TRANSKRYPCJA 🔗
00:00:02
Speaker 1: Bloomberg Audio Studios. Podcasts. Radio. News. Hello and welcome to another episode of the Odd Lodge Podcast. I'm Joe Wiesenthal.
00:00:22
Speaker 2: And I'm Tracy Alloway.
00:00:24
Speaker 1: Tracy, we're recording this November 12th. And basically in the last hour, it is confirmed Jensen Wong out at NVIDIA. I know.
00:00:33
Speaker 2: So we were both sat in the newsroom when the headlines came out and you could hear the sort of like uproar from all our colleagues. And we're like, what?
00:00:40
Speaker 1: We heard the ooze and azz before the alert popped up on our screens. But there are certain sounds when you're in a newsroom for a long time. You're like, oh, there must have been a headline that just dropped.
00:00:50
Speaker 2: You know.
00:00:50
Speaker 1: And then you infer not only like you can. Get good. It's like, oh, it's probably an NVIDIA related headline. Yeah.
00:00:57
Speaker 2: And I mean, we've been talking about NVIDIA already because a lot of these, you know, rumors of scandal have been out there for a bit. But it's worth reiterating that we've described it previously as NVIDIA is the sun around which the whole AI ecosystem basically orbits.
00:01:13
Speaker 1: Right.
00:01:14
Speaker 2: Like not just through the chips themselves, but through the financings.
00:01:17
Speaker 1: Totally.
00:01:17
Speaker 2: The vendor deals, all of that. And so this is a huge moment for the market and also for geopolitics.
00:01:23
Speaker 1: And geopolitics. And that's the key thing. It's a huge moment at every level.
00:01:27
Speaker 3: Yeah.
00:01:28
Speaker 1: Even if you just confined it to like, well, AI is a technology that exists. Then you say, OK, this is a really big moment. Then you say, well, it's not just that this is a big, important company in AI. It's an important financial player within the AI company. And AI itself has not just contained itself because it is a major economic driver and maybe arguably because there's so much AI spending to like build boxes in which you can put NVIDIA chips in that contributes to inflation and contributes to scarcity. And then it's a geopolitical story because everyone around the world wants AI technology, et cetera. And so there's this dual use element. where it's a little bit like building out technology, but it's also like a little bit about building out a weapon. And that's very strange. So like shades of arms races and so forth. So then to have the CEO at the center of it all, admit to having an affair with an important buyer, but then also for the U.S. apparently to believe that there was a national security, that there was a spy angle. just got it all. It's literally everything you could possibly think of. It really does.
00:02:36
Speaker 2: And just on the geopolitics of it all, I mean, we've recorded episodes about what a U.S.-China war over Taiwan would possibly look like. When you read the way people game theory that scenario out, NVIDIA features really large, right? It's like, well, what would happen to the fabs in Taiwan? This is a physical asset that the world is extremely focused on.
00:03:01
Speaker 1: And the one other thing, you know, just speaking in context, is this all, you know, this is nine days after a huge election that happened. And the news that Jensen Wong was having an affair with the executive of Megaspeed had apparently been understood for quite a while. Almost certainly, you know, we'll figure out why it leaked now and hopefully we discover that, but probably not a coincidence that it happens right after the election and so forth. We know that there are divisions within the White House, for example, on things like chip export controls, etc., But it really just comes at an extraordinary time. We are still at war with Iran. We get headlines on.
00:03:44
Speaker 3: That every day.
00:03:45
Speaker 2: I was going to say, the other thing we should mention, this headline has gotten buried in the Jensen news, but Donald Trump saying that he doesn't need a deal. With Iran, that moved oil higher as well. So you have all these different things going on at the same time.
00:03:58
Speaker 1: So that contributes to the building inflationary pressure. We got a very hot CPI print this week that saw traders ratchet up their bets and more aggressive Fed hikes now. So there's just like so many different strains all kind of hitting at once. And granted, like, you know, the stock market itself is actually still fairly close to its all time highs. The AI trade specifically has got sort of hit, but by and large, you know, good year for the market, even as like all of these like pressure points are building up.
00:04:28
Speaker 3: Right.
00:04:29
Speaker 2: And all of them are in tension with each other. I think that's the key thing.
00:04:31
Speaker 1: This is the singularity.
00:04:32
Speaker 2: Yeah.
00:04:32
Speaker 1: This is actually what they meant when they talked about the singularity.
00:04:36
Speaker 2: That's right.
00:04:36
Speaker 1: They actually meant like a bunch of stuff would all happen at the same time, all in tension with each other. It's all one single story.
00:04:42
Speaker 3: That's right.
00:04:43
Speaker 1: Anyway. It's impossible to synthesize it all into one thing, but maybe we'll give it a try. We'll give it a try. I'm really excited to say you have the perfect guest, someone we've wanted to talk to for a while, interesting thinker on who could think at almost maybe all of the levels that we've hit. Tianyang, CIO, manager of the VPX Fund at Variant Perception, has written a lot about markets and also geopolitics and the stresses that are building markets. within the US and other Western democracies and so forth. So Tian, thank you so much for coming on Odd Lots.
00:05:16
Speaker 3: Yeah, thank you for having me.
00:05:18
Speaker 1: What a week, huh?
00:05:19
Speaker 3: Couldn't have timed it any better.
00:05:21
Speaker 1: Why? When you say, I mean, I agree. But when you
Load full transcript
think about the work that you do, like, I agree that we couldn't have timed it any better. But when you talk, when you think about the work that you do, talk to us how you see this Jensen Wong story sitting at the center of everything.
