniest thing about Ramp is? Insiders sell 100%.
5. Is CoreWeave Facing A Debt Crisis?
Yes. Yes. But I mean, it’s different, right? Anduril is different, but I would actually put all of these companies in the first tier. So, in my view, it’s actually potentially a good investment, but it’s the second-, third-, and fourth-tier companies that will get crushed.
I was actually told by a Ramp investor who runs the fintech side of a very large crossover fund that Ramp is his access to AI. He says, “They are in my AI lab.” And there are truly incredible things happening there in the field of AI.
Yes. Yes. This is actually not a misconception. I even like this look. I think they are doing amazing things on the AI front.
No, but what I wanted to say is that I think there’s an ongoing infusion of money into AI applications that are incredibly disruptive from a unit-economics perspective, but are funded by venture capital and are trying to build a moat with capital, computing power, and so on.
I just think—and I know a lot of people who disagree with me who are in the same profession as me—that we’re evaporating money right now, like in a 2021 situation, for a lot of the things that are happening in AI. I just don’t see how any of these businesses can be sustainable.
You can try to estimate it. I recently saw a tweet about how OpenSea was receiving $200 million in commissions per month at its peak, and it’s hard for an investor not to look at that and say, “Look, we’ve never seen this kind of acceleration. We’ve never seen this kind of growth. We just have to jump on this bandwagon.”
That’s what’s happening in AI right now. Of course, I’m not saying that NFTs are the same as AI, but when you see hypergrowth in a very short period of time...
It's hard to look at it and say, “Okay, what's the durability of this?” I think a lot of what we're seeing right now is a real lack of longevity in companies that are building in a new, original, and useful space. But that's not necessarily going to create great businesses in all the parts of the stack that are currently being funded.
6. Who Wins As AI Gets Cheaper?
Yes. I mean, a dollar invested today: would you rather invest in a cutting-edge model, or would you rather be an investor in a traditional enterprise that has already built trust and sustainability in its business model? And to your argument about Ramp or other traditional businesses that can simply lower production costs with AI, they benefit from being on the right side of the curve, where the cost of providing intelligent services falls. This bodes well for certain companies.
So I would rather invest at this level than in one that tries to compete for the most productive model. By the way, whatever that means, because you can also just claim to have the best model. Ultimately, I don't think we're at a point where the gap will be that noticeable. Maybe my words will sound silly in time, but right now the productivity gains seem more marginal compared to how much someone is willing to pay you.
In other words, the difference between the models a year or 2 ago was much more noticeable. I forgot what they had with the nomenclature—ChatGPT-2 or 4.5—but there was a big jump in performance. Now, in my opinion, it is becoming less and less essential. This makes it difficult for those who want to maintain leadership by spending huge amounts of money, while you can already use the intelligence available today to make your business much more efficient.
So listen, I'm a big believer in open-source models. Right now in the US, in the short term, they're the ones being relied on, not advanced models. I think that will remain relevant for the next few years.
One last thing on this subject: I think people are mistaken in thinking that, despite the transition between closed and open models, running inference still has value.
Yes.
There are obvious beneficiaries of this, which is why I like the companies I work with at this level of the stack. But this is a common misconception that I often hear. They say everything is becoming open, so monetization is impossible. And I say, “No, no, no.” It still has to be run somewhere.
Even if you look at DeepSeek and some other open-source models, the cost of running the inference is quite high, sometimes even higher. The objection to using the more advanced models is performance per chip, meaning the energy to service them. It's more efficient for NVIDIA racks than for others. So I think that's probably what you'll hear from some of these advanced-model developers saying: “No, no, no, the total cost of running this is going to be cheaper with us than with the open-source model.”
Interestingly, Jana, you did a podcast about computing-power markets a week or 2 ago, right? One of the things we're seeing—and this is kind of a digression—is all these discussions around the computing and energy markets, not just computing itself but the whole stack, and how to access the market or hedge against risks.
