Scott Bessent Could Change Everything w/ Andreas Steno

17:23Andreas Steno:I've been banging the drum on this short dollar trade since mid-July. It's starting to work. We've also seen spillovers to gold and Bitcoin and other such short dollar bets. And I still think there's more juice in that trade. And the issue, and interestingly, I think less than 24 hours after this is released, we'll get the Jackson Hole speech by Kevin Walsh, right? We're yet to see the Fed truly playing ball with Scott Besant on all of this, in my opinion. Kevin Walsh has been radio silent on everything related to productivity and all of that since he actually took office, right? He toured with that message throughout the whole audition period.

18:07Andreas Steno:But he's been very silent on everything, basically, since taking office. But if you look at the theme for Jackson Hole, I think it's called financial innovation and how it impacts payments and inflation and policy, something like that. To me, that sounds like AI, blockchain, stablecoins. Stablecoins clearly is another big demand.

18:29Raoul Pal:That's why Besson wants stablecoins at scale. He thinks it can get to$3 trillion. Well, that's$3 trillion of bills that can be held by the stablecoins. Exactly. But maybe it's also an opportunity for Walsh to talk about productivity, right? That's what I think is coming here. He has to talk about, because he's got a job, Besson's got a job, and it's kind of an agreement, as we said. And his job is to say productivity is going to rescue us, and we need to reframe inflation to something that works for the modern age. And beneath that means we can have lower front-end rates. and maybe he'll give us the first few hints into the work groups that look at the current inflation

19:16Andreas Steno:measurements and how to update them to the modern world and all of that i think you're absolutely right that they're trying to find a new way of measuring it also a much more live way of doing it so that we don't have to wait you know six seven weeks for lacked information that moves the market and all that i actually think it's a pretty good vision that he's laid out um but we need the details and friday is obviously a great occasion to deliver the first few details yeah i

19:42Raoul Pal:mean they've got the committees going out to actually do it but he needs to talk about why he's doing it um and not to say i want to lower rates but it's like you know we need to update the system and this is the way of doing it so yeah i think it makes a lot of sense to see what Walsh is going to do. I think everybody's got a role to play here, and it's going to be really interesting to see what plays out.

20:07Andreas Steno:But admittedly, Raoul, say, a week and a half ago, when Besant announced this Treasury buyback change and all of that, I found it to be material, but not something that would necessarily alter the asset market trajectory a whole lot, but oh boy, we've seen a rotation since, right? I mean, we've seen a rotation out of CapEx trades into duration trades. So everything that needs compressed yield curves and all of that to work. Gold has done well. Bitcoin has done tremendously well since this announcement. And all of the AI hardware bets have been sort of since, right? Because I don't necessarily think that those two trades go hand in hand as they did prior to all of these yield curve gymnastics.

21:02And I've also been fairly vocal about that over the past couple of quarters that the CapEx trade, all of the hardware names and AI,

21:14Andreas Steno:they need a steep yield curve, but they're also okay with the bear steepening yield curve, the stuff that we've seen for a while. That's not a good environment for software and Bitcoin and gold and all of that. So, of course, in the scenario that you laid out where Kevin Walsh ends up codding interest rates while Scott Besson is sort of in charge of the long run, both will work in tandem, right? You could both buy hardware and you can buy Bitcoin. But now that we're kind of in between stools here in a sense where either it's kind of a compression scenario for interest rates, or it's a scenario where the yield curve will steepen in a more aggressive way, in a more hostile way.

21:53Andreas Steno:I think it's a bit more either or, CapEx or Bitcoin in those two scenarios. But I'm personally positioned for the tandem scenario where both can work in tandem because of a bull steepening yield curve where interest rates move lower from the front and outwards.

22:09Raoul Pal:RAOUL PAL Yeah, I agree. I do think there was a catch-up trade. I wrote an article called The Great Rotation in GMI about this and saying, look, the front end of all of this is the semis, and they're probably going to slow down because the next tier goes up, stuff like the applications less. So I think about healthcare and biotech and stuff like that. And that's really worked well. And crypto was the really undervalued part of that whole kind of tech ecosystem. And it's likely to catch up the moment liquidity starts returning. Now, liquidity is not fully returned yet because we've just not seen the banks taking on the treasuries in any size yet and stuff like that.

