
There is a specific kind of poker game that only happens at a table full of teenagers, and if you have ever sat at one, you know it well. The chips don’t mean anything. Nobody at the table has rent to pay. The buy-in was ten dollars that somebody’s mom fronted, and the pot could be a small fortune or a fun-size Twix — it changes nothing about how the hand gets played. What these players are protecting is not their money. It is their standing.
Which produces one iron law: you cannot bluff. Not really — not the way it works in the movies, where some James Bond-type slides his stack forward on seven-high and watches a better hand muck. Push your stack forward at a high school table and someone will call you. Not because they’ve read you. Because they refuse, on principle, to let you get away with it. The possibility that you might be lying and win anyway is physically intolerable to them. So they call with nothing. They call with worse than nothing. They will set fire to their entire stack for the pleasure of flipping your cards face-up and being right.
The math of this is brutal and predictable. If a bluff never works, the only winning bets are made hands, so the bluffers go broke, sure. But the callers go broke faster. They bleed chips into pot after pot chasing the satisfaction of the reveal, calling down king-high because someone had the nerve, and mistaking that reflex for toughness. They think they are hard to fool. What they actually are is easy to provoke.
I have started to think that a large and very loud part of the market is playing the AI trade exactly like a sixteen-year-old with a $100 pot.
June was the worst stretch for tech in months; the Magnificent Seven shed something on the order of $2 trillion in market value in a matter of weeks, and the same chip names and hyperscalers that had led the melt-up led the way down. Ray Dalio was on television with his bubble indicators flashing 1999-and-1929 comparisons. Jamie Dimon was muttering about froth. Goldman’s own head of equity research has been running the in-house bear case for two years. And Michael Burry — the man whose entire brand is being right while everyone else is wrong — has been shorting the rally and posting charts to prove the point, blaming the whole thing on “fear of missing out, the greater fool theory, and public commitment bias.”
Sit with that last phrase. Public commitment bias. Burry means it as an indictment of the bulls, and fairly. But it is also the most precise description I know of the person calling every bluff at the teenage table. The need to be publicly right. The need to have called it. The need to not let them get away with it.
There is a cliché I want to use backwards on purpose. You have heard that a person can miss the forest for the trees — get so lost in the detail that they lose the whole. The AI bears have the opposite disease. They are missing the trees for the forest. They have arrived at one enormous, satisfying, categorical judgment — this is a bubble, this is hubris, the emperor is naked, we cannot let them get away with it — and in service of that one big call they have stopped looking at the individual trees entirely. The specific company. The specific margin. The specific cash flow. It has all collapsed into a single word, and the word is “bad.”
Now — and this is the part that separates analysis from pride — the bears are not wrong about everything. There is a genuinely bad hand somewhere in this pot, and it’s worth naming precisely. Hyperscaler capex is running toward roughly $700 billion this year while enterprise AI revenue is a fraction of that; Bank of America has flagged that capex has climbed to nearly 100% of the group’s operating cash flow, gutting the free cash flow that used to fund buybacks. Burry’s circularity point is real: chip stocks rise because the giants spend, equipment makers rise because the chip companies build capacity, and every new spending announcement gets treated as fresh proof that demand is infinite. Some of that is a snake eating its own tail. Some of the second-derivative names are running on story and acronym alone. Those are bluffs. A disciplined player folds into them, or fades them.
But that is the entire point. The disciplined move is to distinguish the bluff from the made hand. What the bluff-callers do instead is shove all-in against the whole table, because they have decided the whole table is lying. And when you do that, you end up shorting the trees along with the forest.
Look at what the trees actually are. Nvidia — the supposed epicenter of the mania — grew revenue something like 65% year over year and trades around 40-something times earnings. At the top of the dot-com bubble, Cisco traded at 472 times. You can believe the AI stocks are expensive and still notice that “expensive and wildly profitable” is a fundamentally different animal from “expensive and pre-revenue.” One is a rich price for a real hand. The other was a bet on a hand that didn’t exist.
Or look at where the real money is quietly stepping in while everyone else calls the bluff. Bill Ackman spent years saying Alphabet, Amazon, and Meta were “never cheap enough” for him. Then the market got so consumed with punishing Big Tech for its AI spending — so eager to make them prove it — that it handed him the entry he’d been waiting a decade for. Meta at roughly 18 times forward earnings. Cloud businesses running 33-38% operating margins that monetize the AI boom no matter which model wins. His thesis is almost aggressively boring: don’t try to pick the frontier lab, just own the compute everyone has to rent regardless. He points out that the market wrote off these same franchises the way it wrote off Berkshire as “old economy” in 2000 — right before the “old economy” quietly compounded through the wreckage.
That is the trade the pride-callers are missing while they’re busy flipping over Nvidia’s cards to prove a point. The value didn’t disappear. It got repriced because the table decided, categorically, that it wasn’t going to let anyone get away with anything.
Thanks for reading Passing the Torch Newsletter! This post is public so feel free to share it.
Here is what the winning high schooler eventually figures out, usually after going broke a few times. It is not the kid who is impossible to bluff. Being unbluffable is not a strategy; it’s a personality trait, and it’s a losing one, because a table full of people who know you’ll always call will simply stop bluffing and value-bet you into oblivion. The kid who wins is the one willing to look like a coward — willing to fold, over and over, and eat the occasional indignity of being bluffed off a hand — and who, when he finally pushes his chips in, has the cards to back it. His edge is mathematical. The caller’s edge is emotional, and an emotional edge is not an edge at all. It’s just the market charging you for the privilege of being right out loud.
Shorting a category because its arrogance offends you is calling with an empty hand. It might feel like conviction. Pride — at the table and in the trading book — is the single most expensive trait you can have.
So by all means, fold the froth. Fade the circular financing. Short the specific bluffs when the price and the read actually line up. But do not confuse that with the reflex to call everything, because someone, somewhere, might be getting away with it. In this game, the people who cannot stand to be bluffed are the ones who pay for everyone else’s pots.
Thanks for reading, and as always, stay curious folks!
—J&E