00:05:37
Speaker 3: Yeah, so I would say this is an anecdote within a broader primacy of sovereignty theme. And this has been an ongoing multi-year theme that investors have had to slowly but surely confront and realize. And it's a very distorting force on all asset markets. The basic idea is that we're in a US-China rivalry that's escalating. It's driving a lot of tensions. And ultimately, the needs of the sovereign come first, right? We've gone from an era of globalization where everybody's friends, you're not really worried about But sovereignty means you have to be able to say no to your rivals. And to say no, you need to not be super dependent. Hence, you have the mobilization on tech, on AI, on food, on energy. And you also need some domestic cohesion. When you put the two together, that's where the impossible trinity that we're going to talk about today comes from.
00:06:25
Speaker 1: Yeah, we got to get into this, your notion of the impossible trinity. By the way, headline, NVIDIA now down 14%. It's below $ 4 trillion. You know, I really do think it's like, I understand sovereignty is important, Tracy. I understand like, oh, this is important, but we can't have the stock market go down. Like that's, we can't, we can sacrifice a lot, but are we really going to let stocks fall?
00:06:45
Speaker 2: We should, we should maybe talk about how like all of US society is predicated on the line going up. But I'm going to save that for later. Let's talk, you mentioned the impossible Trinity. Let's talk about the other impossible Trinity. Like can the CEO of major global company, I guess, abide by US law while having an affair with an alleged Chinese spy? Like just on the surface. When you look at the Jensen headlines, what does that tell you about the state of the company?
00:07:12
Speaker 3: Well, so first of all, I think there are historical analogs. So what jumped out at me when this news came out was the actual Oppenheimer analogy. You had essentially this pivotal figure, the center, obviously, creating the nuclear bomb. Originally, you know, national hero, all that, right? Rock star. Obviously, it's been a while since the movie, but over time, people start to question.
00:07:31
Speaker 2: I fell asleep during the movie, by the way, so I don't remember any of this.
00:07:33
Speaker 1: You're going to have to, all right, spoiler time for Tracy. Okay.
00:07:36
Speaker 3: But essentially, later on, when it comes to not just atomic bomb, but the H-bomb, he starts to doubt it. There starts to be questions about, is it a communist sympathizer? You can see this overarching lens of sovereignty today. We don't know what Jensen divulged or not, but at the margin, these are the pressures that come through, right? There has to be total trust from the sovereign to strategic assets. And it's fairly clear that with the way we're going, this is nuclear bomb plus commercialization potential, right?
00:08:05
Speaker 1: Right, because it also, and it's not binary. It's not like, okay, we achieve a really good chip and now we're done investing. It's a race with no obvious finish line. Theoretically, there may literally never be a finish line. Let's talk about the impossible trinity as you see it. Let's go big picture and then we'll sort of drill down. What is the core thing that from time to time countries really have to wrestle with?
00:08:28
Speaker 2: Yeah.
00:08:28
Speaker 3: So as I started saying a bit earlier, it's really about sovereignty's idea of being able to say no. So you don't want to feel dependent. And what we've come from is that, you know, as we've seen with President Trump's trade war, there's a lot of hidden vulnerabilities in supply chains that countries might not have realized. So if you go all the way back to, say, Joe Biden with the IRA, the mobilization, right, when you start reindustrialization. I view all this as part of this theme of reducing dependence. But if you actually look at China, this has been like a decade-long process. China has been working on this since essentially the Syngenta-China deal all the way back in the 2010s because China was worried, hey, they were dependent on the West for food, right? So they said, you know, do some seed technology acquisition. But that was the last thing the West tolerated. SIFIA signed off, and then what happened? the ZTE sanctions, the Huawei sanctions, right? The ongoing things. So it has been a kind of a two-way thing where on the international front, it's not so much, you know, there's a friendly rivalry, but there's really this sense of acute vulnerability that the person can cut you off.
00:09:30
Speaker 2: Yeah.
00:09:30
Speaker 1: But what's the trinity? So, okay, there's sovereignty, but what's the trinity?
00:09:34
Speaker 3: So the trinity is this idea because you have this rivalry against foreign powers, you're spending on guns and your mobilization, that's locked in. It's non-negotiable. But at the same time, if you need to have sovereignty, you have to be, have domestic cohesion. And the problem is in the age of social media, we obviously live in a very polarized society. So the way you get around that is a lot of fiscal deficits. You just keep spending, keep doling out the butter because you don't really have time to resolve it. Then the tension becomes, if you're committed to your guns spending and you're committed to your butter spending, then history shows that typically the leg that gives way is sound money. And that's what's putting us into a very kind of 1960s vibe kind of roadmap. And you're really starting to see a lot of the kind of macro samples come through that.
00:10:18
Speaker 2: Okay, so talk to us about some previous, I guess, historical examples of governments having to
choose between, I guess, sovereignty, butter, and sound monetary policy.
