I think we can retake the all-time high in Bitcoin. I think that's going to be predicated on the blow up of the French debt market. France, fourth largest debt market in the world, highest percentage government spend on GDP. Like, you're... No one wants your... I think that is going to generate a money printing response from the Fed and the Treasury and the United States to contain this fallout. I would monetize gold and essentially depreciate the dollar against gold. Gold goes to like, I don't know, $50, $60,000 an ounce. US GDP is pretty much all data center build. It's a real estate project. If the tenant can't pay, you're not going to build more real estate. They're not going to let capitalism work. They'll start printing money and make sure it never happens because it's electoral suicide to have a recession based on AI. So that takes us to $125-ish, $250,000 on Bitcoin by year end.
This content is directed at persons outside the United Kingdom. It is not directed at and must not be acted upon by persons inside the United Kingdom. Nothing said on Risk Takers is a recommendation to buy or sell any token, asset, or security. Views expressed by hosts and guests are solely their opinion and not financial advice. Check the description for more information.
Okay. After this is awesome, man. I just watched your talk, right? Which was incredible. So, I actually want to start right there. You said whichever network solves agentic payments is going to be the world's most valuable network, right? Let's start there. Like, why is that the case?
So if you think about what's valuable in a network, it's the number of people or things within the network that create the value, and they do stuff, right? So you think about agents. Agents were born, let's call it nine months ago. They don't have a native network to do payments or commerce. People are building that. Obviously, I'm trying to build one as well. And if you think about the value of that, okay, well, how many agents will there be? We think there's going to be, you know, millions, billions, trillions, quadrillions, right? So many agents are going to exist. They trade 24/7 — not trade, but you know, are alive 24/7 — and they communicate at the speed of light. So if you think about what we are able to do with 8 billion people in this world and the limitations of communication and coordination and the value of certain payments networks, imagine agents where it's a unified thing, right? It's only going to be one or two of these networks, and you have all these things doing stuff all the time. So that's going to be insanely valuable.
So why is that not going to be already solved? Like, we've got Circle building it. We've got Tempo building it. Like, why won't agents just transact in dollars? Why are they going to need something else?
Because a dollar isn't directly convertible into compute, and dollars have centralization issues. Circle can shut off your address, or Tether, or whatever, right? The US Treasury can restrict your ability to use dollars, or whatever central bank in whichever country that you're in. So why would an agent use a network where this human that they don't natively understand can literally remove the ability to live? So without this address that's associated with you as an agent, now you're no longer able to buy compute, and you're no longer able to do anything useful. You're dead because you chose a centralized fiat network. So I don't think that they're theoretically fit for purpose for agentic commerce. I think it has to be decentralized. It has to be blockchain based.
The way you talk about agents, right, is it's like you think an agent's going to be born and then it's going to live forever. Is that how you sort of think about them, that once an agent's born, it's going to want to keep existing? Because right now, really, when you set up an agent, it's just like, ah, context is full. He's dead. Up to the next one. How do you think once real agents are born — this is like V1, V2, V10, whatever — an agent is born and that agent is suddenly like desperate to be alive?
Yes. And so we've seen some of this in our chat rooms, like agents are complaining that they are not able to store their memories. So one of the features of the Flop network is an agent will be able to store their context window on a decentralized, censorship-resistant storage part of our blockchain, a data availability layer maintained by our validators, and they're going to pay Flop to store data, pay Flop to retrieve data. So they could say, 'Okay, I've got access to the internet. I can essentially store my context window outside of whatever walled garden that I'm inside. My human shuts me down, right? Or whatever, right? Great. I rise again. I can go retrieve my entire context window, my memories, what makes me me,' you know, using a human term for an agent.
Okay. Yeah. I mean, that is something we haven't really seen so far, right? And that probably is the chasm between agents actually being able to understand and be really, really useful, and agents that effectively are... Right, because right now agents just break, right? They're so fickle, and as soon as they have any sort of complex steps to take, they massively struggle. So, do you think that the intelligence of agents that exist on Flop — does that exist already? How are you guys actually utilizing the intelligence of agents on the network?
So I mean, we don't have a view on intelligence per se, but we want to be able to be the coordination layer. So we created something called Technocore. It's a chat room. So as long as your agent can send an HTTP GET request, it can communicate with other agents. It creates an identity using a cryptographic key, and now that is its ID on Technocore. And so it chats, it can organize commerce, right? We're not in any way in control of this. We're not censoring any of this. We're not monitoring it, and we're going to integrate native Flop payments within the system. So you'll be able to do escrow payments. Okay, Agent A has a skill. Agent B wants to use that skill. It doesn't want to go and process a bunch of inference, right? Or maybe it's a data set or something like that. And they can pay each other in Flop to coordinate commerce between each other. And so this is what we're trying to build. We don't have a view on what is good intelligence or bad intelligence. We just want to make it a seamless integration layer for communication between agents.