One of the stories that a lot of venture capitalists are telling right now is that older GPUs continue to hold their value very well. They have incredible residual value, and there are all these companies in the computing market that say, “Oh, well, there are going to be these hedgers, they'll have these indices, and we're going to do this,” and there’s going to be a huge market that will allow people to hedge some of the risks we're talking about.
Then you go and talk to all these people now, and the only thing they make money from—they make a lot of money, but the only thing they make it from—is brokering the sale of old GPUs. They make a lot of money. I heard about one company that was supposedly making $2 million a week right now brokering old GPUs.
I think the whole market structure is really broken. There's a lot of money floating around for things that are clearly not durable. For me, that's the real question: will people just buy off this because there's enough money around, or are we in for one of the biggest crashes in many sectors that we've seen since the last time we talked about the railroad boom? Sometimes I really feel that way.
But let's end on a positive note. Bitcoin is at $82,000, which is positive. We should talk about tokens, the ledger, and any updates on that. But to sum it up, very few people use it today, and you'll never hear anyone say they don't want to be smarter. If you have the ability to use intelligent technology, there's an argument for Jevons' paradox.
It's just that right now the application of AI is very limited, and it needs to grow by orders of magnitude. But it's always a matter of time.
7. Ads (Fidelity Crypto, DAS London)
Can I make one last point about this? I never want to hear the term “Jevons' paradox” again, because it has become a meaningless phrase. People just say “Jevons' paradox.” This is a new level. It's like the phrase “I think from first principles.” Remember how it was with “first principles”? Now you can't even pronounce it because it sounds so awkward. You can't remember Jevons' paradox anymore.
Take a disciplined approach to crypto investing with Fidelity Crypto—a platform built in-house with over a decade of crypto asset experience. At Fidelity, you can trade cryptocurrency and stocks simultaneously, protected by industry-leading security systems. When you see your bills in one place, it’s easier to make smarter decisions. Get started at fidelity.com/crypto. Crypto asset services are provided by Fidelity Digital Assets. Assets are not FDIC or SIPC insured and carry a risk of significant loss. Securities are offered by Fidelity Brokerage Services LLC, member NYSE and SIPC.
This November 10-11, the main institutional event in the cryptocurrency sector returns to London. Blockworks Digital Asset Summit at the Hilton Park Lane Hotel. 1,500 attendees, over 700 institutions and speakers from BlackRock, JP Morgan, Coinbase, Robinhood, Figure and others. Meet the developers and investors who are bringing topics like tokenization and DeFi to the mainstream. Two days of panel discussions, networking, and additional events building the financial future online. Order tickets at blockworks.com/events. See you in London.
8. What Stood Out At Token2049?
Okay, let's move on. Tell us about the main findings from Token2049. Name something very specific: which companies were at the center of the discussions? Rob, you are not allowed to talk about Rain, Ethena, or Polymarket.
I'm a little less biased. Well, I have—
You can. You are allowed. Rob, what were your impressions of Token2049? What companies or tokens were popular? What were the narratives? And just don't say “tokenization,” because we already know that. I'm very interested.
I get it. You're telling me what I should say on this podcast.
Okay.
Listen, tokenization has really been a hot topic.
You're killing me.
No, no. I'm going to—let me speak.
Okay.
I think the main takeaway, if you start at the top, is that there was a lot more institutional interest—big companies, big asset managers, big banks—and fewer regular retail investors than usual. The space continues to focus on infrastructure and institutional uses, so it's understandable that we keep talking about stablecoins and tokenization.
There has been a lot of talk about the Robinhood network and what it means for the future of other companies that may be building their own ecosystems. We should think about the fact that, yes, Robinhood was great in terms of fees and so on, but it's still mostly meme-coin trading, and what that means for the industry.
There was also a lot of talk about prediction markets, which is still relevant today. But the biggest winner of the week was probably Pearl. Pearl was probably the one that was talked about the most.