00:10:29
Speaker 3: Yeah, so I'd say the big one right now would be the 60s with LBJ, when you had the Great Society, you had the Vietnam War. And when you rewind back to the beginning of the war, he essentially said, hey, we can pay for both. We don't need tax rises. And over time, the fiscal deficit expanded, the pressure built, and... Ultimately, there was a famous anecdote about him confronting the central bank chair. So again, you can see this is not new in history. Apparently, the then Fed chair, William McChesney Martin, hiked rates against his wishes. He called him to his Texas ranch, physically confronted him. And so I think that's the kind of the time we're in where we've seen these underlying tensions come through. And I think why it's very pertinent today was Back then, when you got all the way up to 1968, the London gold pool broke, right? So obviously, no one's really talking about the gold rally anymore. So, well, back then, gold was pegged, right? So it was pegged at $ 35, just globally. But over time, whilst that peg has been placed, slowly people have been reducing the foundation of sound money. So, for example, there was something called gold cover that was repealed, so you needed less gold to back your currency. And as your policies became more inflationary, countries like France started... selling down their dollars and removing gold from the gold pool. And so essentially what happened was that you had the separate, essentially free market price of gold that basically broke away from the peg. And this was 1968, right? In the grand scheme of things, it took until Nixon, 1970, 71 for Bretton Woods to end and for Obviously the gold peg to end. And so I think that's the echo I see with, if you go back to 2023, we've had this gold rally, right? Come through. You've had this persistent concern about fiscal deficits. I, you can't cut the butter, you can't cut entitlement yet. We are in this mobilization, right? On tech or military. So these pressures are coming through and you're getting the historical echoes of warnings from the market that something's going to, something has to give.
00:12:24
Speaker 2: Right.
00:12:25
Speaker 3: And finally you have a Iran war. you have these things that force a lot of simultaneous physical constraints to hit. And then suddenly it becomes like a big kind of inflation problem and we'll have to confront it.
00:12:35
Speaker 2: Right.
00:12:35
Speaker 1: And we got that hot CPI report this week. So like hopes that we were still on this like glide path downward seem to be fading. And this is not just a US problem. Like with this, like these challenges, like it's, we obviously, these are real stresses in the US, but like, I don't know, it seems kind of worse in Europe.
00:12:56
Speaker 3: Yeah, it's a global problem, right? And this is linked, again, you can't get away from this idea of sovereignty because ultimately we've been super used to this globalized era, just-in-time supply chain, everything, the system works itself out. The whole point of sovereignty is the market prices no longer function as a signal because it's about the physical movement of things, right? The price of money can be a rationing mechanism when everything can move, then move interest rates. But now if there's hard just borders, essentially physical borders blocking things, you need the price of things to be the rationing mechanism. And that's where we are.
00:13:44
Speaker 2: So when we talk about the potential for sound monetary policy to give way, and you bring up the late 1960s, early 1970s, I mean, everyone's going to think, well, this means rampant inflation. But it doesn't necessarily have to translate into that. There are other ways to deal with high debt. I wrote a piece, I guess, a couple months ago now in the Oddlots newsletter, which everyone should subscribe to, called The Other Way to Reduce Government Debt, which was all about financial repression, right? And we've seen hints of that already with, you know, banks being encouraged to like load up on more government bonds and things like that. Could we not just see a system where people are forced basically to hold government debt?
00:14:27
Speaker 1: Yeah, absolutely.
00:14:28
Speaker 3: And this is obviously the difference between the US, which has its own sovereign currency and say, you know, the issue France has right now, right? You can't print your own currency. But the tricky thing with financial repression is, yes, I think it works. The basic idea is, you know, run the inflation of 4% to 5%, right, trying to suppress interest rates, or amortize away the debt. But the angle maybe we don't talk about as much that's going to be uncomfortable for asset owners is the potential for capital controls. Because ultimately, if you need this to work, you can't just let the money flow out first. And that would obviously be, I don't think there's any story in which that isn't a, that's a bullish thing for markets, right? That's almost certainly a universal negative.
00:15:05
Speaker 1: The thought of capital controls in the U.S. is unthinkable, right? Yeah.
00:15:08
Speaker 3: Yeah, but I will say this is the slow creep. For example, like I said, we have a microcosm of how these things work, right? So Iran will start this year. China did diesel export bans. Russia did a diesel export ban. Very quickly, President Trump said, hey, maybe we need to do it here. Oh, yeah. So other countries do capital control still today. Who's to say as they do more?
00:15:27
Speaker 1: It's not like capital control in the sense of money, but it is just like this ongoing creep. And if we get like, and I guess if we get a... You know, a much tighter clamp down on chip exports. It would be the same thing. Yeah, it's the same, right? It's not money controls, but it's like if we're making it here, we're keeping it here.
00:15:47
Speaker 3: Yeah, it could be money
flow. It could be the flow of goods. But it's more this idea we are clearly moving away from the free movement of everything. And it is a trend. So all these little anecdotes fit into that trend, right?
00:16:01
Speaker 2: Weirdly, it could also be the more politically appealing choice, I guess, right? Because sovereignty, okay, so like governments aren't going to give up on the sovereignty issue because they framed it as this existential race. They can't give up on butter entitlement spending and social services and things like that because then everyone will get upset. But will the voting population, would they like would they get worked up over capital controls or like different leverage ratios for banks? Is that something people are actually going to vote on?
00:16:31
Speaker 1: Yeah.
00:16:32
Speaker 3: So I would say the banking stuff, I think there's some hope financial pressure works. So again, there's analogs to the 60s. Back then you had Regulation Q, which capped interest rates you could pay.
00:16:41
Speaker 2: We did an episode on Reg Q.
00:16:43
Speaker 3: Yeah. So there you go.
00:16:44
Speaker 1: Right.
00:16:44
Speaker 3: So in a sense, if you look at stable coins, it's like, hey, you can't pay interest. You have hints of that. You had Operation Twist in the 60s. Today, we obviously issue a lot more bills rather than long duration treasuries. You mentioned obviously SLR reform, free bank balance sheets. So there are at least the tools are in place. And obviously as the politics shifts, as say, once again, you get a change election, unite, or maybe all the parts of the government become unified in 28, you can see the tools of government being used to do that.
