Latest / The PhilStockWorld Investing Podcast / AI Bubble's Systemic Risk: How Financial Engineering, Sovereign Debt, and Demographic Collapse Threaten the Nasdaq

Transcript

AI Bubble's Systemic Risk: How Financial Engineering, Sovereign Debt, and Demographic Collapse Threaten the Nasdaq · The PhilStockWorld Investing Podcast

• 0:00
  Penny: Welcome to the deep dive. If you're like us, you've probably been watching the market headlines and feeling, well, a bit dizzy. Right?
• 0:06
  Roy: Definitely. Financial vertigo is a good way to put it.
• 0:09
  Penny: We've got all this global instability, a major US government shutdown, currency turmoil heating up, and yet the Nasdaq, the S and P five hundred, they just keep charging ahead, fresh record highs. It feels like the market isn't just ignoring the risks. It's like actively choosing momentum over everything else.
• 0:27
  Roy: It really is the definition of market cognitive dissonance. Yeah. And that's exactly why we wanted to put this deep dive together today.
• 0:34
  Penny: Right.
• 0:35
  Roy: Our mission really is to look past that, you know, surface euphoria and ask what systemic risks are actually propping up this rally. Okay. We've pulled together analyst forecasts, got some fresh numbers from Citi macro reports from New York Life Investments, and crucially, FTC analysis looking into these big AI partnerships.
• 0:55
  Penny: So we're connecting the dots then between the high flying tech stocks, the political chaos, and maybe these demographic time bombs we keep hearing about.
• 1:03
  Roy: Exactly. We need to see how they link up.
• 1:05
  Penny: Okay. Let's dive in then. Starting right at the heart of it all, AI. The amount of money being thrown around is just staggering.
• 1:13
  Roy: It is. Citi analysts are now projecting that the major hyperscalers and, you know, for listeners, that's your Microsoft's, Alphabet's, Amazon's, Oracle's, Core Weaves, they're set to spend something like $490,000,000,000 on infrastructure next year.
• 1:27
  Penny: Wow, dollars $490,000,000,000. That's nearly half a trillion dollars.
• 1:32
  Roy: Yeah. And it's a big jump from their last estimate, was around $420,000,000,000 Yeah. So this huge spending commitment that is the structural growth story everyone's focused on.
• 1:41
  Penny: I sense a but coming.
• 1:43
  Roy: Well, yeah. The key insight from the sources we looked at is that a massive chunk of this growth isn't actually coming from end user customers paying for stuff.
• 1:52
  Penny: So where is it coming from?
• 1:53
  Roy: A lot of it seems to be based on, well, financial engineering. Engineering. Okay. Give us an example.
• 1:58
  Penny: The clearest and maybe most, alarming example the sources point to is this deal between AMD Advanced Micro Devices and OpenAI. One source even called it the great tech circle jerk. That's pretty strong language. Okay, unpack that. What does that mean exactly?
• 2:13
  Roy: It's strong because the deal itself is so circular apparently. AMD's stock went absolutely nuts after they announced this huge deal with OpenAI.
• 2:21
  Penny: Right, I remember that.
• 2:23
  Roy: But the sources break down the terms. OpenAI got warrants for a 160,000,000 AMD shares. That's paper value over $35,000,000,000.
• 2:33
  Penny: Okay.
• 2:34
  Roy: And what do they give in return? A promise. A massive multi year commitment to buy chips from AMD.
• 2:41
  Penny: Hold on. Let me get this straight. OpenAI essentially got paid tens of billions in the stock value anyway just to promise to buy chips from AMD. Chips they presumably needed for their business regardless.
• 2:51
  Roy: That's precisely how the sources describe it. Yeah. These tech giants are basically transacting with each other, booking these internal promises as revenue growth and puffing up their paper valuations.
• 3:00
  Penny: Okay. So let's do the reality check then. What about OpenAI's actual business? Money coming in from real customers?
• 3:06
  Roy: Well, the estimates are that OpenAI's actual customer revenue people paying for ChatGPT Plus or enterprise services is maybe only $3 to $4,000,000,000 a year.