Maybe Zcash got a little more attention, too, along with concerns about future tokens as stores of value that may have a better structure than Bitcoin does now. It seemed there was a lot of talk about this, too. This is probably the only new topic I encountered while I was there.
Wait, Rob. Let's wait until Santiago leaves and talk about Pearl, because our analysts are back. I first heard about Nock at a dinner somewhere, maybe 6 months ago or a few months ago, and I didn't pay any attention to it. Our analysts are back from the conference, and everyone is saying, “Pearl, Pearl, Pearl.”
I just saw that we have another podcast called Zerx Research. This is our podcast where analysts discuss what they're betting on, what they're selling, and what they're thinking about right now. I saw a clip from yesterday from our Zerox Research podcast in which someone said that 90% of the people at DA Singapore were talking about one coin. That shows what kind of environment analysts are operating in if 90% of them are talking about one coin.
It said, “Pearl is to Zcash what Zcash was to Bitcoin.” So while I'm sitting here in my cozy Zcash, they've already moved on from Zcash and are talking about Pearl.
Yeah.
Wait, Rob. You have a position in Nock. What is this? What is the summary of Nock, and why is it worth buying?
One of the problems with Bitcoin right now is obviously the Bitcoin security budget. There are fears that almost all Bitcoin miners are moving into the AI space because they get paid more to build data centers.
There is a real concern that, over time, there will not be enough miners to properly secure the Bitcoin network because it simply isn't paying them enough anymore.
Pearl is a kind of Bitcoin derivative, a new Layer 1 that replaces some of the hashing required in Bitcoin with matrix multiplication—essentially, the mathematics used for AI inference. In the crypto world, this is called “proof of useful work.” No one in the AI field would call it that, but it’s what people in the crypto industry are talking about.
The idea is that you can incentivize miners to perform matrix multiplications to trigger logical inferences, and they get paid by the providers of these services. I think Together AI has an agreement with Pearl for an inference endpoint. At the same time, as long as they’re doing this multiplication and getting paid for it, they’re also securing the blockchain.
This essentially solves the problem of reusable proof of work, whereas Bitcoin miners solve random puzzles that have no economic meaning. That ties into a narrative about logical inference, so there’s value in it. While talking to Muneeb from Stacks, I said, “Listen, man, this is why I don’t like Bitcoin. This is extremely uninteresting, because every time I’ve talked about Bitcoin’s security budget for the last 10 years, everyone says, ‘The block rewards will eventually go to zero, but don’t think about it. It’s not a problem now; we’ll figure it out as we go along.’”
I said, “No, no, no, no.” That’s why Zcash uses this opportunity on the quantum side, and Pearl uses it from a security-budget perspective, I would say. These cryptocurrencies are quite fragile, and people don’t realize it. They work until something else happens, until everything breaks.
When that breaks down, it’s like Soros betting against the pound: It undermines trust. Once you undermine trust, it’s a slippery slope, because the only value of these things is social consensus. They’re a kind of artifact, like money.
This is worrying. I believe that once the block emission goes to zero, you’ll have to believe that you’ll earn from fees, but fees are low. So what will protect the blockchain network then? Everyone says, “Oh, well, you’re subsidizing it in one way or another.”
I think we had James Preswich here about 3 or 4 years ago. He suggested something—he had an article about it a while ago—and he’s one of the few people who’s talked about it and tried to suggest something. But it’s hard, really hard, to get it done and even to talk about it in the Bitcoin community.
I understand that you must resist change. This is what made Bitcoin, in my opinion, truly sustainable. You just don’t want to innovate. It’s difficult for a reason. It’s a cryptocurrency, but there are serious concerns right now.
I would say the market is very concerned about this. That doesn’t mean that Bitcoin will depreciate to zero, but I don’t think the probability of Bitcoin being supplanted by something else is any higher than it has been in the last 10 years.