22:49Raoul Pal:But that's when they're trying to orchestrate. So the market's sniffing that out. The financial conditions are easing again. RAOUL PAL Yeah, they are. And to some degree, I think a deal that solves the lack of oil products coming out of the Strait of Hormuz

23:08Andreas Steno:unlocks all of this during the autumn, right?

23:11Raoul Pal:RAOUL PAL Totally agree. And that feels like they've got to do something in the next two or three weeks before the electioneering starts for the midterms. So a quick break in your regular programming. If you're serious about your future, grab my free report called Prepare for 2030. I think you've got five years to make as much money as possible. And this guide will help you navigate what's coming. The link is in the description. Download it now.

23:38Andreas Steno:because we've seen a wave of disinflation coming from the flat oil market, but it's just not disinflation at the pump yet because of the very, very elevated crack spreads. If you look at the diesel price, it's still very elevated. If you look at the gasoline price, it's also higher than it should be given where oil trades. The reason is that we have a lot of refiners capacity locked in in the Middle East because of a lack of shipping transits. So by the end of the day, they also need a solution to the Strait of Amoose to really unlock all of this. But if we get that, and I could easily say that we'll have maybe three, four months in a row with negative CPI prints of the month.

24:23Andreas Steno:And that will really allow the Fed Reserve to do something. RAOUL PAL Yeah, and they've got to be aware of it.

24:30Raoul Pal:It's gone on longer than they expected, for sure. I think they were very close to a deal. I don't know what the hell fell apart. Something fell apart. I think it's old regime versus new regime in Iran. And now Besson's got his nuclear war from the Treasury on them. I don't know if that's going to work or not. It never stopped the Russians. I don't see what's going to stop the Iranians. But in the end, they have to find a solution here. They have to walk away from this and get the sort of deal that they want out of it. Yeah. So, you know, my take on it is that this first 30, 60 day memorandum about outstanding that allowed for, you know, a decent amount of exports out of the Strait of Moose kind of allowed both the U.S.

25:16Andreas Steno:and the Iranians some more patience in these negotiations. And that was probably why we ended up in a round two. The Republican Guard in Iran, their income source is basically exports to China through the strait, right? So it bought them a couple of quarters of time at max, the income that they could take in during that first Mirando von der Stenberg, because they're currently not exporting anything at all, more or less, right? So that's also why both sides do have some strategic patience right now, but they also need a solution, both sides, right, by the end of the day, because the Iranians cannot, essentially, they cannot fund their operations.

25:58Andreas Steno:And even though the Republican card is very ideologically founded, also religiously founded, they also need to be able to pay their soldiers, right? and they cannot keep paying their soldiers with a closed strait. We also know that all countries surrounding Iran are building solutions to avoid the strait now. So obviously, they also need to strike a deal at some point because otherwise the strait will become irrelevant. So I simply cannot, I look at this through a very cynical game, theoretical setup. And I simply cannot imagine a scenario where they don't end up striking a deal

26:39Raoul Pal:because both sides need a deal. Exactly. I just think there's a very high chance. And China wants a deal. Everybody wants this deal to happen because it removes the last choke point in the global oil market. And therefore, oil just doesn't become a geopolitical issue anymore. It just becomes straight supply and demand. And it all goes away. And OPEC kind of falls apart. and everything just clears. Venezuelan oil starts coming to market. You get decent refiners to be able to clean up the heavy crude. And before you know it, oil doesn't become an issue for the markets ever again. And that is a very likely scenario.

27:14Raoul Pal:And it's incredibly desirable for everybody. Yes. Yeah, so I've got that. Here's another interesting thing that I looked at, is I've been thinking through this productivity thing, and the data's slow, and it's not great. and it dawned on me recently so what are we trying to look for right we're trying to look for the application of ai and robotics on an economy and it dawned on me that there is a perfect example of this that has led since about 2018 was amazon so amazon have been the forefront of everything. So they have now almost more robots than humans that will cross next year. So they've scaled like an economy, right?

28:04Raoul Pal:The size of an economy as well. I'm a gigantic company. They've implemented AI. They've got self-driving fleets. They've got drones. They've got, I mean, every component part of the exponential age that you can possibly install. And they're running it. So I went back and looked at it in a big article I did for GMI. And what you saw is that profit margins troughed. They put their big investment into all of this. And out the other side is their margins have just been increasing all of the time, which is productivity, essentially. And I thought it was just fascinating to say, well, look, there's an obvious example that it's going to work.