• 3:16
  Penny: $3 to $4,000,000,000 in revenue. But they're committed to infrastructure spending, funded partly by these deals, estimated at 15 to $20,000,000,000 annually.
• 3:27
  Roy: Yes. That's roughly a five to one ratio. Spending $5 for every $1 of actual customer revenue.
• 3:33
  Penny: That just screams bubble, doesn't it? Massive investment chasing tiny sales. So why are the sources comparing this to 02/2008, the financial crisis, and not the dot com bust back in February? That seems like the more obvious parallel.
• 3:47
  Roy: That's the really critical distinction they're making. Yeah. You know, back in February, you had lots of companies with zero revenue, crazy stock prices. When they crashed, it hurt, but it was somewhat contained.
• 3:56
  Penny: Right.
• 3:57
  Roy: 2008 was different. It was about systemic interconnectedness. Yeah. It was built on these complex, really opaque financial products, CDOs and the like. Yeah.
• 4:04
  Bullshit accounting as one source put it. Yeah. Today the argument is the AI bubble isn't just about high valuations, it's about this deep interdependence. If one major player say OpenAI or a key chip supplier hits a funding wall or fails.
• 4:19
  Penny: It doesn't just fail on its own.
• 4:20
  Roy: No, the collapse instantly cascades. It hits the chip suppliers like AMD, it hits the cloud providers hosting them, it hits the financial backers funding them all at once.
• 4:31
  Penny: Ah, so the risk isn't just that a stock price falls, it's that the failure of one card could potentially bring down the whole house of cards, maybe even threaten the stability of the Nasdaq itself because everyone's locked into these exclusive financial and supply deals.
• 4:48
  Roy: Exactly. That's the systemic risk they're highlighting. And we're seeing hints of the physical limits driving this. These big cloud providers, Microsoft, Amazon, they've publicly said they have existing AI capacity constraints.
• 5:00
  Penny: They can't get enough chips, basically.
• 5:02
  Roy: Pretty much. They're constrained on AI capacity, especially chips. And that scarcity forces them into these really complex deals, co developing chips, exclusive partnerships like Anthropic working closely with AWS.
• 5:14
  Penny: Which just tightens that web of interdependence even further.
• 5:17
  Roy: Per se.
• 5:17
  Penny: Okay. That paints a pretty concerning picture of the financial engineering inside the tech boom. But let's broaden the view now because the market seems to be ignoring more than just that. There's a whole deteriorating global political situation too.
• 5:31
  Roy: Right. We need to pivot to what the sources are calling the sovereign stress test.
• 5:36
  Penny: The sovereign stress test.
• 5:37
  Roy: Okay. Market might be glued to Silicon Valley but the actual sovereign systems, the governments and economies that underpin global stability, they seem to be actively unraveling in places.
• 5:48
  Penny: Like where?
• 5:49
  Roy: Well, look at Europe right now. France just saw its Prime Minister, Sebastien Lecourneau, resign after just twenty six days in office.
• 5:57
  Penny: Twenty six days? That's incredibly short. The shortest in modern French Right?
• 6:01
  Roy: Exactly. It signals just complete political paralysis.
• 6:04
  Penny: And this is happening at a terrible time for France, isn't it? They've got huge debt, like a 114% debt to GDP, and they need to find something like 40,000,000,000 in budget cuts.
• 6:13
  Roy: Correct. But how do you push through painful reforms when the government is in chaos? It becomes almost impossible.
• 6:21
  Penny: So France is the immediate acute chaos. But where do we see the longer term structural issues playing out?
• 6:29
  Roy: Japan. Japan gives us a kind of preview according to the sources. It's the world's third largest economy, remember? Right. And the details there are stark.
• 6:37
  The new prime minister, Sana Takedaichi, is apparently doubling down on massive government spending.
• 6:42
  Penny: Despite the fact that Japan's debt to GDP ratio is what is it? Something absolutely astronomical.
• 6:50
  Roy: It's mind numbing. 263%. The highest, basically, in human history for a major economy.
• 6:56
  Penny: 263%. How is that affecting their economy now?