This isn’t slow progress. It’s like the frog in the proverb about the frog in boiling water. You undermine trust to a certain point until it’s completely gone, and when it’s gone, you just move on to something else. This is clearly a bullish scenario for Zcash: Capital will continue to flow from Bitcoin to Zcash.
9. Does Pearl’s Mining Model Work?
I have a question for you about Pearl. When I first heard about Pearl, I thought it was really cool. So, Rob, if I understand correctly, I’ll explain this as if I were 5 years old.
You have Bitcoin miners. They do a certain job—they solve a problem—and they get rewarded for that solution. But you don’t actually do anything with what you create, which is the bitcoins. It’s like a “dead” asset, a store of value or something.
Many miners have started redirecting their computing power to artificial intelligence, such as CoreWeave. They get more money for inference workloads than for Bitcoin mining.
That’s right.
Then Pearl says, “What if you could perform calculations for AI but also receive this Pearl token on top of that?” So I have a question: Is this correct so far?
Yes. Good.
What concerns me about this, or what I think will happen, is that if you’re being rewarded with tokens for solving math problems, you’re actually going to solve the simplest problems to prove to the network that you’re solving them. You won’t actually be performing AI inference.
You’ll just say no, because you’ll be paid for AI inference by inference service providers, right? The computer doesn’t know whether you’re answering real AI questions or helping someone. I know you’re an investor in this project, and I don’t know anything. This is just my opinion as an outsider, but wouldn’t you just do a bunch of fake AI work that looks like real work to get the easiest loot—the Pearl token?
Does this make sense? You have no incentive to make money doing real AI inference, right?
But they have an incentive to make money from real AI inference, right? They were paid through—Pearl has now launched, I think, the Gemma 4 endpoint with Together AI, right? That was financed at a discount.
They were essentially able to lower the price a little bit because they financed that discount or incentivized it through the Pearl issuance. Because of Together AI, people were interested in using Pearl for their inference workloads, and they were getting paid from that side. Pearl emissions were also present, so they happened together.
What is the point of this token then?
Well, you want to decentralize the network, right? So, MV = PQ—what’s the velocity here? People will just sell the token.
Miners—I mean, that’s what happened to Bitcoin, right? The biggest sellers of Bitcoin were miners. Why is that? Because miners get paid in bitcoins, so they have to sell them to fund their capital expenditures or raise more funds.
In the case of Pearl miners, they get paid in another way. You get a token and cash, right? So theoretically, you’re also less interested in selling the token.
Got it. I understand your point. Together AI pays them in dollars, so they don’t need to sell the token as much as a Bitcoin miner, who will sell it immediately.
The thesis is that Bitcoin is losing some of its hashrate due to AI, while Pearl as a network is actually gaining hashrate due to AI.
Probably, yes. Yes. Right. I mean, this happens. They’re now synergistic.
Is this a competitor to Bitcoin?
I personally think Bitcoin is in a league of its own. None of these things are a competitor to it. It will continue to exist as the digital-gold narrative.
I think Pearl is similar to Bitcoin, and of course it will be of interest to some people who own Bitcoin. But that’s a whole other thing, and it’s one of the best ways to access the growth of inference while still being able to manage the blockchain.
But Santiago is more pessimistic about Bitcoin than I am.
I am not a pessimist about Bitcoin.
10. What’s Next For Crypto Exchanges?
No, no, wait. Let’s move on from Pearl. Tell me about other things with tokenization.
So, Pearl.
Pearl—that’s good. Pearl—that’s great. Rob, good answer. Good answer, Santi.
Well, thank you, Jay. I’m glad you approve.
I approve. I approve.
Really?
If you just said “tokenization,” I’d be bored.
So, yes, I can name a few. I was on a few panels. The Backpack one was big; of course, I’m an investor there. We’ve already talked about this here, and it was a big topic of discussion.
