Fund Manager Who Called Rally Warns: The Great Rotation Is About To Begin

ELI5/TLDR

Hedge fund manager Thomas Hayes thinks the stock market right now is a coiled spring held in place by one event: the Iran war. While the war drags on, oil and inflation fears are high, so investors have crammed all their money into AI stocks — the only thing that seems safe in a slow-growth world. Hayes argues the war will end soon (an election is coming, the President needs a win), and the moment it does, oil falls, interest-rate fears fade, and money floods out of crowded AI names into everything that’s been ignored — boring “defensive” stocks like alcohol, food, and auto parts. He’s not calling a crash, just a rotation: same money, different pockets.

The Full Story

The setup: one war is holding the whole market hostage

Hayes splits the year into two halves. In January and February, his earlier call was paying off — the “other 493” stocks in the S&P 500 (everything outside the handful of giant tech names) were outperforming. Then the Iran war started, and that trade died.

His explanation is mechanical. War pushes oil up. Oil pushes inflation expectations up. Higher inflation expectations push bond yields up, because the market stops expecting interest-rate cuts and starts pricing in hikes. And almost no business does well in that environment — slow growth, high borrowing costs — except the AI names, which investors treat as recession-proof.

“Since the Iran war started out of left field, nothing has worked other than the AI trade.”

So money piled into AI not out of pure enthusiasm but as a hiding place. That’s the key reframe: the AI rally is partly a fear trade.

What’s actually under the hood of “record earnings”

This is the most useful stretch of the interview. The market is celebrating big earnings growth from the AI giants (Hayes cites roughly 24% this year). But he argues a large chunk of that isn’t real operating profit — it’s paper gains.

The giants (Alphabet, Amazon, and the other “hyperscalers” — the companies that own the massive data centers AI runs on) hold equity stakes in AI startups like Anthropic. When those startups are valued higher, the giants get to mark up the value of their stakes and book it as income, even though no cash changed hands. Hayes’s numbers:

“60% of Alphabet and Amazon’s combined net income came from these windfalls which is simply paper marked to markets. This is not their operating improvement.”

In plain terms: the engine looks more powerful than it is, because a lot of the horsepower is accounting, not operations. He expects next quarter’s earnings to expose this, because the paper gains won’t repeat and the real return on all that AI spending won’t show up for another 12–18 months.

The “ducks are quacking” tell

There’s a Wall Street saying: “When the ducks are quacking, you feed them” — meaning, when buyers are desperate to buy, sellers should sell. Hayes sees this everywhere. Anthropic filing to go public near a $1 trillion valuation (on ~$35bn revenue — about 35x sales). SpaceX at a reported 100x sales while losing money. And most tellingly, Alphabet selling $80 billion of its own stock — the largest equity raise in US history — to fund AI data centers.

That last one is his loudest warning. Companies sell stock when their stock is expensive and they’ve run out of cheaper money:

“When you see companies like Google that can no longer finance their capex commitments through cash flow or through debt issuance and they now have to dilute their shareholders by $80 billion raising equity at the top. It’s something to pay attention to.”

He’s careful not to call it a bubble — “there’s froth in it in price relative to underlying value in the short term.” His baseball metaphor: we’re in the “fourth or fifth inning” of a nine-inning game. Middle of the story, not the end. Just due for an intermission.

Other froth signals

He rattles off classic late-cycle tells: retail call-option buying at the highest level since the 2021 peak (right before the 2022 tech crash), record margin debt (investors borrowing to buy), and a “collapsed put skew” — meaning downside insurance is record cheap because nobody wants it.

“The time to buy insurance is before the house is on fire. No one wants to buy insurance. No one believes there’s going to be a fire.”

The trigger: a tweet ends the war

Hayes’s whole thesis hinges on one prediction — the Iran war ends soon, because politically the President can’t afford it dragging into an election. Whether it’s a deal, a withdrawal, or a declared victory doesn’t matter to him:

“This is going to be a man-made created resolution. And the minute that resolution is tweeted… oil is going to be down, yields are going to be down, the AI trade is going to be down, and the everything trade… is going to rip higher.”

This is “buy the rumor, sell the news” — the AI trade has rallied on war fear, so it sells off when the fear resolves, while everything that was abandoned snaps back. He calls it “January and February part two.”

Where he’s putting money

The actual rotation call. Defensive sectors — staples, healthcare, utilities — are at a 25-year low weighting in the S&P (down from 35% to 16%), even lower than at the March 2000 dot-com peak. His historical rhyme: in 2000, Buffett looked washed-up buying Coca-Cola and P&G while everyone chased Cisco and Micron. Then the NASDAQ fell 80% and Buffett had three of his best years.

