Buyers disappear amid the AI sell-off! Goldman Sachs' trading desk phone is "blown up" but no one is taking over, hedge funds and retail investors are all withdrawing

On Thursday, November 5, Eastern Time, as technology stocks in the U.S. faced a fierce sell-off, the phone lines at Goldman Sachs' trading desk were nearly "overwhelmed," with investors calling in to inquire how much further the market might drop, but there were very few willing to buy on the dip.

Goldman trader John Flood described in a market commentary sent to clients that Thursday's trading environment was completely opposite to the previous day: Wednesday was one of the busiest trading days of the year for the bank's stock trading desk, while Thursday was "completely the opposite," with traders receiving numerous calls about market conditions but hardly seeing any buying interest.

This description reflects a sudden cooling of AI trading, with market sentiment shifting rapidly from chasing gains to seeking safety. The Nasdaq Composite Index fell 5.52% on Thursday, the Philadelphia Semiconductor Index plummeted over 12%, and the overall market capitalization of technology stocks evaporated by about $2.5 trillion. Meanwhile, volatility surged, retail investors concentrated on selling, and hedge funds reduced their positions, further amplifying the market's decline.

### Market Reversal in the Eyes of Goldman Traders: Active Buying on Wednesday, Urgent Retreat on Thursday

Flood noted that after the market rose on Wednesday, some investors who had significantly reduced their risk exposure ahead of the U.S. midterm elections re-entered the market and bought back some stocks. However, on Thursday, the market suddenly reversed, and some hedge funds began selling stocks they had just bought the previous day, particularly large tech stocks and mega-cap tech companies.

He pointed out that long-term investment institutions were once again in a wait-and-see mode, and there were no buyers willing to step in to support the market. At the same time, brokers catering to retail clients also reflected that selling pressure was significantly greater than buying for the second consecutive day.

Trading activity also showed unusual patterns. Flood stated that as of Thursday's intraday trading, the trading volume of the S&P 500 ETF SPY was about 160% higher than the average of the past 20 trading days, the trading volume of the Nasdaq 100 Index QQQ was about 100% higher, the semiconductor ETF SMH's trading volume was about 150% higher, and the trading volume of DRAM-related stocks was about 75% higher.

The surge in trading volume reflects that investors are intensively adjusting their positions, but it does not mean that all trades are net selling. Combined with the lack of buying interest described by traders, a more apparent signal from the market is that when stock prices fall rapidly, the funds that might have stepped in to support the market chose to wait.

### Institutional Selling Pressure Concentrated, Retail Investors' Single-Day Net Selling Reaches Nearly One-Year High

The retreat of buying interest observed by Goldman’s trading desk was also corroborated by data from other institutional trades.

According to Reuters, UBS trader data showed that on Thursday, the active trading flow of the bank's institutional clients was clearly biased towards selling, with selling pressure mainly coming from hedge funds, particularly in technology stocks, especially in the memory chip and semiconductor sectors, which faced concentrated sell-offs. UBS traders described the trend of emerging cloud service providers and semiconductor stocks as a clear "surrender-style sell-off."

Goldman traders also pointed out that on Thursday, hedge funds were selling some large tech stocks that they had just bought the previous day. The market not only saw internal rotation within sectors but also investors directly reducing their overall risk exposure.

Retail investors also joined the retreat. Data from Vanda Research indicated that on Thursday, the net selling scale of retail investors in individual stocks was expected to set a record for the largest single-day net sell since the pandemic began. Data from JPMorgan as of 11:30 AM Eastern Time showed that when including ETFs, the net selling scale of retail investors reached about $1.3 billion, the highest in the past year.

Notably, retail investors typically buy stocks to "catch the bottom" during market declines, but this time, the buying interest in individual stocks has significantly weakened. Vanda Research strategists pointed out that in traditional downtrends, retail investors often buy on dips, but on Thursday, such buying interest had effectively disappeared from individual stocks.

### VIX Surges to Eight-Month High, Investors Rush to Buy Protective Options

Beyond the stock sell-off, the demand for investors to hedge market risks also rapidly increased.

The "fear index" that measures market volatility—the Chicago Board Options Exchange Volatility Index (VIX)—rose by as much as 11.75 points during Thursday's trading, reaching 28.36, an eight-month high, marking the largest intraday increase since the U.S. imposed reciprocal tariffs in April 2025; it closed at 28.18.

Changes in the options market also indicated that investors were buying protection against further declines. Reuters cited data from options analysis firm Mercury Alerts, stating that on Thursday, the premiums for put options on major index products were about twice the usual level, reflecting a significant increase in demand for downside protection.

Steve Sosnick, chief strategist at Interactive Brokers, stated that market sentiment changed significantly in just one day: on Wednesday, investors felt there was no need to buy protection, but on Thursday, they suddenly realized the need to hedge risks.

Additionally, leveraged and inverse ETFs, as well as some quantitative strategies, typically adjust their positions based on market volatility as the close approaches, which may further amplify end-of-day fluctuations. Nomura Securities estimated that leveraged ETFs might need to sell about $47 billion in stocks before Thursday's close, marking the second-largest such sell-off since the firm began tracking related data in 2006, with selling pressure primarily concentrated in tech and AI-related funds.

This means that some of the selling pressure did not entirely stem from investors actively changing their long-term views but could also be a result of market declines triggering strategic reductions, thereby creating further downward pressure.

### Concerns Over Data Center Construction and AI Investment Returns, Sell-Off Spreads from Chips to the Supply Chain

The direct backdrop of this round of sell-offs is that AI-related stocks had previously accumulated a large amount of capital and high market expectations, while the results of the U.S. midterm elections cast a shadow over the prospects for data center construction.

After the Democratic Party achieved significant victories in the midterm elections, the market began to worry that local governments and regulatory agencies might impose stricter approvals, energy, and environmental requirements on data center projects, thereby slowing down the construction of AI infrastructure in the U.S. Semiconductor, memory chip, emerging cloud service providers, and data center power suppliers were thus impacted.

Goldman trader Flood also raised a more specific concern: by 2028, the market may face shortages of land, power, and construction capacity for data centers. If the pace of data center infrastructure construction does not keep up with chip deliveries, the market could shift from worrying about insufficient chip supply to worrying about chip oversupply and a lack of sufficient data centers to accommodate them.

This also explains why Thursday's sell-off was not limited to chip manufacturers. Data center construction contractors, power suppliers, and emerging cloud service providers also faced significant selling pressure. Investors began to reassess the pace of expansion in the AI supply chain and whether massive computing power investments could be converted into actual revenue and profits as expected.

Meanwhile, recent earnings outlooks from Qualcomm and Arm failed to meet market expectations, further undermining investor confidence in the semiconductor sector. The logic of long-term growth in AI demand has not disappeared, but the market's requirements for the speed of growth realization, capital expenditure returns, and profitability across various segments of the supply chain are increasing.

For the market, the key question has shifted from "How much more will AI investment grow?" to "Can these investments be completed on time, obtain sufficient power, and generate returns that support current valuations?" When investors simultaneously choose to reduce positions, hedge, and wait for buying opportunities, the phones at traders' desks are overwhelmed, yet buyers are nowhere to be seen, becoming the most intuitive reflection of market sentiment during this round of AI sell-offs.

Risk Warning and Disclaimer

The market has risks, and investment should be cautious. This article does not constitute personal investment advice and does not take into account individual users' specific investment goals, financial situations, or needs. Users should consider whether any opinions, views, or conclusions in this article align with their specific circumstances. Investing based on this is at one's own risk.