By Jae, PANews
U.S. stocks went through a violent two-session reversal. On Nov. 5, the market suffered its strongest one-day selloff in nearly 19 months, led by the artificial-intelligence supply chain. Social media quickly filled with claims that hedge funds were being wiped out in a broad wave of liquidations.
The reversal came the next day. Chip stocks led the recovery on Nov. 6, with the Philadelphia Semiconductor Index rising 4.2% and AMD gaining 11%. AI-linked names including Nvidia also recovered. The Nasdaq Composite rose 1.1%, while the S&P 500 added 0.3%.
The rebound had some warning signs in the previous session. At the same time, capital was moving across asset classes.
Forced selling eased, but the larger risks remain
The Nov. 6 rebound began to show itself in the final 10 minutes of the previous session. Vanda Research data showed that U.S. retail investors turned noticeably toward buying near the close, recording more than $100 million in net purchases during those final 10 minutes.
Leveraged exchange-traded funds were another important variable. Nomura estimated that leveraged ETFs needed to sell about $47 billion in stock exposure before the close to complete their net-asset-value rebalancing after the Nov. 5 move. It was the second-largest one-day sale in the firm’s records, which date back to 2006.
Once that large volume of passive selling had been completed, short-term pressure eased. Dip-buying, short covering and institutional portfolio rebalancing then combined to support a price recovery. JPMorgan said in an after-hours comment that the sharp deleveraging on Thursday had cleared some crowded trades and left the market’s positioning structure healthier.
The rebound was therefore more consistent with a technical repair in short-term trading. A shift in policy expectations and doubts about the returns on AI capital spending were not resolved by the move higher.
Claims of a hedge-fund wipeout overstated the October deleveraging
Hedge funds became the main focus of attention during the latest shock in U.S. equities. Bloomberg reported on Nov. 5 that Tiger Global’s flagship hedge fund lost about 8.5% in October, leaving its year-to-date return roughly flat. Walleye Capital fell about 5.9% during the month and was down about 6.8% for the year. Lone Pine Capital lost about 8.8% in October.
Social media then amplified a panic narrative about Wall Street hedge funds being wiped out together. Reports that Walleye Capital, Engineers Gate and Samlyn Capital had cut teams, streamlined strategies or returned capital to investors added to concerns about a chain reaction of institutional liquidations.
Business contraction, however, is not the same as a forced liquidation or a fund failure. The moves were described as active risk control and changes to operating models, rather than bankruptcy proceedings caused by insolvency.
The timeline also matters. Hedge funds had begun cutting positions and reducing leverage in October. The Nov. 5 plunge may have represented the concentrated realization of that process rather than the start of a single-day collapse.
Goldman Sachs Prime Brokerage data showed that global fundamental equity long-short funds lost about 3.5% on average in October, while systematic long-short strategy funds lost about 3.6%. It was the first time since Goldman Sachs began tracking the two groups in 2016 that both strategies had lost more than 3% in the same month.
A less visible adjustment took place in mid-October. Over a two-week period, Goldman’s hedge-fund clients were net sellers of $20 billion in technology stocks. In the five trading sessions from Oct. 8 to Oct. 14, they erased all of the net long position in AI stocks accumulated since the start of 2025.
By the end of October, hedge funds’ net leverage had fallen to 73.4%, down 4.2 percentage points for the month. That was the largest monthly decline since March of this year.
Morgan Stanley data showed a similar shift. Stock-picking hedge funds experienced their largest active reduction in exposure in more than a decade during October. Their market exposure had fallen by about 9% in total since June. By the end of October, more than 40% of the funds in the group had either a flat or negative return for the year.
Performance was not uniform across the industry. Coatue Management posted a 0.2% gain in October, Citadel rose 1.7% and Quantedge gained 7.2% for the month. Some managers were caught in crowded trades; others made money from the volatility.
Capital moved rather than disappeared
The loss of trillions of dollars in U.S. equity market value was accompanied by a large shift in capital. Money moved out of high-volatility, highly leveraged technology stocks and into assets viewed as safer and more liquid.