00:17:15
Speaker 2: I hadn't thought about stable coins as like a tool of financial repression, but you're right. Like they are a huge source of demand for U.S. treasuries.
00:17:22
Speaker 1: So the reason why people like independent central banks is they say that democratically elected officials will never pull away the butter, right? This is like, and so you need like this outside thing to just clamp it all down. What does history say about that? Like there have been candidates who have run on things like spending reduction. It's not completely unheard of. And as recently as like, you know, The early post-GFC, arguably the worst time in history to do spending cuts. That was actually a pretty popular political movement to cut government spending. That seems like ages ago. Could you see a political appetite for curbing spending reemerging?
00:18:03
Speaker 3: So, yeah, there's been signs, right? Like, technically, President Trump ran on, you know, Treasury Secretary Besson had the 333 plan, right? They tried Doge. You don't have the America's Party when they fell out. It was a lot of things about the deficit. So there are echoes. So, you know, I've laid out that potentially the early 90s with Ross Perot, a very strong third party emergence, it's possible. But the problem is
you usually need a big crisis first for that to actually be implemented. It's one thing to say something. You know, Marine Le Pen is talking about a deficit in France. Who would have thought? But if you look at the details, that's what matters. You need the kind of population to be willing to accept pain to do that. So if you look at Reagan and Thatcher, they did do it. But that was after some horrific events. crisis in the 70s, right? And so I think that's the problem. Yes, we can get there, but things probably got to get a lot worse.
00:18:51
Speaker 1: Well, it could like, we just have a thing where young people are like, we can't buy a house because mortgage rates are so high and we need to relieve some of the pressure. So there's a movement to, you know, cut spending on various things to reduce inflationary pressure in the economy.
00:19:07
Speaker 3: I think that would be easier if we didn't have this simultaneous guns and butter pressure, right? I think the foreign aspect is of it means that, you know, you have very limited fiscal resources. I think that's the tension. So, you know, if I look at Peter Turchin's work, some of these long cycles of history, the basic idea is if you can have your economic elites voluntarily give up something to kind of support the bottom, that's how you get a peaceful reform and we transition. You can see there's token signs. Elon talking about giving away SpaceX shares. We've had Anthropic and companies talking about giving the public a stake. I think those are the marginal signs, but to really get there you know, for the magnitude of kind of redistribution we're talking about, I think that that's the challenge, right? And I suspect you need a crisis first before we can get there.
00:19:53
Speaker 2: Wait, can you say a little bit more? Because this was going to be my next question. Like, if the problem is physical binding constraints, then why wouldn't it, I guess, dampening public consumption help with the trilemma?
00:20:05
Speaker 3: I think in theory you could, but it's about imposing suffering essentially on some part of your society, right? And so ultimately there's a cost and you have to crowd someone out. So if you look at this guns and butter kind of idea, China has the same trilemma, but their choice is sound money and guns. And then they suppress the households. They're suppressing butter, right? So it's ultimately a choice. It's more, I think the challenge in the West is because of how polarized society is, it's very hard to pass a sweeping bipartisan policy.
00:20:34
Speaker 1: Thing, right?
00:20:35
Speaker 3: Like Ross Perot managed to kind of pull the parties back towards the center, but he also had a very big kind of macro tailwind because the Soviet Union ended, you had a big peace dividend, and you had the technology over the top. I think without that, given how polarized things are, it's usually easier to, you know, that's why divided governments has this seasonal impact on, you know, equity markets going up, right? It's because we're divided, so it's easier to keep the butter flowing.
00:21:00
Speaker 2: Yeah, Joe, this is one of my big realizations of 2026 is that you know, people always talk about, oh, China wants to boost domestic consumption. I don't think they do. Like, I actually don't think they do. Yeah.
00:21:10
Speaker 1: Yeah, I.
00:21:13
Speaker 2: Agree. Okay.
00:21:16
Speaker 1: I also just, my other take on this is that even if they want, even if like, quote, Beijing, unquote, wanted to, I don't know if they can. Because if you think about how we know, like, policy is done there, which is, like, everything is done through, like, regional KPIs, et cetera, they might say we want to increase domestic consumption. But if every party leader in every province still thinks that, like, building out factories is their path to political promotion, then, like, it's not obvious how they actually....
00:21:49
Speaker 2: Well, also now I'm thinking of like every local government basically competing on retail sales or something, which would be interesting to say the least.
00:21:57
Speaker 3: Okay.
00:21:58
Speaker 1: Can we talk stocks?
00:22:00
Speaker 3: Yeah, sure.
00:22:00
Speaker 1: What do you make of NVIDIA being down like 14% and the S & P 500 being down 0.17%? We're describing this company for a long time as the beating heart of the most important industry of the market, maybe world history. And its future has never been more uncertain. And the stock market is down like not even a quarter of a percent.
00:22:28
Speaker 3: Yeah, so I think you guys alluded to this a bit earlier, but the key here is AI financing, right? So in my mind, this is almost like you're running a marathon because just the sheer amount of financing we need to raise in the next few years. But because of all these pressures, every marginal fundraise is like a sprint. And in a way, if you look at, the circular financing around NVIDIA, I think it's because they're sprinting. You don't have time to bring people in. That's why you do all these, you know, inverted balance sheet type things.
00:22:54
Speaker 1: Explain that. What do you mean they don't have time to bring people in?
