Transcription
With these AIs, you replace more knowledge work within human civilization and demand keeps growing. So I don't think we're anywhere close to the saturation point of AI capex build. If this is proving that you need to build even more capex because at the end of the day you need to sell to anthropic, you need to sell to OpenAI, Gemini and XAI and all these sorts of players.
Now we are thrilled to welcome back to the podcast Maelstrom founder Arthur Hayes. He's back in the virtual studio again for the next expert council iteration. Now we're going to dissect the macro forces driving this massive AI capital expenditure bubble. Uh, is there potential for like some s looming mega market disaster up ahead with all the highly anticipated things like the SpaceX IPO and the upcoming OpenAI IPO and those outlandish valuations that accompany these IPOs. Arthur, welcome back to the show.
>> Thanks for having me.
>> Thanks, Arthur. Arthur, let's look at this AI orgy around uh these mega IPOs um and whether it's kind of like Dreamworld type stuff. I read through a little bit of the the the the SpaceX filing and it was all speculation about uh internet in space uh asteroid mining all these different things. And then also you look at some of these large language models uh you know 90% of usership is basically uh not uh paid subscribers. So if these tech giants are heading toward what might look like a digestion disaster for the market uh and alongside uh Trump in America and Xiinping in China being sold by all the tech bros at kind of a fairy tale that their national security relies upon this insane capital inflows into these AI companies. So my my question to you is like when do you think this whole bubble might pop?
So I don't think it's we're anywhere close to the end of this as you said capex orgy and AI you know Nvidia earnings were out last night and the beat across the board Jensen talked about the massive upsurge in inference the amount of money spent on inference the growth of the AI agentic economy uh someone sent me a news clip about from the FT detailing how anthropic is now profitable right like you know they were burning billions of dollars up to a few months ago and now on a you know go forward basis Anthropic is now profitable because of the surge of demand for inference due to people using um agentic solutions within their business and personal workflows and so the revenues are there okay maybe you're paying a little bit too much for the forwarders but it's not as if the the high-f flyers are cash flow negative companies that's you know apart from open AAI frappic which has the weeding models and the most popular models out there right now is a profitable company. And as the price of intelligence declines, the usage increases, right? The more the more that data centers that are built, the more chips that are sold, the cheaper it becomes to run these models. You just get more complex models and you do more complex reasoning with these AI. You replace more knowledge work within human civilization and demand keeps growing. But I don't think we're anywhere close to the saturation point of uh an AI capex build. If if anything, this is proving that you need to build even more capex, right? Because at the end of the day, you know, if you you need to sell to anthropic, you need to sell to OpenAI, you need to sell to uh Gemini and XAI and all these sorts of players and the demand is only increasing. I think you know Meta is firing another what 8,000 workers. I think Zuckerberg announced overnight. I read, you know, Jamie Diamond at JP Morgan is saying he's going to hire more AI specialists and bankers, which basically means the financial services houses are getting into the adding agentic AI into their own workflows. And you can imagine the amount of demand that there is for compute when you talk about high frequency trading. I think I read uh Hutchin River Trading has now their own cluster. Uh I forgot which data center again. So demand is rising as the price of intelligence declines. We're using more of it. I don't think we're anywhere close to the saturation point for this capex buildup. And so I think that while there was a brief dip in the market as people were expect you know looking for Nvidia earnings and you know reading which tweet from Trump or the IRGC and sort of getting a little bit of cold feet I think that is over and it's back to the original plan of of only for a lot of these AI stocks especially if you talk about some of the smaller cap companies that are in the supply chain of building chips building data centers who is going to have more inference requests as demand grows.
I think that's interesting actually talking about the smaller cap companies that are in the mix because not all these companies are equal like anthropic I think you know they their kind of narrative that I sort of sense around that is vibe coding and just basically uh uh you know agentic coding and things like that Apple is building AI directly into their hardware. Google are sort of embedding AI into all of their different products. But but you know what is the narrative for say open AI? It's doesn't have hardware. It's just sort of like software. What is the narrative say for XAI and things like that? Do you think there's going to be some sort of bottleneck when it comes to the plethora of the big mega cap companies out there?