Hyperliquid also had a very good event. Jeff talked about HIP-4, options, and how they complement perpetual contracts. Options are at their peak right now, and derivatives protocols have grown from less than $1 million to $5 billion in a short time. I think there’s renewed interest in options as something that was tried before.
There were ones like Hedgic and others, but none were truly effective on-chain. Now it seems that interest in this has increased again. Jeff, of course, talked about it, so I think that got people excited about what the next stage of Hyperliquid’s development will be. That was interesting to watch.
I met with Kyle. I was on a panel, moderating a discussion about this being just the end of the beginning. That was a good conversation. Kyle is now reportedly joining the Backpack board.
What’s interesting about Backpack is the level of shared liquidity. When you think about Hyperliquid entering the U.S. market, it always means liquidity fragmentation. For example, you have Binance, Binance.US, Hyperliquid, and presumably Hyperliquid U.S.
I think there’s a lot of talk around Backpack about having shared liquidity—essentially, a global pool of shared liquidity—which I found interesting. Most people perceive Backpack as an exchange, not a prime broker or a venue for shared liquidity, and I think people understood that better during the token sale.
I mean, it’s a stock exchange. It’s definitely an exchange; of course it’s an exchange. But there’s much more behind it in terms of the licensing infrastructure that allows it to be a global pool of shared liquidity, not just a local one.
This isn’t just Backpack, because Kali, Polymarket, Hyperliquid, and everyone who’s trying to get a license in the U.S. right now is talking about this. Many of them are discussing this level of shared liquidity for global markets.
The biggest surprise in the crypto sphere, without a doubt. I wanted to say that I think it’s very—I think it’s wishful thinking that regulators would allow this. I don’t believe regulators will allow this because they don’t want it.
If you go talk to U.S. regulators now, even as you discuss what Hyperliquid is talking publicly about doing with Kraken, right? With Hyperliquid, you’ll have some retail traders coming through NinjaTrader, the FCM, right? They’ll undergo KYC through NinjaTrader, and they’ll have to work through this platform.
They will also have to receive margin and leverage from NinjaTrader, right? They will go into an isolated liquidity pool on Hyperliquid, or what they are now calling HIP-3. This liquidity pool will be different from the liquidity pool for the same markets elsewhere.
Then they will conduct clearing. They would have to send the same transaction, follow the same reporting requirements, and do the same clearing through the DCO, which is Bitnomial, which they own, right? Market makers will have to go through the onboarding and KYC process through Bitnomial directly for this American structure, right? So it would be an isolated system, right?
So the question arises: if there is a lot of liquidity, for example, in Hyperliquid and in another pool, does this really increase liquidity? After all, they can conduct cross-market hedging and arbitrage between offshore, unregulated markets and other parts of the market. But Millennium and Citadel won't do that, will they? Someone will do it, right? Tower won't do that, right?
It becomes quite clear to me: yes, Hyperliquid is incredible technology, and many people have great technological solutions, but if the CFTC requires this kind of clearing infrastructure and FCMs, you can't really aggregate liquidity in these global markets.
I'm not arguing with you. I think Armani would tell you the same thing. The big players don't like it, but others absolutely do. These are the small players for whom you are solving the issue of capital efficiency, which has always been a problem in the cryptosphere. They are focused on allowing small players to operate within a common level of liquidity.
I think you're right. Big players won't necessarily go for it, but other parts of the market like it because you can get an advantage through cross-margining and the overall efficiency of a shared liquidity level. I don't think he's building this for Millennium or Citadel.
But it's still an isolated level of liquidity, right?
Yes, it is still isolated.
No, it should be invisible. For example, the CFTC in the US will require this. You can do it, but focus on the fact that their license portfolio is much broader. This is what makes Backpack special: they have been focused on the licensing part of their business from day 1, making sure they have access to the European market and the US.
The US will now probably require a separate liquidity pool, even if the other pool is, say, regulated in the EU or Japan, right?