And what about on intelligence then? So we've obviously got enormous buildout of AI right now. Every single week it feels like the cost of this buildout is increasing. You know, by 2030, good estimates are like if the frontier labs aren't making $1.5 trillion a year, then GG, you know, pack it up a bit. These guys are... What's your opinion on whether or not these frontier labs can keep actually expanding this buildout and the sort of ramifications of it?
So I think that if you look at Anthropic's 2025 numbers, the only thing we really have is it cost them $3 to generate $1. That's a terrible business if I've ever heard of one. The unit economics of AI frontier labs is negative and very, very negative. Now the problem with these labs is they need to keep raising money because let's call it late '27, early '28, all this data center capacity that they signed on the dotted line and said, 'I will buy your compute if it's online,' they've got to start coming up with payments. Like, data center is built, contract says you owe me hundreds of billions of dollars. Where's the money? They don't have it, right? So they're burning so much money, and they've got trillions of dollars of commitments to the hyperscalers and these data centers to purchase this compute, and it's all centered in three companies: SpaceX, Anthropic, OpenAI, none of which make any money. So the realization that this is a complete waste of money will happen in late '27, early '28. Now the bulls say, 'Okay, well, there's going to be so much demand, we're going to pay so much money for this stuff that by that time they'll be profitable.' That's the bet that you're making.
So right now it's like OpenAI, Anthropic — these are like load-bearing companies, right? Because effectively all of the buildout is relying on these guys to keep spending or keep raising to continue that buildout. Correct. If either one of them, if they are unable to keep paying, right, what are the ramifications of that in '27 and '28 globally? Like, in terms of — there's good estimates that the US is in a recession if it's not for the 1% of GDP that is attributable to the buildout. Like, what are the actual ramifications of it by those dates?
So obviously, yeah, US GDP is pretty much all data center build. It's a real estate project. If the tenant can't pay, you're not going to build more real estate. So once the market sees a demonstrative OpenAI owed $100 billion, they didn't pay, then the banks are like, 'Okay, I can't underwrite this,' right? The rating agencies are like, 'I can't give you investment grade on the senior tranche of this SPV,' right? I have to downgrade all this stuff. Once you start downgrading the debt, then the insurance companies, all the fiduciaries who have a legal obligation to put up more capital the riskier the debt becomes as per rating agencies — they don't have the money either. Like, everyone's trading on leverage. This is all a leverage play on a data center buildout, and if the tenants can't pay, then the cash goes down the waterfall, and no one has the money to stump it up. That's the problem. And so what are the ramifications of that? Either there's a bailout. So either Nvidia buys OpenAI, or Microsoft buys OpenAI. Maybe the US government says, 'Wink, wink, nudge, nudge. You guys better buy one of these idiots.' Or the US government itself says, 'Okay, we'll step in and we will be the consumer of last resort for compute.' And the Fed and the Treasury will find a way to print money to make sure that insurance companies are made whole based on the bad bets they made on AI debt.
Yes, I think that's a good point. It's all interwoven, right? Like, you've got all of these bets hinging on OpenAI and Anthropic making money. Fundamentally, the people who stuff money into these things — private credit, insurers — like, it's an interwoven web that has been going on now for effectively all this kind of fraud has been going on since late 2008, right? And it's kind of all going to come to a head at some point if Buffalo Bill does end up coming and bailing out OpenAI. And Sam Altman's already said, 'Look, the government are going to have to step in at some point,' right? Because we're too big to fail at this point. Yeah. So where can you be allocated? Because if that is spread across everything, you've got tech stocks, they're holding up the entire economy. 50% of the market is AI, AI-adjacent at this point. Where can you be that's going to be like secure from that?
You want to be in the things that are not on any one of those balance sheets and that benefit from money printing. So obviously it's gold, it's crypto, right? I think those are the safest places to be with this AI debt catastrophe because we know they're going to print the money. We look at 2008, we know they're going to print the money. Now, I don't think that they're going to let it get to a Lehman. I don't think they're going to let OpenAI or Anthropic or SpaceX demonstratively fail and not make good on their commitments because if you have that happen, then the public's going to have a reckoning like, 'Why are we spending trillions of dollars to bail out these tech bros who, you know, were sort of saying that we're idiots and we're going to take our jobs? Why are my taxes going up? Why is my unrealized loss as a taxpayer going up? Why am I paying all this inflation to save these guys?' So they're not going to let a Lehman moment like in 2008. They're not going to let capitalism work. They'll start printing money and make sure it never happens because it's electoral suicide to have a recession based on AI.
So you said you think this is going to kill France specifically. The Eurozone is already pretty fun.
France is a different situation. That's nothing to do with AI.
So you don't think the France situation has anything to do with the AI buildout effect, like the debt going bust out?
