This post is opinion only. See full disclaimer below
Zero Hedge—much like this site—frequently likes to give its readers a heads-up on key breaking news and higher-level structural concepts. Sometimes it is superb in this regard, and other times it features material I disagree with quite a bit. However, I definitely recommend considering a very interesting recent essay they posted: “Forget CDOs, Meet CCOs: This Isn’t A Tech Cycle… It’s 2008 With Silicon.”
The argument in that piece could be entirely wrong or right in important ways. Its discussion of second-level derivatives and complex financial instruments is outside my primary wheelhouse, which is macro-political. Nonetheless, it is conceptually fascinating, well worth a read, and invites an extension regarding its base argument.
The 2008 Analogy & The Deceleration Trap
The essay’s core thesis—and I’m painting in very broad brush strokes here that do not fully enough consider its deeper important details—is that the correct historical analogy for the AI market bubble is not the 2000 internet boom, but the 2008 housing crisis. It puts forward the highly advanced argument that the 2008 housing bubble crashed not when growth went negative per se, but simply when the rate of growth decelerated to a point where debt could no longer be rolled over, causing the whole house of cards to crumble.
What I find fascinating is what happens to this framework when you introduce a key variable we have discussed often here: It is impossible to scale energy at the same rate as computing power because energy does not follow Moore’s Law. Wall Street prices the future on a software spreadsheet while ignoring the laws of thermodynamics.
Thermodynamics vs. Silicon
The market narrative treats AI as a perpetual motion machine—scaling exponentially like silicon under Moore’s Law, assuming infinite digital growth in a vacuum. But artificial intelligence doesn’t run on pure code; it runs on massive, uninterrupted, 24/7 baseload energy. Here is where the fantasy hits a physical wall: oil and hydrocarbons do not scale exponentially.
- Computing (Moore’s Law): Computing power may follow Moore’s Law and double in capacity for the same cost every 2 years, but energy follows no such law. In the last 50 or so years transistor counts and computing performance has gone up by a factor of over a billion through exponential compounding..
- Hydrocarbon Extraction: Oil extraction and recovery efficiencies have seen linear, incremental improvements, moving total reservoir recovery from roughly 10% historically to an average of 30% to 50%—a fractional, multi-fold efficiency gain rather than an exponential curve. Nor will new energy sources save the day providing Moore like gains. Guess how much over the same period a solar cell has increased its energy output? Only by a factor of two. That’s a really big problem.
Now, layer in ongoing geopolitical friction and critical transit constraints across key maritime bottlenecks like the Strait of Hormuz and the Red Sea. With a major slice of global oil supplies and critical LNG transit throttled by kinetic conflict, vacation-season market assumptions of quick normalization face a harsh test. When virtual AI infrastructure demands meet a constrained physical energy grid under the pressure of active geopolitical supply shocks, the underlying structural tension remains unresolved.
The Inevitability of Deceleration
If the core premise of that article is correct—that merely a deceleration in the rate of growth is enough to crash an AI bubble akin to the 2008 housing crash—then that deceleration in terms of that model becomes a very real threat. Energy needs, perhaps, simply cannot keep pace with computing scale demands. Add a potential severe structural spike in energy costs from wartime constraints, and the process becomes exponentially riskier.
The Counterfactual: State-Backed Intervention
The Zero Hedge article considers certain counterfactuals to its own argument, but we must add one more critical variable: AI is structurally linked to state-based intelligence needs and power structures.
As I have noted before, it is entirely possible that major state actors will simply not allow AI to crash the way traditional market bubbles do. The real question is whether major governments with deep pockets can forestall such an outcome, or whether—like in 2008—they will ultimately have to step in to save the market and exert direct control post facto.
Conceptually, these are vital issues the outcome of which are in no way certain but that every market actor should take very seriously.
Disclaimer:
The content on this blog is provided for informational, educational, and entertainment purposes only. It represents the author’s personal opinions and should not be construed as financial, legal, investment, trading, or tax advice.
By using this blog you agree that:
–You assume full responsibility for any outcomes resulting from the use of this content.
–You are solely responsible for all investment decisions and any gains or losses you incur.
–Past performance is not indicative of future results.
—Under no circumstances shall the author assume any liability for financial losses, capital allocation damages, or adverse market outcomes resulting from positions taken by readers, who execute any investments or trades entirely at their own risk and sole discretion.
Any discussion of specific securities, sectors, or strategies is offered strictly for educational purposes and does not constitute a recommendation to buy, sell, or hold any investment. You should always:
Consult a qualified professional before making any financial decisions.
Consider multiple viewpoints and counterarguments.
Conduct your own research
I do not collect analytics or advertising data; any such information is gathered and processed by WordPress (including Jetpack) and Blaze Advertising under their own policies.