28:44Raoul Pal:And people don't believe it yet. No, they don't.

28:47Andreas Steno:And of course, if you look at some of the official productivity statistics, which I, by the way, don't really buy, it doesn't look like we've had a revolution. But if you look at the micro cases, we've clearly had a revolution already. We have record high profit margins everywhere. It's not even only in technology. and yet the statistical bureau tells us that is that it has nothing to do with productivity

29:19Raoul Pal:i don't really buy that because and considering we've had the economy's not been growing super fast household spending has not been great so where the hell are these profits coming from

29:32Andreas Steno:and there has to be productivity has to be it has to be i i perfectly agree with that the the thing that probably also wrong-footed a couple of the IPO pitches from the big LLM companies is that we haven't seen a big impact across the labor market yet. But my point is, and you have decades of evidence for this, is that when profits go up as aggressively as they do now, you don't see layoffs, even though you could lay off people without probably harming your business. you'll see them once the profit cycle rolls over. And that's where you'll harvest the productivity for the next cycle.

30:20Raoul Pal:One thing we have noticed is that people won't hire us aggressively. So the labor market doesn't expand in line with GDP or the business cycle. But you don't see that divergence. Because as you say, people are making record profits. Yeah, sure, some of the tech companies have laid off a lot of people. But overall, people won't do it until they're forced to do it. And then they'll really dig in. And that's where productivity almost accelerates. Once you get an economic downturn, everyone realizes you can operate with half the number of staff.

30:52Andreas Steno:But there was an amazing interview with the business line manager in Scandinavia. And he said that, well, he's basically instructed all of his employees to optimize their tasks with AI. And he aggressively just said, use AI, use AI, use AI every time they had a new task. But then he got the question, why haven't you laid off your employees if they have automated everything? and it took him a little while but then he kind of indirectly admitted to well it's actually a pretty nice thing to be a manager for like 20 people right so and and i think that's exactly why you need to force those outcomes top down in organizations because a line manager he's not incentivized to lay off people even though they've automated a lot of stuff it will only happen once the CEO is forced to talk down.

31:53Raoul Pal:Managing people is power and ego. In the end, I mean, you should be, if you're ruthless, want to manage the least number of people possible. But then you don't feel powerful because what people like to say is, oh, yes, I have 500 reports in my business area, as opposed to saying I've got 5 ,000 agents reporting to my AI and I'm orchestrating that at top level.

32:17Andreas Steno:people haven't got to that yet they will do yeah at some point i guess you can um brag about managing 5 000 agents in the bar when you're trying to hook up in town right uh i there is a big you know generational cliff uh in that question i guess because i guess millennials and gen c's they'll they'll end up bragging about how many agents they have quite pretty soon right um But I also see the same thing related to digital art versus physical art, right? I've had quite a few employees who, you know, when they were like five, seven units of alcohol deep, they started showing me their weird coins in their wallets on their phones, right?

33:02Andreas Steno:As some sort of art, right? See, I have this shit coin, right? And they kind of loved talking to me about that. I'm even too old for that. You know, I would never do that.

33:13Raoul Pal:One of my biggest investments that I have is my digital art collection. Yes. So even though I'm not quite a boomer, but I'm Gen X, but I've gone down that rabbit hole deeply. But yes, I mean, digital things can have value in this world. And yeah, other people just don't get it yet. No. You're right. So yeah, it's going to be fascinating to see how companies adapt to all of this and how the economy adapts to this. The other thing is, I don't know if we measure it in GDP.

33:49Yeah.

33:49Raoul Pal:Because, I mean, if we look at the sort of, if you've got record profits and they're accelerating, you've got record CapEx, you've got sort of records of everything, yet the economy's not really growing, which is not measuring something.

34:08Andreas Steno:but it's a really good question right because if if you hire a bureaucrat it's very easy to measure that in gdp terms right because you pay that bureaucrat a salary and that pay slip will directly go into the gdp calculation as his output is or his output um how do you measure an agent I don't think it's easy. And I'm not sure I have the answer. But look at the statistic put out by OpenAI. The amount of agents live in the OpenAI ecosystem has quadrupled, I think, over the course of two months.

34:54Raoul Pal:So something is clearly happening there. RAOUL PAL I mean, AI bots are now by far and away the largest part of the internet. That is not far and away the largest part of the payment system, but they are creating economic value, you and I know, because we run them.