Specific names he likes (all “turnaround” stories — beaten-down companies cutting costs and returning to growth):

Diageo — alcohol giant, $3bn free cash flow, new management, betting premiumization beats the “nobody drinks anymore / everyone’s on Ozempic” fear.
Alibaba — “the cheapest way to play AI globally,” with the same hidden equity-stake upside as the US giants, plus a weak-dollar/emerging-markets kicker.
Hormel — protein foods, 65 straight years of dividend increases, ~6% yield, turning the corner.
Advanced Auto Parts — sold off Canada, closed bad stores, back to growth; he thinks it could triple.
Estée Lauder — beauty, recovered from COVID overstocking, shifting to online/TikTok shop.
Taylor Morrison (homebuilder) — notes Berkshire bought it the same day it bought Google, calling it the smarter of the two bets, since housing (18% of GDP) booms when rates normalize.

The Fed wrinkle

He expects the new Fed chair (referred to as “Walsh”/“Worsh,” i.e. Kevin Warsh) to cut rates before year-end — a non-consensus call, since the market is betting on a hike. His logic: once the war ends, inflation fear fades and a cut becomes possible. He also flags the bear case honestly — if the Strait of Hormuz stays closed, inflation hits 5%, yields hit 5.5–6%, small caps get “decimated,” and a recession starts. He just thinks that’s improbable because it’s politically self-destructive.

Key Takeaways

The AI rally since the Iran war began is partly a defensive trade — investors hiding in the only sector seen as recession-proof, not pure conviction.
A large share of the AI giants’ recent “earnings growth” is paper gains from marking up equity stakes in AI startups, not operating profit. Hayes claims ~60% of Alphabet + Amazon’s combined net income came from such mark-to-market windfalls.
Alphabet’s $80 billion equity raise — the largest in US history — is Hayes’s biggest red flag: it means a cash-rich giant has exhausted cash flow and cheap debt and is now diluting shareholders at peak valuations.
Anthropic filed confidentially for an IPO near a $965bn–$1tn valuation (~35x sales); SpaceX cited around 100x sales while losing money.
Late-cycle froth signals: retail call buying at highest since 2021, record margin debt, and record-cheap downside protection (collapsed put skew).
Hayes is NOT calling a crash — he says we’re in the “fourth or fifth inning” of the AI story, due for an intermission, then a resumption.
The rotation trigger is the Iran war ending, which he treats as near-certain before the election. Resolution = oil down, yields down, AI down, “everything else” up.
Defensive sectors (staples, healthcare, utilities) sit at a 25-year-low weighting in the S&P 500 — 16%, below even the March 2000 peak.
He expects the Fed (Warsh) to cut rates before year-end — against consensus, which expects a hike (poly-market put a Fed hike before 2027 at ~38%).
His stock-picking frame: “good businesses at great prices versus fair businesses at ridiculous prices” — turnaround names like Diageo, Hormel, Advanced Auto Parts, Estée Lauder, Alibaba.
Bear case he concedes: if the Strait of Hormuz stays permanently closed — 5% inflation, ~6% 10-year yield, recession, Democrats sweep the fall elections.

Claude’s Take

The thumbnail-grade framing — “fund manager who called the rally warns” — is the standard finance-YouTube setup, and it should make you read with one eyebrow up. The whole thesis rests on a single binary prediction Hayes states with near-certainty: the Iran war ends soon because the President needs the win. If that’s wrong, every downstream call (rate cuts, the rotation, the snap-back) collapses. He waves at the bear case but treats it as nearly impossible, which is exactly the kind of confidence that doesn’t survive contact with reality. Predicting a man’s tweet is not analysis.

That said, the middle section is genuinely worth the time, and it’s the part that doesn’t depend on the war prediction. The point about AI-giant earnings being inflated by mark-to-market gains on startup stakes is a real, checkable accounting observation — if you only remember one thing from this, make it that. The “$80bn equity raise = they’ve run out of cheap money” read is also sharp. These are structural facts, not vibes, and they’d matter even if the war ran another year.

The stock picks are where to be most skeptical. He runs a fund and was on the show partly to talk his book — every name is a “turnaround story” he presumably owns, narrated at its most flattering (Diageo “returning to glory,” Advanced Auto Parts “could be a $150 stock”). Treat these as leads, not recommendations.

Score: 6. Better than the average doom-thumbnail interview because the earnings-composition argument is substantive and the rotation logic is internally coherent. Marked down because it’s one giant conditional bet dressed as multiple insights, it’s lightly promotional, and the confidence wildly outruns the evidence on the load-bearing prediction.

Fund Manager Who Called Rally Warns: The Great Rotation Is About To Begin | Thomas Hayes