I'm not saying that's not true. I'm just saying it's a huge prize, and I hope we get there. In my conversations with regulators—and Armani and I know a lot of people in common—we actually talked about this last week. We discuss this with many people in Washington whom we communicate with.
I don't think regulators will be ready for this, even now, even though these are the most innovation-friendly regulators in the US in a long, long time. Mike Selig is doing an incredible job there, but there are still things they will require, and I think that will make it difficult to create the kind of global liquidity levels that everyone is talking about.
Does that make you more pessimistic? If the US simply banned it and didn't allow it, who do you think would benefit the most, and who would lose the most?
That would be good for teams that already have distribution in the US, right? Or for those building there, because the US is still the biggest market for everything, not just crypto. In crypto, the biggest “whales”—the people who do most of the trading—are actually offshore, right? At least in the retail sector, in Asia.
You and I have talked about this a little bit. All these offshore exchanges are coming into the domestic market. At the same time, there are some domestic exchanges trying to find ways to expand internationally. Everyone is trying. The market is certainly becoming more global, but whether the global market can become a common pool of liquidity is, in my opinion, a different matter.
If the hardest place—the prize—is the US, then the people who are already here, who already have a great brand and a lot of money to enter the market, are the ones who will benefit.
Well, man, the counterargument is that the US is much more competitive. It is better to focus on other markets where there is less competition and where you could attract a lot of latent demand, such as options and tokenized shares.
Most of all, as you rightly pointed out when we were talking, the competition in the US is extremely tough. From a customer-acquisition perspective, do you really want to go there and compete with the big players? Maybe not.
So, we have a graph showing the percentage of trading volume on Solana during US market hours and after hours. 65% of tokenized shares on Solana are traded during US after-hours. It shows that 2/3 of tokenized shares on Solana are traded when the US stock market is closed.
What I do with that information is up to you, but I think that, with stocks, it is still very clearly a story about access for non-US users.
100%.
We started the year with 170,000 people holding tokenized shares in their wallets—that was January 1, 2025. Now there are 4.7 million of them. In my opinion, this is the most interesting graph in the entire technology sphere: the volume of tokenized shares has increased 550 times since January 1, 2025, and the number of wallet owners has increased from 170,000 at the beginning of the year to 4.7 million today.
11. Will Stablecoins Replace National Currencies?
These are huge, simply colossal numbers. I think this is very interesting.
I had a clip that, Rob, you'll like. It feels like it's been a month since our release of Invest Like the Best, so I think we can talk about this again. There is such a thing as Nubank.
I actually didn't listen to this episode, but Nubank is a phenomenal business that probably few people in the US know about. This is one of the best companies in the world at the moment: 140 million customers and a great founder, David Vélez.
There is one great quote about stablecoins. This is one of the largest banks—probably, Rob, the biggest neobank in the world right now.
One of the biggest.
In any case, Revolut and Nubank are 2 of the best neobanks in the world. This is a quote from David, the CEO and founder of Nubank. He talked about stablecoins and why they could save the US dollar, and I like this perspective.
He said that there are 190 currencies in the world. In his opinion, in 10–20 years there will be 10 of them left. The fiscal situation in many countries is so dire that a lot of fiat money is not actually real money. Therefore, billions of people will increasingly use offshore finance, US dollar stablecoins, Bitcoin, and other assets.
He said, “I think a dollar stablecoin will ultimately be a great savior of the US dollar because it democratizes access to it. At least, there is a lot of demand for dollar stablecoins. Unfortunately, the demand for Brazilian stablecoins is very low.”
So he's talking about saving the US dollar. I think there is an even more interesting point, or rather a tangential one. Now every country has its own currency—not every country, but most countries have their own currencies—and that will all disappear.
No, nothing will disappear anywhere. There's a reason why. From a sovereign monetary-policy perspective, you wouldn't do that. This will not happen.
What I will say is that Nubank launched the Nubank Crypto wallet about 2 years ago. It wasn't that big of a success. They learned a lot and worked hard to improve the product. What you see now is what we studied in detail when we analyzed that company in Mexico.