>> I mean I I honestly it doesn't matter. I don't think you can't right now. None of us can invest in open AI or anthropic right on a liquid basis. Yes, there's all these secondaries floating around. Maybe they're legit, maybe they're not. But if you talk, if you think about it, like as an investor, what can you invest in? You cannot invest in these front-ier models, it's not clear to me whether or not they are going to be the ultimate winners. Um, obviously, if you move more to the source of intelligence generation, right, on the the hardware side, yes, there are clear winners. Nvidia, Intel, TSMC, right? Like it's pretty clear that these guys are are doing things and you can't replace them. But on the anthropic side, like if your frontier model drops behind a few percentage points in effectiveness, then the corporate client switch, right? And then all of a sudden all that money you spend is worthless. And so this leads to the, you know, I need to train the next best frontier model. I need to spend the another 100 billion, 150 billion, whatever it costs to train one of these models that continues to drive the capex cycle because you cannot fall behind. If you're in the game of providing a front-term model, you cannot fall behind unless you're going to go the Chinese way which says I'm going to undercut price, open source everything, I'm going to make money in a different way on volume, right? So, you know, you have to choose your poison.
>> Yeah, that's interesting. Yeah, you can do the bargain basement like the Chinese one or else you're up there and it's just a constant uh 100 meter sprint basically for all those companies. Now you you mentioned recently in your newsletter which I have to like commend you on like they're always good but that one like that so many points in in it that I I'd love to discuss now with the red queen effect that I'd never heard of before but like forcing companies into this endless hyper expensive race with AI where you know the tech that they're developing rapidly becomes obsolete. So, in terms of that, like how much real long-term economic productivity do you think will actually survive uh once these massive capital expenditure cycle cycles potentially start to cool off? Because, you know, can it really sustain itself uh you know, in terms of this rapid expensive constantly trying to keep ahead?
>> I mean, if you think in a very long-term perspective, not to say that this means that you need to buy every single AI stock, right? You know, our entire value that we've basically created since the mainframe computers is all intelligence work as a human civilization. So, think about the billions of people who do something related to manipulating symbolic things whether in their minds using a computer. If we could do all of that with AI and we can do much more of it than we could do with the stock of human intelligence that's around, imagine the demand for that. that doesn't necessarily mean that it's a great investment. I think you know there is val validity to the point that yes in a medium to longer term scenario there's an insatiable demand for intelligence to learn more about the you know our universe around us now that we have these super intelligent beings that can recursively improve themselves and what do we need to provide we need to provide you know energy to them to learn and to experience the universe and to give us answers to questions at the end of the day. So yes, I think in a longer term scenario, you know, it's a very big number in terms of how much larger this could grow. That doesn't necessarily mean it's a great investment though.
>> Yeah. Yeah. I know it's it's just, you know, the playing field just seems quite confusing at the moment. And also I'm just sort of u probably in the minds of lots of audience as well. It's kind of where's my job going in all of this? So in terms of you know this expansion of intelligence and you talked about the layoffs at different companies do you think there's going to be sort of like almost like like permanent early industrial revolution style layoffs of older industries and will people be able to find new types of employment in this new who knows what type of environment that's going to be created?