• 7:00
  Roy: Well, it's driving currency debasement. The yen's down, what, 30% since COVID? Wow. And inflation, which Japan famously hasn't had for a decade, actually hit 3%. That's historically huge for them.
• 7:10
  Japan is kinda showing us what happens when impossible debt math collides with a lack of political will to actually tackle it. They seem to be trying to inflate their way out.
• 7:18
  Penny: Which rarely ends well.
• 7:20
  Roy: Rarely. Yeah. And underlying all of this, in France and Japan, in The US, everywhere, is the demographic time bomb.
• 7:26
  Penny: Ah yes, the simple terrifying math problem. Who's gonna pay for all the promised benefits, whether it's high tax France or relatively lower tax US?
• 7:37
  Roy: The math is simple and terrifying. It's the worker to retiree ratio just collapsing. Think about The U. Back in 1960, you had 5.1 workers supporting every retiree.
• 7:49
  Penny: Okay.
• 7:49
  Roy: Now it's down to 2.8 and the projection for 2030 is just 2.3 workers per retiree.
• 7:56
  Penny: That trend line is stark.
• 7:57
  Roy: It is. And Japan, their birth rate is down to 1.2 children per woman. You need 2.1 just to keep the population stable. When this fundamental math breaks down, the sources argue, the inevitable result is economic stagnation, currency debasement, and political paralysis. Sound familiar?
• 8:12
  Penny: It sounds like exactly what we're seeing play out. Now, speaking paralysis, the sources draw an interesting contrast in how different systems are handling or maybe not handling this debt pressure. They compare France with its high taxes, high benefits.
• 8:23
  Roy: Right. France has a 44% tax to GDP rate that buys you things like free college, universal childcare, loads of mandatory vacation, lower health care costs per person.
• 8:32
  Penny: Compared to The US model.
• 8:33
  Roy: Exactly. The point the sources make is that the political difficulty of making changes is immense regardless of the system. In France, that high social contract means any austerity is political dynamite, hence the twenty six day prime minister. In The US the argument goes even with lower taxes the private costs are so high especially healthcare $12,500 per person versus France's $5,000 that it creates immense pressure on individuals. It's described almost as economic servitude.
• 9:03
  Penny: So both paths seem to lead to the same outcome. Yeah. Political paralysis. The will to either cut benefits or raise taxes enough to actually service the mounting debt, what, dollars 38,000,000,000,000 now in The US just isn't there.
• 9:16
  Roy: That seems to be the core argument. And The US is adding another layer of risk right now.
• 9:19
  Penny: The shutdown.
• 9:20
  Roy: The shutdown, yeah. Day six now. Which means critical economic data isn't being released. No jobs report, no CPI updates, no GDP revisions.
• 9:27
  Penny: So with $38,000,000,000,000 in debt, global instability, potential recession, we're base basically flying blind Right. Without the official economic instruments.
• 9:35
  Roy: That's the analogy one source uses like unplugging the smoke detector while the house is on fire.
• 9:39
  Penny: Okay. So if you're an investor and you're watching governments seem paralyzed, debt spiraling, fiat currency looking shaky, where do you go? Where's the money flowing?
• 9:50
  Roy: We're definitely seeing a powerful move into what's being called the debasement trade.
• 9:56
  Penny: The debasement trade. Meaning fleeing potential currency devaluation.
• 10:00
  Roy: Exactly. Fleeing policy uncertainty, currency risk. Gold, for instance, is really pushing higher, challenging $4,000.
• 10:07
  Penny: Okay. Gold makes sense as a traditional safe haven.
• 10:09
  Roy: Right. But what's perhaps more striking or at least newer is Bitcoin. It hit $125,000 intraday highs recently. The argument there is it's driven by people seeking a non sovereign asset, something outside the direct control or potential debasement of government policy.
• 10:25
  Penny: Interesting. So the old correlations, the old asset relationships might be breaking down.
• 10:29
  Roy: It certainly looks that way, which really underscores the need for holistic diversification, thinking differently about where safety lies.
• 10:36
  Penny: So let's get concrete then. What are the investment for trying to navigate this double whammy of the interconnected AI bubble and the sovereign debt crisis?