35:12Andreas Steno:And they're creating huge economic value for us, but it's not captured. RAOUL PAL And they're even writing Stan Drunkenmiller's opinion letters now. They also write part of my stuff. I perfectly admit to that. Why shouldn't I use the technology? I get that feedback in my comments. Yes, why have you used Claude to write this? Well, because then I can write about five topics for you instead of one. I mean, why wouldn't I?

35:40Raoul Pal:Yeah, it's not like you don't put your thinking aside to them. You still do your thinking, but they do the writing and the other stuff or putting together a bunch of thoughts that you put. They put them together in a coherent way. that's just efficiency you're not you're not outsourcing your thinking because they can't they're not very good in that respect it's very good if you want them to be a partner

36:04Andreas Steno:it's like you know this is not my first language right when i write an article in english it's probably a good idea that i allow clock to work with them and improve it and all of that right it's like telling me that i can't use addiction when i'm not a fluent it's just nonsense when's the last time you did long division on a piece of paper.

36:27Raoul Pal:Nobody does anymore. It's like transitioning from pens to typewriters. Oh, well, that's cheating. I was like, no, it's just technology. It's just how it works. But people are very weird about it all. So let's, because we haven't got long to go, but what's your view for the US economy now going forwards in terms of the cycle, liquidity, the things that matter to markets? Because this is what everybody ever wants to know is when's it going to stop? What's the risk here? What are you looking at?

37:01Andreas Steno:So if you look at a little bit more old school business cycle models, my best guess is that the PMI cycle will peak probably in November or December, or something like that. But it doesn't mean that 2027 will be a negative growth year. I think it would be slightly more sluggish from a rate of change perspective, but I don't think that we've seen sufficient tightening for me to really worry about the cycle into next year. And everything Besant has done over the past couple of weeks, we obviously still need to see the details in the coming weeks, tells me that they're trying to prolong the cycle throughout, Trump's remaining period here.

37:47Raoul Pal:Rolling bills at the short, issuing bills at the short end, that's more ongoing liquidity. The changes, as you say, sound like these are not short-term fixes. These are longer-term extensions of the business cycle. We've got the largest CapEx boom in history, and half this stuff, no, three-quarters of the stuff hasn't been built yet. So I'm struggling to see how the cycle can stop. and we're not seeing inflation so it's not like they're going to jack up rates in the middle of this in any meaningful way so to me this looks just like the 90s and we saw ism go up a lot it cooled off a bit it then went up again it cooled off a bit then went up again before the final peak feels like this is what's going to play here i don't see it playing any differently

38:35Andreas Steno:The funny thing is that if you look at forward expectations for many of the hardware companies, for example, everything related to the CapEx build-out, it seems like most people agree that 2027 will be slightly bigger than this year. And then if you look at 2028 and 2029, it's just a flat line from 2027. And that is the only scenario that won't happen, right? It's not like we'll just plateau and then, okay, this was the exact amount of spending needed for the economy. Either we accelerate or decelerate. And I still think it's vastly underpriced that we actually accelerate in nominal terms, which is the point I've been making basically two or three months in a row.

39:20Andreas Steno:Yes, we will peak in rate of change terms this quarter. It's simple math. I mean, we cannot grow at 200 % forever in CapEx terms, because that would lead to numbers that even you and I could not imagine, right? So obviously, that cannot be the case. So even if we grow CapEx 25 % from this base, it's massive in nominal dollars. And I think that's likely what's going to happen the next few years, that we'll grow maybe 15%, 20%, 25%, something like that, which is huge. I mean, it's a bigger CapEx increase in nominal terms than we've seen ever outside of the last 18 months, right? So it's still huge in nominal terms.

40:08Raoul Pal:And CapEx have been so lackluster over the last decade or two that the rates of change looks large anyway. but as you said it's actually going to be the nominal change in the end that's going to be bigger than people expect and i just don't see what's interesting is everybody all the naysayers on the capex boom just like this all circular financing this is all nonsense and then you listen to and you know all the depreciation of this stuff and then you hear people like andy jassy amazon saying no we just keep using the same chips we need more and more and more and This is never going to stop. And all of them are saying, these guys aren't smart enough not to need to hype cycle.

40:50Raoul Pal:I can understand that Anthropic and OpenAI need a hype cycle, but these guys don't. No, they don't need a hype cycle.