I'm sure there'll be new employment, but I think that it's a bit silly for people to claim that, oh, just you lose your job yesterday and tomorrow you get another job in AI, whatever that means, right? Talk to all the folks in Europe and America who lost their manufacturing jobs because of China and other Southeast Asian countries. How quickly did they find another manufacturing job? A lot of them didn't, right? A lot of them just got addicted to opioids and killed themselves essentially if you think about America. So I think that's a bit silly to say there's this creative destruction and all of a sudden all these workers are all going to get these magical AI jobs. In the short run there's going to be lots of uh dislocations and it's going to depend on where you're based and what the social contract is in the country that you're in. You think about the United States, right? If you're talking about knowledge workers getting replaced by AI, well knowledge workers are the most politically active people in the United States, whether it's a Republican or a Democrat, they're going to find a way to get protected. um they're the ones out in the streets. They're the ones who the mainstream media caters to. They're the ones who sell products to. And if they're saying that they're long no longer gainfully employed and it's the tech overlords that have taken their jobs through efficiency gains, I think there's going to be a political backlash in the United States and maybe they're going to be some sort of you know AI to you know labor social contract reform whatever that means. But if you think about places in, you know, near where I live in Southeast Asia, talk about India, Bangladesh, Philippines, these places. Um, these workers are essentially cough centers for American and Western European companies. Nobody cares if they starve or not. Oh, okay. I'm going to lay off my entire Filipino staff that was doing all this manual work that I didn't feel like automating because they were so cheap per hour. Well, now an AI can do all that. Goodbye. And nobody cares. They lose their jobs. Starvation, right? And so they and you know the Philippines and a lot of these countries don't have the the wherewithal to make a situation better. And so I think there's going to be massive social strife in a lot of these third world countries that are essentially back office processing centers for a lot of western companies. So again I think the the displacement and what that means for a worker is going to depend on where that worker is located.
No, it's it's interesting what you said there and compounding this kind of vulnerability especially in people in Southeast Asia and different countries is the fertilizer shortage as well coming out of uh the closure of the streets of Hormuz and combined with a super elino this year. Uh there's actually I was speaking uh my other half she works in an NGO and they're already trying to sort out kind of famine relief uh efforts in different countries around the world. So do you think like you touched upon something quite interesting there in in in Europe and America we're kind of not really living in the reality of what actually this all this means this this this you know the AI obviously the AI capital expenditure but also the political situation that's happening especially in the Middle East but
>> Sorry, what's your question?
>> Oh yeah. Sorry. So like were you sort of thinking yourself like that the this market exuberance for AI companies and stuff like that uh it's kind of drifting further away from the reality of what it's like in the ground especially you like you mentioned those countries in in uh Southeast Asia. Uh it's just got like no relevance or you know they're not even thinking basically about all these the strife that could evolve from all of these changes.
I mean, if you're in a western advanced economy and you know, you have enough food to eat, you have place shelter for your head, and even if you lose your job, there's some sort of government assistance, purely academic, the suffering that exists thousands of miles away, somebody you've never met, will never meet in your life, you have no friends that have a connection to that area. And so when you talk about there might be a fertilizer shortage or famine or a bunch of, you know, young men without jobs who used to work at a call center, it's like, okay, that's really bad, but I'm going to go continue to get my Frappuccino latte for $15, right? No one cares. We're human. That's not the point of view. Unless you live there, it doesn't matter. And the financial markets certainly don't care.
>> Um because, you know, the financial markets are geared towards big techi, whether that's in the United States or China. Who cares if a million Bangladeshies who lost their job are rioting in the streets whether you know with the capital of Bangladesh burns to the ground no one cares.
>> Yeah. I suppose this Adam Smith's the invisible hand of capital doesn't care. It it's not really conscious in that in that respect. Uh let's can we bring it into like crypto assets digital asset space? So you've made the strong case in all of this that's happening around the world in terms of uh you know the the the strife in the Middle East and things like that of Bitcoin's explosive trajectory this year. Now you're also eyeing uh privacy narratives especially around tokens like uh Zcash. So given the ongoing regulatory crackdowns uh and d-listings is happening to these privacy privacy focused protocols, are you still confident in uh your narrative around uh privacy tokens or do you think they're going to suffer from regulatory um input?
>> Regulatory input means nothing to these coins. If if Zcash needed a positive regulatory environment to survive, Zcash is worth nothing. The whole point is that it exists outside the system that provides a useful service to those who don't want to be surveiled by big tech, big government, and to soon to be agentic AI agents crawling around uh the internet. So, the fact that Zcash and Monero and a lot of these coins are getting delisted proves that they're valuable because if they didn't do if if they're privacy coin that is listed in a place that bans privacy coins and obviously it doesn't work. So yeah, I mean it's a bit counterintuitive, but if anything that validates the thesis that Zcash does something valuable.