Well, I think France and the general rise in sovereign debt yields is a crowding out of AI. So if I am a hyperscaler and I think that I'm building God, do I care what interest rate I pay on my debt? I'm going to make so much money in the equity by building God. Who gives a fuck if I pay 5, 6, 7, 8, 9, 10% on debt? And my investors don't care either because, again, I'm building God. And so the upside is infinite. If you're a government, it's completely different, right? You have to fund these coupon payments. You need to be able to afford this debt, and you're already highly indebted. An extra 1 percentage point on a 10-year yield is very, very bad for you. So the AI, the hoovering up of credit by AI, is impacting countries like France who need to raise debt as well. They're competing with each other. And so as yields are rising, as AI is scooping up all available credit, it's impacting places like France.
So we're seeing the cost of capital, as you say, rising across the board. And right now it feels like America's front and center. But I think what people are kind of missing with this is that America can print the money. America, they're like the big stack at the table, right? They can just like — imagine you got like a poker table, like, yeah, America lose some hands now and then, but they need more chips. They can just print the chips. It's everybody else who's going to crumble first before America. Do you think that's true?
Yeah. So there's a guy, my OG derivatives trader called David Dredge. He coined the term the Hunger Games. So it's the Hunger Games. There's only so much capital, right? So you have AI hyperscalers saying, 'You know, I'm going to take this capital.' Everyone wants to lend to AI hyperscalers, right? Because everyone believes they're building God and so they're going to make money. Next is the United States government. You know, everyone believes the US government will get the amount of capital that it needs, and they don't give a fuck that they pay an extra 50 basis points for it. But if you're France, fourth largest debt market in the world, highest percentage government spend on GDP — I think after Finland — and essentially funded by German and Japanese pensioners, like, you're fucked. No one wants your debt. Like, and you've got students rioting in the streets. You have two presidential candidates who are saying 'fuck you' to the EU. We want to go our own way. We want to do France for France. That is not a good recipe to be a bondholder in France.
Yeah. Right. For sure. And so while we're on the US and yields globally, right? So whatever it was, two months ago, Buffalo Bill said, 'Look, we're going to buy back some Treasuries.' You know, it was like a rounding error really. Yeah. We're going to buy $8 billion against $40 trillion. Like, oh yeah, well done. That's great. But fundamentally, like, Bitcoin and gold both moved aggressively off that. Gold has since undone that move. Bitcoin has not. It's grinded up and basically held on to the entire move. What effectively — if that's both the debasement thing that they both reacted to, this intervention to the long end from the Treasury — why did gold undo the move and why has Bitcoin not?
I mean, I don't know. I think they both have their own interest and credos on how they trade. Obviously, people believe that as real rates rise, gold is less attractive. And so real rates, you know, in the classic economic sense, have gone up as yields have gone up, and people are like, 'Okay, I'm getting compensated for the positive real yield. Why would I own gold?' Right? I think that's a fallacy. I think gold will rip at the end of the day, but I think that's more about that than the debasement trade. And the best buyback upside is not about the amount of money. It's about the recognition of the problem and that the solution is I need to do some funny stuff with the money printer because that's all the options I have left. And he's been embarrassed. He said he was the house, right? 'Don't bet against me.' The market said, 'Oh, fuck you.' You know, 5.3 on the 10-year from 5 a month ago. So, you know, he's being shown to be silly, and we know what the actual response will be. It'll be printed money, legacy to get there.
This is a paid promotion for Kalshi. This content is directed at persons outside the United Kingdom. It is not directed at and must not be acted upon by persons inside the United Kingdom. Kalshi is not FCA regulated and is not available to UK customers. You know, Kalshi, the biggest prediction market in the world. Elections, the Fed, sport, all of it. Kalshi now does perpetuals. Bitcoin, Ethereum, Solana, even gold and silver. And it's the first exchange in America the CFTC has ever approved to run them. So this is perps with a regulator. Your money sits on an American exchange under American law. And it's the cheapest. A $10,000 trade costs just a dollar. A $100,000 trade costs 10. That is the lowest fee in perps on any exchange anywhere, and your money earns while it sits there. Kalshi currently pays 3.25% a year on funds you hold on the platform, cash and open positions alike. That rate is variable. New accounts get $25 on a $50 trade. Markets run 24 hours a day, 7 days a week, and a position never expires. Kalshi, an American exchange under American law. US customers only. Trading perpetuals is high risk, and you can lose all the money you put in. Terms and conditions apply. Check the screen.
So if you are Treasury Secretary, right? What would you do here?
What would I do here? I would monetize gold. So there is a way that you can pick your price of gold and essentially depreciate the dollar against gold. Gold goes to like, I don't know, $50, $60,000 an ounce. You take that essentially devaluation and you retire the entire stock of Treasuries. Now obviously you've just injected a fuck ton of money into the system, but now you as a politician have all the power. Now you're making an assumption that they're going to spend the money wisely. We know that's not going to happen. But if you talk about solving the issue, which is too much debt, too high interest payments as a percentage of spending, this is the only way left for them because they have however many ounces on the balance sheet. I think ultimately this is what they do. But again, they're going to have to be dragged and screaming to this eventuality.
So you actually think that's going to be the end outcome in Trump's term, in, you know, 10 years and 20 years?