U.S. Treasuries were the first major destination. On the day of the Nov. 5 equity selloff, the 10-year Treasury yield briefly fell to 4.84%, its lowest level since September. Because bond prices and yields move in opposite directions, the move pointed to a large inflow into the Treasury market.
The dollar was another recipient of demand. Alex Kuptsikevich, chief market strategist at FxPro, said the technology-stock panic had weakened risk appetite and brought substantial safe-haven buying back to the dollar.
Chris Turner, global head of markets at ING, also said investors currently preferred to keep cash temporarily in dollar assets. In his view, the market was moving closer to dollar cash rather than moving away from the dollar.
The Korean won and the New Taiwan dollar, which had been hit by risk sentiment, also drew renewed attention as the dollar came under demand. Claudio Piron, head of Asia foreign-exchange and rates strategy at Bank of America Global Research, said the two currencies presented buying opportunities after the AI-stock selloff.
Piron pointed to the sizable current-account surpluses of Taiwan and South Korea as a strong base for investors seeking to establish long positions. Another Bank of America view was that AI supply-chain companies had already built substantial capital expenditure plans and export orders. Those orders would not be canceled simply because U.S. stocks had fallen.
Continued export-related foreign-exchange conversion over the next six to 12 months could provide a new underlying source of support for the New Taiwan dollar and the Korean won.
The selling peak may have passed, but the position unwind may not be over
A fast rebound after a crash is not unusual. The harder question is whether the move signals the end of the decline or only a temporary recovery.
August 2007: quantitative-fund deleveraging
From Aug. 6 to Aug. 9, 2007, several equity long-short quantitative funds suffered concentrated losses as markets experienced severe dislocation. Research by MIT scholars Amir Khandani and Andrew Lo found that the concentrated unwinding of crowded trades, combined with a temporary retreat by market-making capital, contributed to the abnormal liquidity conditions.
The most intense phase of the decline lasted only four days. Some strategies began to recover after Aug. 10, but the broader deleveraging process continued from July through the end of the year. The sharp fall was only the opening phase; clearing positions and repairing risk appetite could take months.
March 2020: a broad selloff in risk assets
From Feb. 19 to March 23, 2020, the S&P 500 fell about 34% in total. The adjustment lasted more than a month and included several violent one-day rebounds. The index did not recover its previous level until August of that year.
In the near term, the first two sessions of the latest U.S. stock move look closer to the first example. The most concentrated selling may have passed, but a broad portfolio rebalance and the repair of risk appetite are not complete.
That is where the market’s disagreement lies. Investor Jason Calacanis said a market bottom may already have appeared, while adding that technology-stock valuations could still fall another 10%. He described that point as an attractive buying opportunity.
Tom Lee, chairman of Bitmine, offered a more optimistic view. He said Thursday’s U.S. stock selloff looked like an aftershock from a 「blue tsunami」 triggered by the result of a midterm election. He described it as a combination of 「last year’s tariff shock + the DeepSeek moment + this June’s Situtational Awareness liquidation」. In his view, concentrated selling of this type often runs its course within days and creates buying opportunities.
U.S. stock analyst Mark Minervini was more cautious. He said the decline might not reach the scale of the 1987 crash, but could lay the groundwork for some form of market collapse. That collapse could last one to two weeks. If the decline intensified, the worst-case scenario for SPY would be a 10% to 15% fall from its high, while the Nasdaq could decline 20% to 25% and enter a cyclical bear market similar to 1998.
For now, the Nov. 6 gain was not large enough to offset the previous session’s decline. An asset that falls 20% must rise 25% to return to its starting level. The Nasdaq fell 5.52% on Nov. 5 and gained 1.1% the following day, leaving it well below its earlier level.
After two sessions of sharp volatility, U.S. stocks are facing a new test. The narrative that drove AI technology stocks higher for months is being tested by the delivery of industrial policy and questions over the returns on AI investment. Short-term leverage reduction may be nearing its end, but the medium-term debate over those issues is still underway.