00:22:56
Speaker 3: So normally if you want to raise financing, you can, you know, go to everyone, get everyone to do their analysis. It's time to wait, right? But the way the AI financing is, we have to raise so much money so quickly to keep the game going that every sprint is kind of this rush. Okay, what can we do? Everyone's very short-sighted. We just got to get the next fundraise done so we meet the next tranche of obligations and whatever we committed. You know, the analog is almost a bit like, say like, you know, government deficit is the same thing. It's a marathon to have to keep financing the government, but every auction is a sprint.
00:23:28
Speaker 1: I get that, but explain what does that have to do with Nvidia being at the center of it all? I get all they're
all sprinting, but what does that have to do with Nvidia?
00:23:35
Speaker 3: Well, it's because the AI ecosystem is essentially about can the frontier labs and the other guys raise enough money to buy the compute to train smarter models. That's the confidence for the market. For the market to have confidence for these guys to raise financing, they clearly need these deals to have an anchor person, right? There's a reason the AI ecosystem guys are stepping in to be the anchor to deliver the confidence for the market so that ultimately they can raise money from outsiders. The problem is, I think this works, but now that the Fed's hiked, now that we have these inflation pressures, now the interest rates are tightened, this becomes much harder. That's why I think this analogy of you're going to have to sprint even harder now against the headwind of higher rates But it's still a marathon, right? It's not like you do one sprint.
00:24:20
Speaker 2: It's over.
00:24:21
Speaker 3: I don't know which specific sprint makes the runner collapse, but you can see it getting harder and harder to sprint.
00:24:28
Speaker 2: You run a systematic fund at Variant as well. So going back to Joe's point about, you know, NVIDIA is down like 14% now. Is that right?
00:24:37
Speaker 1: Yeah, let's see. Yeah, it was the last time I got a headline.
00:24:40
Speaker 2: Yeah. It's one of those days where we're going to have to start saying the exact minute that we're recording the episode. It's 9.55 a.m. It's only down 7% now.
00:24:48
Speaker 1: All right.
00:24:49
Speaker 2: Well, I mean, still pretty... Still, this time last year, if people had said that this would be happening at NVIDIA, I think everyone would have been very surprised. But can you talk to us about, like, I guess dispersion in the market and why we aren't seeing more of a ripple effect on some of the AI companies that NVIDIA is actually financing?
00:25:10
Speaker 3: Well... Maybe the Oppenheimer analog is appropriate. Ultimately, the H-bomb was built. Once Oppenheimer was out of the way, things were still done. There's still a need to win this race. And if the financing market struggles, I think the governments probably will step in and backstop. But the problem is, obviously, if the governments get very involved in financing, the winners and losers change, especially if you're in a slightly marginal part of the AI ecosystem. I suspect the government's not keen to finance that. But if you're the top two frontier labs... And you have evidence that when you put the compute in, your models are getting smarter. I think that will be backers. So I think that's why it's somewhat more of a, I would say, a rational response at this point. I think the structural theme is still intact. The question we have is around financing and it's reflexive, right? Let's say that fraud pick IP goes well next week and it's a big bump. Then in two weeks, open air announces, hey, we close our next private round. Suddenly soft banks, collaterals, money good again. And then all the pre-commitments are real, right? Because that's real money flowing. So I think it's just more acknowledging the two-way risk that I would have said there was less of, say, back in July when the situational awareness happened. Back then, the macro was very benign. Today, it's pretty unambiguous how much the macro cycle has shifted.
00:26:22
Speaker 1: Do you think, okay, so the macro cycle has shifted. In your view, has there been any shift in the AI fundamentals story? Because, you know, there's all these questions about, like, are the customers really getting ROI and so forth? Is it just a macro story in your view? Has there been no change really to the sort of like AI fundamentals part of the leg?
00:26:44
Speaker 3: I would say the macro and the AI are finally intertwined. So the reason the market's, I think, been very bifurcated is you've had a classic macro cycle in the non-AI stuff, right? Like, you know, the fast-tracking rates close down, but AI has been the secular thing. What connects them now is this financing need. And the financing need is a function of macro interest rates and where we are. And that's what's connecting them now.
00:27:08
Speaker 1: I guess I'm just saying the secular story about the AI build-out, like that this investment will be worth it. People are going to want to pay for it. That companies will like, oh, I'm getting value from my AI spending. That maybe consumers will like maybe pay for like an assistant bot. Like some of these theses that people have, like, Do you not see at this point any weaknesses in these stories? Oh, yeah.
00:27:36
Speaker 3: I think clearly there's some optimistic times for the markets. But I would say, again, if you put a sovereignty lens on, There's a lot of assumptions about which of the markets and where things are going to leak, right? So let's say we move to a world where governments have to get involved because we're strong into finance. There's going to be a lot of conditions attached to that government financing. The government might mandate everything you buy comes from the US locally. So then you inject a lot of stimulus into the system. It's all kept here. So asset prices here might be okay. But if you're a foreign supplier, you have other things, that goes away, right? I think that's more likely the way we're going, that sure, if it's a pure private sector, you know, kind of monetization is tough. Right. But as, as we said, if you're trying to, the nuclear bomb has no economic value, but we've put lots of resources into it. So you have a nuclear bomb today, plus this deal use technology. And clearly I will say there's a key assumption we have to make here, which is, is this still the case that if you put in a lot more compute, you get a lot smarter of a model.
00:28:34
Speaker 1: Yeah.
00:28:35
Speaker 3: I would say that's open question. The evidence so far is still yes.
00:28:38
Speaker 2: Yeah.
00:28:38
Speaker 3: If
that's true, then you have to back the big guys and put a lot of resources behind it because clearly whoever wins and has the best AI will hack the other country, right? And then the world will split. Every other country has to pick which side they go in. There's still a lot of markets to be addressed and it's backed by governments.
