Wall Street Warns of Crowded AI Trade as Short Sellers Use Derivatives to Hedge Systemic Risk
Miles Bennett
Institutions, retail, and leveraged ETFs are all piled into AI. Goldman warns herding risk is flashing red — if the crowd heads for the exit, $9 billion in forced rebalancing and $21 billion in vol-driven selling could turn a pullback into a stampede.
Everyone is buying the same thing — what's the problem?
Goldman partner Bobby Molavi flags that active funds, passive index funds, hedge funds, and retail are all adding AI exposure at the same time.
This means → when every portfolio looks the same, every sell trigger fires at the same time — and there is no one on the other side to absorb the pressure.
Cross-strategy correlation keeps climbing; the market has shifted from "diversified risk" to "collective bet."
Why are leveraged ETFs the biggest detonator?
Roughly $200 billion in leveraged ETFs exists today; tech and momentum products account for 85% of that total.
In plain terms = leveraged ETFs are built to chase: they buy more as prices rise and sell more as prices fall — no judgment, just amplification.
Under deep negative gamma (a mechanism where options hedging automatically amplifies swings), every 1-percentage-point move in the underlying index triggers about $9 billion in same-direction rebalancing.
This reflects a strong positive feedback loop: price swings create more trading, and more trading creates bigger swings.
Can a routine round of profit-taking trigger a chain-reaction stampede?
Nomura's cross-asset strategy team finds that even normal profit-taking could set off a cascade of delta-hedging — forced selling by market-makers adjusting their risk exposure.
If the S&P 500 swings 1% daily for two consecutive weeks, volatility-control funds would be forced to dump roughly $21 billion.
This means → the threshold for a stampede is not high — no "black swan" required; an ordinary wave of selling could light the fuse.
Is the smart money already buying insurance?
Lee Robinson, the hedge-fund manager who captured a 900% return during the 2008 financial crisis, is ramping up bearish bets.
His target is not the $1.8 trillion private-credit market itself — there is no easy way to short it directly. Instead he is shorting its main funding source: insurance companies such as Lincoln National, MetLife, and Berkshire Hathaway, via credit-default swaps (CDS — contracts that pay out if a borrower defaults).
The trade is spreading: multiple hedge funds have followed, and Goldman and JPMorgan have rolled out derivative tools for clients. Net notional CDS shorts on U.S. insurers rose from under $49 billion at year-end to $55 billion as of May 22.
How deeply are insurers tied to private credit?
Moody's data: at end-2025, 20% of the U.S. life-insurance industry's $4 trillion in fixed-income assets sat in illiquid private credit.
In plain terms = insurers are lending policyholders' premiums into high-yield loans. If borrowers can't repay, the insurers' balance sheets take the hit.
Credit quality in the mid-market direct-lending space is weakening and borrower debt-service pressure is rising — impairment charges are increasingly likely.
Can Micron's earnings keep the AI valuation alive?
Micron's stock has surged 269% this year; the single name accounts for nearly one-fifth of the S&P 500's 2026 gains.
Consensus expects quarterly net income of $23.8 billion and revenue of $35.6 billion; the four cloud giants plan a combined $725 billion in 2026 capex.
Yet options pricing implies a post-earnings swing of up to 10% — and Aptus Capital's portfolio manager says expectations are maxed out with almost zero margin for error.
This means → Micron's report is not just one company's scorecard — it is a stress test for the entire AI-hardware valuation chain. Any miss on results or guidance could trigger a chain sell-off across tech hardware.
Content is for reference only, not financial advice.