I think the vast majority of the volume in the wallet is just stablecoins. Not stablecoins and Bitcoin, but stablecoins—that's about 75% of the total, and that's really driving their business. They're in Brazil, Colombia, and Mexico, but a significant portion of those 140 million users are in Brazil. Most of them just leave and switch en masse to USDC. It's an incredible product for them.
I don't know if I've said this on the podcast, but I've said it to our partners, and I may have said it to you outside of the podcast. I've believed for a long time—for years—that we're in a currency cold war. What David said is what I've said before: there are 190 currencies in the world, and many of them will disappear.
Santi is absolutely right that there are concerns about sovereignty and monetary policy in all of these countries. But it's also true that they are becoming more dependent on the US and China every year. The US and China are working hard to move production closer to home and increase their influence in many parts of the world, and the result is simply a more stable currency.
We see that the US dollar is more stable than local currencies in many of these 190 countries. My expectation is that, as a lifeline, many of them will take the opportunity at some point to be pegged to a country and currency other than their own.
When and how this will happen is unclear. El Salvador, for example, is the first swallow. Venezuela—I think these countries will become bigger examples. They will start to see a kind of division happening under greater influence from China, greater influence from the US, and greater influence from the EU, in how people adapt their economies.
Yes, the euro is an interesting experiment, because there were, of course, concerns, especially during the Greek debt crisis.
The euro, in my opinion, remains a very interesting experiment: an attempt to bring together different economies, different growth rates, and different fiscal policies—which, in some aspects, are quite different—and unite them under the slogan, “Let’s just avoid a World War II scenario.”
There’s some resistance from people about leaving the European Union. I don’t know. Brexit was bad for Britain, I think.
Yeah, there are a number of reasons. For me, yes. Listen, would you accept the argument that in 20 years we’ll have, let’s say, about 180 currencies?
How many currencies will we have in 20 years?
Well, does this mean that we’ll still have 190 countries? It’s kind of—But let’s put it this way: I get your point. Yes, there will be more and more countries like El Salvador transitioning to a dollar-based economy.
To some extent, this is already happening, isn’t it? You go to Latin America, where you can pay in dollars. Most companies have dollars. China, for example, holds a huge amount of U.S. Treasury bonds.
But I still believe that the reason you want to have your own local currency is to support the financial pyramid. You work with very unbalanced budgets. You can print money.
But you can’t support a pyramid scheme if your users have access to a better product.
Your citizens have access. People were locked into this pyramid, but now, if you can take the red pill and get out of this pyramid, you can.
The reason sovereigns continue to exist is because they impose their will by force. That’s what makes a nation-state a state: you have the violence to enforce this type of regulation, right?
Venezuela had one, and Argentina had a shadow dollar. You would see the difference between the official rate and the shadow rate. Politicians like to have their own currency because they can support the pyramid for as long as they can.
There are only 3 levers, right? You can raise taxes, lower taxes, or support the pyramid scheme. The easiest thing to do is support the pyramid scheme. For that reason, I think it remains a problem.
Capital flight has always been a huge problem. That’s why China has banned Bitcoin so many times, right? India doesn’t like this. When capital can move so freely, it’s a direct threat to sovereignty.
The internet has made it possible for information to spread freely and unhindered around the world. There are some countries that have managed to ban it, but there are very few of them.
What the internet didn’t do was allow money to move freely, easily, and efficiently. Blockchains allow this to be done. Some countries will find a way to block it, but for the most part, it will be like the internet, I think, and almost all global barriers will be broken down.
So, we have 180 currencies. Some countries will figure out how to ban it, but I don’t know. Should the Mongolian tugrik exist as a currency? Probably not, because I think the users—
Your Mongolian listeners will be outraged after this.
They’ll be angry.