00:29:08
Speaker 2: This is a question I've wanted someone to respond to on the podcast for a while, but we haven't had a chance. Can you just explain what's going on with gold? Because I felt pretty good about stacking my gold coins earlier this year. And part of your impossible trilemma thesis is, well, if we're going to give if the sound money leg is going to give way, then there's higher inflation and buy gold. But what we've seen this year is like actually gold doesn't seem to be responding that much to higher inflation. when interest rate expectations go down, it's actually moving up. And it also doesn't really seem to be much of a safe haven. I'm looking at it right now and it's down. What is actually driving gold at the moment?
00:29:48
Speaker 3: Yeah, I will say this is where the echoes to 60 is very interesting, because I say the 23 to 25 rally was probably the London gold pool break. And you got that initial warning. But what happened afterwards, there was a slight adjustment. You know, Walsh is not off of burns, but Walsh could be this generation's Martin, right? Where at the margin you do adjust. He did hike rates back then. Inflation comes down. You have a bit of a credit crisis. Things normalize. But ultimately, at the margin, policy is still moving that way. So I think that's the adjustment we're on, on top of the, I would say, meme-like trading that took over gold at the beginning of the year.
00:30:20
Speaker 2: The other explanation I've seen is, well, oil prices have been going up, and so you have central banks or governments that have to liquidate some of their gold positions in order to pay for the higher energy bill, even though we see reports of them buying more gold.
00:30:34
Speaker 3: Yeah, I mean, again, this is where I think time horizons matter. I would say cyclically, interest rates have gone up, right? If interest rates go up, that's probably the biggest thing on, you know, it's a headwind to gold. But I think structurally that that long kind of long cycle analog is you have these counter trend moves quite a lot. But the fact that the gold has behaved is telling you there's underlying pressures. It may take a while before that truly explodes.
00:30:56
Speaker 2: So eventually I'll feel good about my gold coins again.
00:30:59
Speaker 3: Well, I think that's why we have the layout. So yeah, so If all these pressures persist, yes. But let's say suddenly US and China announce an AI safety summit, right? Analogous to kind of arms control. Then that leg comes down. It's not so big. Equally, let's say in domestic politics, you start getting a shift towards, hey, more people talk about the deficit, right? So I think that's why it's a trinity. You have to see which leg's giving way and how to position.
00:31:25
Speaker 2: Can I just press you on this point? So what do you think is the most likely circuit breaker at the moment? Is it something like, you know, a technological arms race accord or a candidate coming in and saying like, no, we're actually going to be serious about the deficit?
00:31:39
Speaker 3: I would say we probably keep, it's probably the arms control analog, that we keep escalating to a point where you create so much physical constraint, both here and in China, that there's acknowledgement, okay, we have to do something about it. And that would be obviously, again, analogous to the Cold War and those things, yeah.
00:31:56
Speaker 1: So in your view, when you say this is creating so much physical constraint, which is that like, I'm just trying to like, your view is basically that even if the AI investment can't be fully backed by the market alone, that it's still going, as long as like the scaling persists, as long as you get more value for more compute, US and China will just have no choice but to keep investing in it regardless. And then this kind of creates like a floor under the industry. Even if it's not economical, because there's this sort of like cybersecurity race and because of scaling, it's just going to have to keep pushing on absent some sort of agreement to like genuinely pace the frontier or whatever.
00:32:44
Speaker 3: Yeah. Again, I don't know for sure, but I think that's the angle markets typically don't tend to you know, give a lot of weight to. And I think this is part of that longer term trend. Like I say, if I told the story on China, all the way back to that, you know, sanctioning of Huawei all the way back, right. That countries need to think very long-term. So I don't think there's a lot of evidence that there's a rapprochement between us and China over time. So, and this is where like simple DCF analysis.
00:33:12
Speaker 1: Pure national defense and geopolitical security will keep a floor under AI spending, regardless of what the market supports.
00:33:19
Speaker 3: So, yeah. But obviously the tricky thing as an investor, though, is even if the government puts money in, the winners will redistribute, right? So we've had a phase where the rising tide lifts all boats. But if we move to a more sovereign-backed investment AI, then let's say if it's a domestic critical resource provider, if it's a domestic equipment maker, they're front of the line. They might become even the monopoly supplier, right? But if you bought the fifth... frontier AI lab that doesn't have enough compute. I don't think that's a good investment, right? It's more, you're going to differentiate a lot more winners than losers. And again, you can look at the Chinese markets, right? That's what's been happening.
00:33:56
Speaker 2: Yeah. So this is something I wrote about a few years ago, but like China is just much better
for obvious reasons at signaling which industries it actually prioritizes from a sovereign perspective. And you can see that, you know, Xi Jinping will say something about, we want to focus on choke point technologies. We don't want to focus on like, e-retail or something like that. And the stocks kind of adjust, the financing adjust. But we haven't yet really seen that in the US. And it feels like that's kind of an adjustment process that's still underway. Yeah.
00:34:29
Speaker 3: And I think that's the way things are going. You've got to be aligned with the sovereign when you invest.
00:34:35
Speaker 2: I still remember Chairman Rabbit put out a tweet during the big tech crackdown. And he basically said like, put your money in what the government is saying to put your money in, which at that time was like hard tech and chips and stuff like that. Like, it's very clear there.
00:34:53
Speaker 1: What's going on with the Chinese market now? The Chinese market hasn't been doing well at all.