40:56Andreas Steno:And by the way, if you look at the latest investor deck that was leaked ahead of the IPO from Anthropic, they're close to life. I think they're close to$70 billion in annual recurring revenue. and 15 months ago, they barely had a client. That is a revenue journey that is second to none in history. Also, if you look at it in inflation, just the terms of it, it's the single biggest revenue machine ever seen. And it will far outpace what we've seen from Google, Microsoft, et cetera, when they were on their initial journey. It is incredibly fast. So sure, do they need something like$250 billion,$300 billion, something like that, to be able to pay for their future compute capacity?

41:47Andreas Steno:Sure. But it's a company that went from zero to$70 billion in a year. And all they say is we are compute constrained.

41:57Raoul Pal:So what they're telling you is the demand for intelligence is outstripping the ability to provide that intelligence. RAOUL PAL And you've highlighted that as well. RAOUL PAL Meaning the air capacity, there's not enough, there's still not enough data centers, there's not enough power, there's not enough. RAOUL PAL Look at the only subcategory in Anthropics

42:23Andreas Steno:accounts that looks a little bit weak from a rate of change perspective right now is called code revenue. But it happened when they became constrained. So it's the exact reason why they lost momentum. And there's a reason why they paid SpaceX to 3x what everybody else pays for compute in this panic deal ahead of the summer, because they were simply too laid back on their compute needs.

42:48Raoul Pal:Because you heard Dario talking about this, and he was boasting like, we're being really sensible at how we think about the compute. And OpenAI, we're just like, we'll take all the compute. And then suddenly, Anthropic are like, fuck, we've got the world's hottest product that's ever been, and we can't supply it.

43:06Andreas Steno:Yes. And for that simple reason, I'm tempted to say that I'm more interested in investing in the OpenAI IPO than Anthropic right now, because they have a super compute pipeline. And Anthropic, they're still behind on that question, clearly behind.

43:22Raoul Pal:Well, you can see it in the models. OpenAI, because they've got enough inference, will deliver an answer much faster than Claude will. And you're not getting rate limited and everything in the rate that you are with Anthropic because they just don't have the compute.

43:37Andreas Steno:Yeah, and I think that's a bigger deal than what most people anticipated six months ago because no one kind of dared to extrapolate that wave of demand into the future. But I still think it's very fair to extrapolate it.

43:50Raoul Pal:My case case is the demand for intelligence is the largest of any single product ever produced, and it will never stop. Because intelligence is going to go into every single thing. People don't realize it. And that chart of the average token price that you first showed me, and we looked at it in detail and realized, you know what? All this is going to show is Jeevon's paradox. All it's going to show is you can bring in as much cheap intelligence as you want, but the entire pie is going to grow so fast that everybody can hold their margins. And that's proven true. I mean, the frontier models are holding margin against all of these cheap competitors because they're not competitors because the pie is growing too fast that everyone can provide intelligence at different prices.

44:40RAOUL PAL, JR.: How long did it take for the Kibi company in China to tell customers to stay away?

44:50Andreas Steno:I mean, no, no, stop, stop, stop, stop. We don't have any capacity for you. It took like

44:54Raoul Pal:12 hours, something like that. It was very fast. RAOUL PAL, JR.: I know. Anyway, my friend, good to see you. Let's see how this plays out, because I'm with you. I still remain steadfast bullish here. September's in mid-election years, and never great. Maybe there's a hiccup on route, but it feels that it's all systems go.

45:16Andreas Steno:I think the best guess would be that we rally hard into year-end after the elections, and then we have a better 2027 than most people project.

45:25Raoul Pal:Yeah, I think that's right.

45:27Andreas Steno:Fantastic. All right, my friend. I will see you soon. Take care.

45:31Raoul Pal:So great conversation with Andres, as always. And you can see we share this commonality of ideas of the potential for an elongated cycle with a few ups and downs en route, that this whole CapEx game is not over yet, that intelligence is scaling faster than anybody understood it to happen, that the Fed and Besant are going to start trying to change the rulebook to enable all of this to keep going. So fascinating times. We'll keep our eye on everything. And on Real Vision, we've both got dashboards where you can track all of the things we're talking about in real time. so enjoy those and i'll see you next time you obviously enjoyed the episode because you're here with me at the end but listen don't forget to go to realvision.com forward slash join and grab a free membership it's an incredible community packed with alpha great investment ideas and the research that you need to help you unfuck your future so get started now go to realvision.com forward slash join side.

46:30Raoul Pal:... ...