Market All-Time High! Risks Even Higher! Who is CRAZY Here?

Duration: 7:36 1063 transcript words EN summary

Video summary

The stock market has recently reached all-time highs, yet this optimism comes with mounting risks as credit markets tighten and the spreads among hyperscalers widen. While banks remain stable, there are growing concerns that mirror the conditions seen in 2007 regarding these massive technology companies and broader AI-related investments. Research indicates that without the substantial spending from both US government deficits and private sector giants like hyperscalers, the economy could face a recession or even depression. However, despite this heavy reliance on debt financing to fund long-term projects, there is significant uncertainty about whether such expenditures will yield profitable returns in the future, creating a dangerous mismatch between asset duration and liability duration that financial experts warn against. A major source of confusion lies in the global distribution of AI growth versus market perception; while US tech giants like Apple, Microsoft, and Meta are pouring cash into artificial intelligence with little concern for dips, Chinese models such as Deepseek and ZAI account for over 50% of token usage despite receiving far less capital. This disparity suggests that Wall Street may be ignoring the reality that much of the current AI boom is driven by cheaper alternatives from China rather than just US innovation. Furthermore, JP Morgan estimates that a vast majority of the projected $5.5 trillion in AI capital expenditure will be debt-financed, raising questions about sustainability when token volumes surge but monetization fails to keep pace with adoption rates. The situation resembles historical infrastructure booms like railroads or electricity, which revolutionized the world yet delivered terrible long-term returns for investors who chased them without a margin of safety. The core issue driving this market frenzy appears to be greed and a collective willingness to gamble on an unproven future rather than sound data justification. Investors are essentially betting that AI will work perfectly over the next five years, but if it fails, they risk losing massive amounts of money due to their heavy reliance on debt and equity issuance. In contrast, true value investing focuses on maintaining stability regardless of market conditions by seeking a margin of safety and utilizing hedging strategies rather than blindly riding out an irrational bubble. As Warren Buffett famously noted, only when the tide goes out will it become clear who is actually swimming naked, highlighting that many current participants are exposed to significant downside risk while ignoring potential long-term profitability issues. Ultimately, the market seems determined to ride this wave for another year or two despite the lack of fundamental data supporting such valuations beyond pure speculation and greed. The disconnect between the massive spending on digital infrastructure, semiconductors, and smart grids versus their actual economic returns is becoming increasingly apparent as free cash flows from these sectors stagnate or turn negative. While some argue that this trend might last longer than expected, a prudent approach would involve recognizing that everyone is currently gambling on an outcome that remains unknown rather than investing based on certainty. The coming years will likely reveal whether the current trajectory of debt-fueled AI expansion can sustain itself or if it represents one of the most toxic financial errors businesses have ever made by borrowing short to invest in long-term projects with uncertain payoffs.

Read the full video transcript

Good day, fellow investors. The stock market is at all-time highs, but the risks are piling. Who is crazy here? Found this great article on Bloomberg. Credit markets tightening a little bit, the spread of hyperscalers going up. Nothing tragic, but it is something. Some research, this from Colombia shows that without the hyperscalers spending so much money, the US economy would be in a recession. I would add here that without the US government spending so much money 25% deficits on revenues piling the debt like there is no tomorrow debt that to GDP higher there would be likely a depression but okay that is some specifics longerterm specifics however those go hand inhand with the current situation the credit risk is still okay but the duration time spread that we see on this chart here is expanding for hyperscalers. Banks are stable but you can see here what was going on in 2007 and that is now going on for hyperscaler and and broader AI related investment grade. So that is something to keep in mind. And then on top of everything, some people said the hyperscaler will spend less, but with what's going on and how they see the world, the spending will never end. And here comes perhaps one of the most overlooked but most genius ways of thinking from Warren Buffett is the mismatch between the assets duration and the duration of the liability. Warren Buffett discusses how companies borrow short to invest long and that's one of the most toxic and fatal errors financial institution or business can make. This we have to change a little bit for the hyperscalers. They are investing long. They are also borrowing long. But we don't know whether the spending here will give return on investment. For now, what's certainty is that the total free cash flows have gone to less than a quarter of what those were just two years ago. Apple is not spending on AI high cash flows. Microsoft is, Oracle is, Alphabet is, Meta, everyone is spending their cash flows. The market doesn't seem to care every dip. If you buy the dip, you simply make so much money. So the buy the dip continues. Investments related to AI are exploding. South Korea, NASDAQ, smart grid, digital infrastructure, semiconductors, SML, everything is exploding. But am I stupid or what? This is the world open routter token usage. Yes, great growth. Everything is growing. Beautiful. But then I check this. Who is growing? Deepseek, ZAI, Quen, Minimax, Moonshot AI. That's more than 50% of it. That's China. And then we have the other ones. I don't know. You tell me. Who's the stupid one here? Or simply Wall Street doesn't care, never cared about China. If you say China, you lose retention, viewership, everything. That's how it goes. But I don't know. And look at this. We have seen 50/50 token and everything. This is the spending from China with IPO and bonds at 200 billion. Okay. US bonds especially exploding now more and more negative free cash flows and everything more than a trillion and we have seen that stabilizing at more than a trillion per year. This is insane. But as token volumes surge, falling prices and migration toward cheaper models suggest monetization has not kept pace with adoption. This is key and this is the mismatch in duration. We do not know whether the invested money here will be profitable in five years. The debt, the issued equity, all that the spent money will be spent. So that's a certainty. However, the return on investment is still unknown, but everyone is gambling on it. And I'm truly thinking, I'm going crazy when I see this that 50% is Chinese for much less money. This is insane. JP Morgan estimates that 4.1 trillion of the 5.5 trillion in AI capex will be debt financed. But if you spend four 5.5 trillion, you need to have a great return on it. So, I don't know. The Bloomberg journalist discusses here how the risk might come from rising long-term treasury yields that will impact the debt, but that's not really what's going on. The risk comes from not knowing what's the profitability of all those investments down the road. We discussed circular financing. This is getting crazy. And then if you look at things a little bit, telecom not great returns. Electrifications peaked in 1911. Worst performing sector for the next 100 years. Did it change the world? Yes. Did the internet change the world? Yes. Terrible returns. Railroad the same. Who is crazy here? I'm thinking, okay, what's the disconnect there? And then I listened to a podcast Steve Eisman legendary the big short and he was discussing Michael Bur calling the top he says that if it's month from now that's one thing if it's a year from now then you still want to play everyone still wants to play. So is it possible that the best explanation for what's going on in the market is this might last another year or two let's go for the ride. To me, that's insane. But it seems that's how the market thinks. There is nothing, no data to justify all of this except greed and riding the situation. To end the circle with Warren Buffett, as he would say, only when the tide goes out, we will be knowing who is swimming now. Naked AI. Greg Ael from Bergkshire has started spending his money. We'll discuss that tomorrow. I'm fine waiting for another five, seven years. But I'm different there because value investing always and ever is about if this happens, I win. If this happens, I win. If this happens, I win. That is value investing because you look for margin of safety. You look for doing okay no matter what happens. Everyone is gambling on AI. If it doesn't work, they lose a lot of money. If AI works, if AI doesn't work, we keep on compounding. That's value investing. You can have hedges. Michael Bur, we discussed it in a video two days ago. Fully hedged, shorts, this and that. There are other options to be hedged, cost-effective hedging, things like that. You can check more on my research platform. And that's it for now. I'm looking forward to your comments on who is crazy