00:34:57
Speaker 3: Yeah, they've been suppressing butter. So if you own consumer tech, it's not exactly aligned to the sovereign. Right.
00:35:03
Speaker 1: It's tough.
00:35:04
Speaker 3: That's why. I think if you take a step back, I think it's just, I'll speak for myself. My entire career, we've all learned from Warren Buffett, all these kind of analysis, there's a core assumption that your regulatory geopolitical environment is stable. Therefore, we laser focus on what's the ROI, right? It's just more we're now moving to a world where it's about... Just these constraints that for a century we haven't had to deal with. Will you get your money back? We'll just have these hard cutoffs. And those are obviously just 1-0 events, right? And I think the problem is the world we're going to, you can't predict when, but every marginal shock will probably hurt you if you're exposed to that. Say if you're a US investor, you put money in China. If nothing happens, great, the valuations are cheap, fantastic value investment. But if there's going to be a shock, is it more likely that your money's suddenly locked up, you can't repatriate, or there's an investigation, right? I think it's just that's the sovereign angle where you can build your portfolio as normal, but the resilience is going to come from understanding the direction of these shocks is always going to be likely inflationary and likely towards some kind of capital control type thing that distorts valuations.
00:36:09
Speaker 2: Is this like a slow creep in the tensions in the trilemma? Or would you expect like a catalyst to suddenly like force a decision on these three factors?
00:36:19
Speaker 3: Yeah, so this is why I wrote the report because the Iran war started giving me the vibes that it's acting as a kind of a catalyst. Because again, the whole point of wars is once it starts, as we found out, it's hard to stop. And it's like a very extreme version of what the sovereign needs, it definitely needs to get,
right? And so if we drag this out, then it's going to have these crowding out impacts. So you have kind of this sovereign crowding out. AI itself is a very crowding out thing. And it will happen as a pace, which is going to force either the populist response. So let's see, do we go for Ross Perot? Do we end up with the extractionist kind of Juan Peron type thing, right? I think this is a forcing mechanism. We're going to see it more quickly over time.
00:36:58
Speaker 1: Or we get the big data center revolt and the domestic population says, we don't want the AI anymore. That has to be at least considered as a possibility.
00:37:07
Speaker 2: Or a productivity boom, which everyone seems to... That's the wild card that everyone is sort of resting their hopes on, right?
00:37:14
Speaker 1: Yeah.
00:37:14
Speaker 3: So this is very interesting. I looked at all their historical analogs and It's obviously hard to measure productivity, but there's things like, for example, if you look at the Spanish and Portuguese empires in the Age of Exploration, you get a lot of gold and silver just get flooded into your system, right? So they were, again, very indebted. Suddenly you get this relief. But the problem is, all it does is exacerbate your existing institution. So if you go back to, this is a work of, you know, Daron Acemoglu, originally, right? Whatever institutional constraints and biases you have, your productivity boom merely exaggerates it and delays it, right? So unfortunately, that wasn't really a good one. Believe me, I try to be optimistic and look.
00:37:52
Speaker 1: First of all, also, another headline.
00:37:55
Speaker 3: Oh, God.
00:37:56
Speaker 1: This is one of those days where, like... The network of companies tied to Alice Wong recently invested in Chinese AI company DeepSeek. So her company is really DeepSeek. This is all getting very interesting. Say more about the downside of a productivity boom. This is a very counterintuitive take.
00:38:15
Speaker 3: As I say, no, obviously the first order impact is it releases your debt constraints. It kicks the can down the road. It's the classic can kicking down the road. But when you have it, it doesn't force you to reform. The whole point right now is what is going to force us to enact reforms to essentially address this guns, butter imbalance, right? And I think that that was more of my point.
00:38:35
Speaker 1: But we can have guns and butter forever. If we just keep getting more and more productivity, right? Like that would be okay. Yeah.
00:38:40
Speaker 3: But I guess I high enough rates, right? Obviously the only day upon really have, and I think Nick bloom of Stanford's not a lot of work on this electrification, right? Ultimately it's worth max one and a half to 2% GDP a year, right? Like you would need obviously a lot bigger. I think that's a challenge.
00:38:55
Speaker 2: What about a productivity decline? Because the other thing that could happen is if the sovereign is driving all this investment and choosing winners and losers, you could easily get overinvestment and oversupply and things not working in sort of economic or market efficient ways.
00:39:11
Speaker 1: Yeah, absolutely.
00:39:12
Speaker 3: I think that's ultimately the end game because in the original phase of mobilization, we're essentially doing duplicates of everything, right? So initial phase is inflationary. Everything's bottlenecked. And then you overbuild, your domestic problems kick in, suddenly you have arms control, and then we overbuild everything. That's going to suddenly become very deflationary at the end, but it's just more in that ramping up phase still.
00:39:34
Speaker 1: Right. Arguably, that is why we still have this just persistent... I mean, arguably, that's another way of describing inflation. Productivity is bad, another way of framing almost the same thing and just saying, yes, this is why we still have this inflation problem because there's just like all this spending.
00:39:51
Speaker 2: Everyone's trying to do everything at the same time.
00:39:53
Speaker 3: Yeah, I think this is why it's very hard to trade long-term bond yields, right? Because I would say, actually, if you take a step back, AI is a secular deflation thing, right? We would hope. Right. Well, I think the evidence probably suggests AI in the long term, I think, is deflationary. But the problem is every geopolitical shock, domestic shock is going to be inflationary against that. And I think that's why the market is going to struggle.
