A few morning things worth noting. Topics:
- Recent media interview.
- Michal Burry's new Nvidia rodeo
- Bragawatts and the O.K. Corral
- The 5% of China argument
Let's do this:
- Me, Me, Me
I was interviewed here and there in the last week on capex exuberance and related topics, perhaps most notably in The Atlantic.
2. Michael Burry's Nvidia bearishness
Michael Burry of "Big Short" fame is in the news today, with a filing showing that he again has puts against AI darling Nvidia (as well as against who-knows-what-it-does darling Palantir). As Bloomberg reminds us, this is not Burry's first anti-Nvidia rodeo, having bet against the company earlier this year, which didn't work out.
Of course, this is only the visible part of his portfolio, so it is possible that what looks like a bet against Nvidia is swamped by other long positions in the stock, perhaps hugely fun synthetic securities he has obtained from investment banks, but ... that seems unlikely.
Does it matter that he is betting against the company? Probably not. While Nvidia is closer to the end of its current run than to the beginning—such is the law of large numbers—it is not clear Burry has a particular edge here, even if it's nice to see he can't stay out of the game.
I have little patience for most private equity musings, as readers will know, but I will concede that KKR calling speculative multi-gigawatt data centers "bragawatts" is something I wish I had come up with. Nevertheless, it is a handy description.
It is handy because it captures the idea that much of what is going is a kind of expensive posturing. Companies try to outdo one another with announcements of ever-larger data centers requiring ever more power. Will they ever be built? Who knows, but if they deter other entrants, then at least some of the mission is accomplished.
But such mass deterrence strategies are inherently fragile disequilibria. If everyone shows up at the O.K. Corral armed to the don't-mess-with-me teeth, the result isn't peace, it's a gunfight. Or, perhaps, it is more like the old joke about always being sure to carry a bomb onto an airplane, because what are the odds of there being two bombs? amirite.
The trouble, of course, is what happens in the aggregate: all these pseudo-rational behaviors incite more such behaviors, leading to, at best, what finance theorists call a "rational bubble". And bragawatts is as good a way as any of capturing that in compressed form.
- Getting 5% of China's market
People try to justify current AI spending to me on myriad bases, ranging from faith-based investing (this will work out because it always works out), to AGI (can you tell me the TAM for superintelligence? no? checkmate!), and more. I wish I had a bingo card.
Among my favorites, however, is the "5% of China" argument. It has to do with capturing a slice of the impending global productivity benefits from AI, and it generally goes something like this:
Step 1: Global GDP baseline (2025): ≈ $110 trillion nominal.
Step 2: 20% productivity gain from AI: 0.20 × $110 T = $22 trillion in incremental output.
Step 3: AI tools/services capture 10% of that: 0.10 × $22 T = $2.2 trillion annualized value.
Step 4: Stand back and make bank.
This is, of course, ludicrous. It is on par with people who say, longingly, if they could just get 5% of the Chinese market for [insert thing here] they'd have a multibillion dollar company.