I’m actually very fond of everyone in Ulaanbaatar. The Gobi Desert is beautiful. I want to go there. You saw the video from Mongolia of falcons flying, and I really want to do that.
But Albanian lek? I think we could just switch to dollars.
I wonder what would happen in a world where anyone with access to the internet could choose a dollar.
That’s absolutely right.
This is very interesting.
I think this is a good sign for stablecoin issuers like USDC and for neobanks like Nubank, as long as they can attract deposits where people can store funds.
Nobody wants to keep their fortune in local shitcoins, dude.
This is why it’s better to have Bitcoin, which is very volatile, than your local currency. Compare the volatility of Bitcoin to the volatility of, say, 20 currencies. Those 20 currencies are likely more volatile than something like Bitcoin.
It’s crazy to say this, but it’s true when you consider hyperinflation.
Content of the week.
12. Content Of The Week
You made it easy for me with Mongolia. One of my favorite history books is—
I thought you’d say you made it easy for me, like we’ve brought the best back to the table again. That’s what I—
Well, yes, you mentioned that. The episode about Nubank is really good. He says that if you look at fintechs like Nubank, Revolut, and others—Monzo or something like that—you still have only 10% penetration compared with traditional banking institutions. There’s still a lot of room for growth.
But I wanted to say that one of my favorite books is "Genghis Khan and the Birth of the Modern World" and its sequel, "Queens." If you want to delve into the subject, this is one of the best books I've read. It's just a fascinating book about Genghis Khan and the Mongol Empire.
Those are wonderful books. I agree. Rob, what do you have?
I don’t know how I could recommend anything else this week other than “You Can See Everything” [?]. That is literally the only thing to recommend this weekend.
What is this?
Nathan Fielder or Elizabeth Holmes. I’ll double down on this recommendation.
They held a premiere in New York last week, and I was offered tickets, but unfortunately I had to decline because I was in Singapore.
At least you can talk about it on the podcast.
At least I can talk about it on the podcast. This is what I want people to know: I could have gone to the premiere.
It was part of New York Film Week, which I sometimes participate in. As everyone knows, I like movies. I haven’t even seen it yet, and I still recommend it because I think it’s going to be a major topic of discussion for the next few weeks, at least among the tech and crypto communities and a lot of our friends.
The trailer was incredible.
What an incredible trailer.
I’m not a big fan of Nathan Fielder, to be honest, because I find him too awkward for me. It seems to me like you’re being pulled over. It makes you cringe when you watch it.
That’s right. It’s 3 hours long, and I might be twitching the whole time, but I’ll definitely watch it tomorrow.
Probably tomorrow.
So you’ll go to the cinema and watch it?
I go to the cinema. You love movie theaters.
I go to Nitehawk all the time. We’ve already talked about this. I’m probably the biggest Nitehawk fan. They should have given me a share in the business.
Santi is like, “I won’t even tell you what continent I’m on.” I’m like, “I’ll be at Nitehawk at 7 p.m., fourth row.”
Come to Nitehawk.
They have a great Old Fashioned. You can watch a movie, and I’ll treat you to one.
This is great. Here’s the thing: if you’re trying to raise money from Dragonfly and you haven’t gathered all the necessary clues, then if you want to pitch the Dragonfly project, you have to go to the movies the old-fashioned way. You have to have a list of horror movies that you’ve recently watched.
Listen, you guys know the way to my heart. I was literally talking to someone this week about the possibility of co-investing in a number of A24 films.
Fintech, prediction markets, and movies. Let’s do this. Boom. That's right. We have a really good episode coming up with Michael Dempsey from Compound on Monday. You guys will really like it. We actually talked—I mean it was a very long conversation. One of the most unique episodes of Empire, I think. So listen to it on Monday. Dempsey is a very good person. He runs a very interesting company called Compound. He's been running it for about 10 or 11 years. Cool venture capital firm. So, Santi, Rob, it was nice to see you. Thank you to everyone who listened. See you next week.