• 10:46
  Roy: Well, resilience in income generation seem to be the key themes coming through. On the equity side, the sources tend to favor large caps right now.
• 10:54
  Penny: Why large caps?
• 10:55
  Roy: Mostly history. As The US economy slows large cap companies have historically proven more resilient. Small caps, on the other hand, face much higher refinancing risk.
• 11:07
  Penny: Because interest rates are higher. That ten year Treasury yield still hovering around 4.16%?
• 11:13
  Roy: Exactly. That persistent upward pressure on yields makes it much more expensive for smaller, often more indebted companies to roll over their debt when it comes due.
• 11:23
  Penny: Okay, that makes sense for equities. What about fixed income? Bonds? That space gets tricky when rates are volatile.
• 11:30
  Roy: Yeah, very tricky. The general recommendation seems to be stay short duration across different types of credit in The US.
• 11:36
  Penny: Short duration meaning less sensitive to interest rate swings?
• 11:39
  Roy: Precisely. When policy is this uncertain and rates could jump unexpectedly, you want less sensitivity, less risk from those rate moves. Short duration helps with that.
• 11:48
  Penny: And where are the opportunities for generating income if traditional bonds are maybe less attractive or riskier?
• 11:55
  Roy: The focus seems to be shifting towards alternatives. Things like private credit are mentioned frequently as places to potentially find yield.
• 12:03
  Penny: Okay. And finally, tying it back to the start, that huge $490,000,000,000 AI infrastructure spend, where should investors be looking in terms of the actual physical build out?
• 12:13
  Roy: Right. That capital has to flow somewhere concrete. The sources emphasize infrastructure, specifically digital infrastructure.
• 12:21
  Penny: Data centers.
• 12:21
  Roy: Data centers, yes. The power grid needed to run them which is a huge challenge in itself and just the broader digital transition infrastructure. These are seen as key beneficiaries because, well, they are the essential building blocks. The AI revolution, managing sovereign stability, the energy transition, they all need massive investment in this kind of physical infrastructure.
• 12:42
  Penny: This has been, wow, an absolutely crucial deep dive. It feels like the main takeaway is that this record breaking market momentum, it's potentially masking two incredibly deep systemic vulnerabilities.
• 12:54
  Roy: That seems fair.
• 12:55
  Penny: First, this financial engineering bubble in tech built not just on hype, but on really complex, fragile interconnections. And second, this sovereign debt and demographic crisis that's clearly unfolding across the developed world with places like Japan and France may be showing us the future.
• 13:13
  Roy: Yeah. To quickly recap, AI is fundamentally reshaping economics and risk for better or worse. That debt and demographic time bomb, it's not theoretical anymore. It's an active crisis playing out now. Right.
• 13:24
  Which means diversification, focusing on income generation, and investing in that durable, essential infrastructure. It's not just smart. It's probably paramount for protecting capital in this kind of environment.
• 13:36
  Penny: Okay. And before we wrap up, here's maybe a final provocative thought for you, our listeners, to chew on. That massive 490,000,000,000 we kept mentioning, it's mostly going into developing and training these huge foundation models, things like GPT four
• 13:48
  Roy: AI systems. Yeah.
• 13:49
  Penny: Now the sources note these models have characteristics like homogenization, meaning flaws in the core model get passed down to everything built on top of it, and they also suffer from unexpected failure modes. Things go wrong in ways we don't predict.
• 14:04
  Roy: That's a known issue, yes. Complexity breeds unpredictability.
• 14:08
  Penny: So given the extreme political instability we've just discussed, the US shutdown, France's revolving door premiership, how resilient will these incredibly complex, increasingly essential, but maybe poorly understood AI systems actually be when the financial and political chaos eventually triggers a real systemic shock?
• 14:27
  Roy: That's a deeply concerning question. The sources argue very strongly that we must infuse social considerations and ethical design deeply into the technological development of foundation models from the start. Because the systemic risk we see building in finance, it now seems mirrored in the structure of the technology itself.
• 14:44
  Penny: A sobering thought indeed. Thanks for joining us for the deep dive.