00:40:17
Speaker 1: Or is it the problem that like, yes, maybe in theory, AI is deflationary. But in the short term, you always have to invest more. And so it's a series of inflationary impulses.
00:40:31
Speaker 2: Because as you say, it never ends, right? You're always trying to develop the next model.
00:40:34
Speaker 1: This is the thing, right? There is no finish line where it's like, OK, we did it.
00:40:39
Speaker 3: Right.
00:40:40
Speaker 1: And so in the meantime, it's a deflationary marathon, but an infinite. It's not actually a marathon in some sense, because a marathon is 26.2 miles. We need a new word.
00:40:50
Speaker 2: It's a series of inflationary sprints.
00:40:52
Speaker 1: It's a Sisyphean marathon.
00:40:53
Speaker 2: That never ends.
00:40:54
Speaker 1: It's actually a Sisyphean marathon, isn't it? Because there actually is no finish line, and so it's just sprinting forever.
00:41:01
Speaker 3: Yeah, well, I think that was the key assumption we have to make. I said it earlier, which is, does a lot more compute get you a lot smarter of a model? If at the margin we see evidence of that slowing down, or in fact stop, that would be a pretty good sign that AI safety is coming, everything is done. Then at that point, I would say it's over.
00:41:18
Speaker 1: Yeah, I guess it's not guaranteed that there will just always be more smarter forever.
00:41:23
Speaker 3: Yeah, I think what's been unique about this AI cycle is if
you're bearish, you could have two years ago thought, hey, if we just put more compute in, it's not going to get that much smarter. But the problem is we've had a lot of evidence. You put a lot more compute, it gets a lot more smarter. When does that end? Obviously, I don't know, but that's why we'll be watching that for the trend to turn.
00:41:42
Speaker 1: All right. Tianyang, thank you so much for coming on the podcast. Kind of crazy times and a fun discussion.
00:41:50
Speaker 3: Yeah, thanks for having.
00:42:01
Speaker 2: Me.
00:42:03
Speaker 1: Tracy, I find it very hard to disagree with this trilemma thesis. I mean, at core, we know a few basic things right now about the world, I think. I think we know a few basic things and they are inflation is hot and people really do not are very upset about the cost of living. There are two major wars going on. One, we're directly involved in one kind of proxy. Three, there is this obvious rising geopolitical tension between the U.S. and China. And it relates increasingly not to the to AI specifically, which is very spending heavy. Yeah. And I don't know, it feels like something kind of got to give here. And I don't know what it'll be.
00:42:47
Speaker 2: No, I agree. I find it hard to ask questions on a thesis that I broadly agree with. And I remember back in 2021, I wrote this piece, The Chokepoint Economy. And this was like my big realization, which was like governments were waking up to their relative vulnerabilities when it came to crucial technology or supply chain components. And so we were going to be in a very like China-esque Style race to build out those resources. I do think like one of the problems here is everything is being couched in these existential terms.
00:43:19
Speaker 1: Right.
00:43:20
Speaker 2: And so if the sovereign says we have to develop AI technology because it's crucial to the future existence of our government slash population. then it's very hard to see anything that actually slams the brakes on that spending. And at the same time, if the technology itself is this iterative thing that never ends with a completely finished product, as we were discussing, you never say like, okay, we did it. Here's the AGI, we're done. Then how do you stop?
00:43:48
Speaker 1: Yeah. And then the other fact about the world that I didn't mention is We're getting older. And we, right? Like us personally? We are. But aging populations more broadly and the obligation.
00:44:02
Speaker 2: I thought you were going to say you're aging, so you're really into gold now.
00:44:07
Speaker 1: That's where we're going. No, I'm aging and I don't really care about the future anymore. No, but like, you know, we have made promises to retirees. A real resource promise. So that puts strains on the working share of the population. That further makes it harder to like get productivity gains and so forth. And so, yes, you have like these deep stresses. And I thought the point, you know, your question about like, you know, worsening productivity.
00:44:35
Speaker 2: Yeah.
00:44:35
Speaker 1: Clearly, like if everyone has to rebuild replica, if everyone has to rebuild their own domestic supply chain. That is a productivity killer.
00:44:43
Speaker 2: Which again is like a classic China problem of overinvestment, right?
00:44:46
Speaker 3: Yeah.
00:44:46
Speaker 2: Anyway, shall we leave it there?
00:44:48
Speaker 1: Let's leave it there.
00:44:49
Speaker 2: This has been another episode of the Odd Thoughts Podcast. I'm Traci Allaway. You can follow me at Traci Allaway.
00:44:54
Speaker 1: And I'm Joe Weisenthal. You can follow me at The Stalwart. Follow our producers, Carmen Rodriguez at Carmen Armand, Dashiell Bennett at Dashbot, Kale Brooks at Kale Brooks, and Kevin Lozano at Kevin Lloyd Lozano. And if you want to keep chatting about all these topics, if you want to discuss more about the Sovereign Trilemma. Check out our Discord, discord.gg slash oddlots. And for more oddlots content, Tracy and I write a newsletter every day, which you can go find at bloomberg.com slash oddlots.
00:45:22
Speaker 2: And if you enjoyed this conversation, if you like it when we talk about the existential race for everything, then please leave us a positive review on your favorite podcast platform. And if you're watching on video, then please like or leave a comment or better yet, subscribe.
00:45:36
Speaker 1: And if you're a Bloomberg subscriber, you can listen to all of our episodes absolutely ad-free. Just go to the Bloomberg channel on Apple Podcasts and follow the instructions there.
00:45:46
Speaker 2: Thanks for joining.
00:45:55
Speaker 1: Do do