Hedge Funds’ Record Tech Underweight Sets Stage for Short Squeeze as Anthropic’s Blockbuster IPO Looms
TREE NEWS reports: Hedge funds have slashed their exposure to technology stocks to near five-year lows, creating conditions for a powerful short squeeze that could be triggered by Anthropic’s record-breaking initial public offering. The firm’s prime brokerage data shows tech now accounts for just 18% of hedge fund net exposure versus a 29% benchmark weight—an 11-percentage-point underweight that Rich Privorotsky, head of the trading desk, described in a Thursday morning note as “extreme.” Nasdaq futures rose nearly 2% in early trading Thursday.
Positioning Extremes Meet a Catalyst
The underweight follows two months of aggressive de-risking in AI-related names, with investors rotating into other sectors. Privorotsky argues that with tech already heavily sold off, any stabilization could force managers to buy back positions. Technical factors reinforce the setup: the S&P 500 is clinging to its 200-day moving average, market makers’ gamma exposure is extremely short, and volatility has recently calmed—a combination that could become self-reinforcing as stability attracts fresh capital.
Anthropic’s IPO adds a powerful catalyst. The AI company priced shares at $83, above its $70–78 range, and upsized the deal from 825 million to 980 million shares. The offering raised approximately $81.3 billion, or $107.3 billion including shareholder sales—43% above the previous global record set by SpaceX in June. Demand was intense: the order book was six times oversubscribed, retail investors were allocated up to 20%, and pre-IPO perpetual futures on Hyperliquid traded near $99 on Tuesday, a 19% premium to the offer price. Vanda Research expects about $250 million in net retail buying on the first trading day.
Liquidity Squeeze Across Markets
Traders attribute this week’s sharp selloffs in gold, cryptocurrencies, copper and Nvidia shares partly to institutions raising cash for the Anthropic deal. The tech selling and the record fundraising may be two sides of the same liquidity adjustment. Privorotsky also flagged valuation risk: at $83, Anthropic is valued at roughly 18 times its $81.4 billion annualized revenue, while its third-quarter growth rate has halved and the company carries nearly $900 billion in committed spending, mostly for computing power. Notably, some of its largest shareholders are also suppliers providing that compute.
Bonds Signal Growth Fears Despite Hawkish Fed
While Fed Governor Waller reiterated his call for a 50-basis-point rate hike and meeting minutes showed “almost all” FOMC members favor further tightening, bond markets are rallying at the long end, with the yield curve bull-flattening. Weak Australian employment data—payrolls fell 43,500 in October versus expectations for a 25,000 gain, following a 39,300 drop in September—pushed the unemployment rate to 4.9%, the highest since 2021. Two-year Australian yields fell sharply as rate-cut expectations rose. Privorotsky sees global bonds pricing in a growth slowdown, driven by capex downgrades, falling commodity prices and easing inflation pressures. He compares the environment to late 2018, when the Nasdaq 100 fell about 25% from its peak before the Fed pivoted—a reminder that positioning repairs can unfold alongside macro stress.
Key Takeaways for Investors
- Watch positioning, not just fundamentals: Extreme underweights in tech create squeeze potential, but timing depends on a catalyst—Anthropic’s IPO could be it.
- Liquidity is driving cross-asset moves: The IPO’s massive fundraising may be draining capital from gold, crypto and commodities in the near term.
- Valuation risk is real: Anthropic’s 18x revenue multiple and decelerating growth warrant caution even as demand surges.
- Bonds are diverging from the Fed: The curve’s bull flattening suggests markets doubt the hawkish path and are positioning for slower growth.
- Macro headwinds persist: A tech rebound may be tactical rather than a return to a risk-on regime while the Fed remains hawkish and growth cools.