>> Yeah. Yeah, I suppose you're right. Like it's it's kind of that's actually it's res on detra. So uh now you've been kind of like talking about how sovereign nations are shifting away from this. You mentioned like from a fugazi of US treasuries to secure physical commodities instead uh like dumping uh sovereign bonds and things like that. And we've seen bond yields like in the UK like just almost gone parabolic. So if the world is genuinely beginning to unwire itself from say PAX America, do you think the Fed's dollar swap lines can actually do enough heavy lifting to prevent any kind of serious systemic financial crisis in the medium to long term?
>> I mean the Fed can take collateral or not take collateral and print as many dollars as it wants. So if you are a friend of the United States like Japan for example or the UAE I don't know if the UAE has gotten a flop line yet but let's take the Japan for example and you know you could sell treasury bonds or equities in the cash markets and that would obviously have a bad impact on the markets or you could swap them with the you could swap dollars with the Fed say oh I need dollars to buy crude oil or jet fuel or whatever it is that you need to buy instead of selling this stuff the Fed says no don't sell it either you know pledge it as collateral and we'll give you we'll give you cash or gives borrow dollars from us and give us yen uh as as collateral right and these are things that are already in place and this can be done and I'm sure this is part of the conversation that Buffalo Bill Besson has with his counterpart at the Ministry of Finance in Japan when he is basically telling them stop selling all these bonds there's another way to do this to achieve the same result and the United States is perfectly happy growing its balance sheet to make sure that the financial asset markets don't suffer so I think you know that is definitely a way that they can do them. They can do infinite amount because they have a prying for us.
>> Lastly, Arthur, did you follow much of the meetings between Putin and Xi Jinping uh earlier this week?
>> I saw the photos.
>> It was some sort of like something came out an announcement about uh a new strategy to create I suppose land trade routes rather than maritime trade routes because of I suppose US Navy being kind of Pax America controls the waves type of thing. Where where do you see that kind of going this interrussia China be you know belt and road initiative type trade and they're trying to denominate more of this in ren minby do you think this is going to have any kind of serious impact for dollar dominance and US dominance across the globe at
>> the end of the day China is the largest trading partner for a lot of countries around the world and so while I don't believe that China wants the yuan to be the global reserve currency uh they do want to put in place the option if something happens between China and America in terms of the geopolitics where countries can settle in yuan they have the I forgot I think I forgot the acronym for the system that they have but they have a sort of a swift alternative that runs in the the Chinese banking rails and that while on an absolute basis it's still a small percentage of global trade flows in terms of payments the rise especially since the war in Iran has been quite marketked
>> off of a low base obviously So again, it's not that China wants to replace the dollar because if you become the global reserve currency, you kind of have to run the same policy that America has and you see the result, right? A de-industrialization, hyper inequality, social strife, you know, essentially a country that's almost at civil war with each other. Um, you know, the politicians try to claim it's cultural wars or other things, but essentially it's you have to print the dollars to go global trade. That's the deal. If you are any in your cap your your capital markets must be open so foreigners have to be able to buy these assets and that creates a lot of issues politically. I don't think China wants to to go down that route especially given the amount of debt that they already carry in their domestic currency. They do not want to open up their capital markets to the rest of the world. So anyone who claims that China wants to be America in terms of global reserve currency doesn't really understand what China is doing. Yes. Are they trying to diversify the way in which they get their calories and get their hydrocarbons whether it's through you know belt and road initiatives over land or the string of pearls the different bases that they have around the Indian Ocean and the South China Sea like absolutely that's um
>> you know the communist party's job is to sort of make sure that people are well fed and have the energy that they need but again I don't think that this means that they want to replace the US dollar if Trump wants to sanction China and try to remove them from the pay global payment system they have an alternative the trading partners have experienced using this alternative and therefore they can seamlessly switch. I think this is a preparedness measure rather than saying China wants to replace the dollar have the yuan and replace the dollar as a global reserve currency.
>> Well, look really astute again. Thank you very much Arthur for coming back on the show. It's always great to chat to you.