Hedge Funds Sold $20 Billion of Tech Stocks in Two October Weeks
Hedge funds sold roughly $20 billion of technology stocks across two weeks in October, Goldman Sachs prime-brokerage data shows, and the selling carried…
AI SummaryAI
- Hedge funds sold about $20 billion of technology stocks over two weeks in October, Goldman Sachs data shows.
- Bullish-to-bearish tech positioning fell to its lowest level since tracking began in 2016.
- 52% of fund managers expect a major cloud company to cut AI spending within six months.
- NVIDIA posted record quarterly revenue of $115.8 billion, up 103% year over year.
Hedge Funds Unloaded $20 Billion in Two Weeks
Hedge funds sold roughly $20 billion of technology stocks across two weeks in October, Goldman Sachs prime-brokerage data shows, and the selling carried on into November. The retreat left funds with their lowest ratio of bullish to bearish technology positions since that tracking began in 2016, and it effectively erased the net AI-linked purchases accumulated since early 2025. The same Goldman dataset records leverage being cut at the fastest pace since March, which means the industry was not only trimming individual positions but reducing borrowed exposure across the whole book.
The exit sits beside a striking contrast: the companies powering the AI boom are still reporting record results, yet the funds that fueled the trade have been reducing their stakes at a pace the bank's own sales data flags as unusual. Part of the judgment is about price rather than performance. Investors are increasingly asking how much they have paid for future AI growth, not whether the earnings are real. Some funds have gone further and opened fresh short positions, betting that individual technology stocks will decline from current levels. The unwind changes the supply picture for the sector's largest names. Positioning that took nearly two years to build was removed within weeks, and the November leg of the selling means the October total of about $20 billion understates the full move. A lower ratio of long to short positions also leaves the sector more sensitive to any earnings disappointment, because a smaller base of committed bulls stands ready to buy weakness. The speed matters as much as the size: selling compressed into ten trading days forces counterparties to absorb blocks that would have moved quietly if spread over a full quarter.
Bank of America Survey Flags AI Bubble Risk
Bank of America's November fund manager survey supplies the sentiment numbers behind the selling. The November 17 post summarizing the survey's headline findings records that 34% of respondents identify an AI bubble as the biggest tail risk to markets. The same survey found 52% of fund managers expect at least one major cloud company to cut AI spending within six months. Allocations tell the same story from a different angle: technology allocations suffered their sharpest monthly reduction since 2006. The concern is specific. It is about spending rather than demand, and spending is the revenue line for chipmakers and data-center suppliers downstream.
@DeItaone · X post
November 17 post summarizing the survey's headline findings.
View on X
NVIDIA results complicate the bearish reading. The company reported record quarterly revenue of $115.8 billion, up 103% year over year, a print that landed despite the uncertainty surrounding Jensen Huang's resignation. It was the largest quarterly top line the company has reported, and it arrived weeks after the October selling had already begun. Demand from cloud customers still underpins the numbers, which is why the fund selling reads as a positioning decision rather than a response to deteriorating fundamentals. The two datasets point in opposite directions and explain the market's split behavior: the earnings argue for owning the sector, while the survey argues for reducing exposure to it. The next test sits with the buyers of all that computing capacity: whether Microsoft, Amazon and the other large AI investors can convert infrastructure spending into stronger profits. If that spending slows, chipmakers and data-center suppliers would face renewed pressure on orders, whatever their current growth rates show. The spending question also sets the calendar. A cut by a major cloud operator would flow through procurement budgets with a lag, which is why fund managers frame the risk as a six-month window rather than an immediate event.
Where the Money Rotated in November
Goldman's latest weekly data shows consumer discretionary stocks, including retailers and other businesses dependent on household spending, attracting the strongest buying interest in the week ending November 12. During October, real estate and healthcare also drew hedge fund purchases, although healthcare selling resumed in November. Goldman described the recent activity as “more rotational,” and the record supports the description: the flows show a reallocation across the market, not a retreat from equities altogether. A crowded trade being unwound can stabilize once the sellers are done, so the question of whether the AI trade recovers or slides further now depends on capital expenditure guidance from the cloud companies rather than on further fund selling. For readers weighing this rotation against risk appetite elsewhere, our explainer on Cryptocurrency vs Stocks: Which Investment Is Right for You? sets out how the two markets price positioning shifts differently.
Primary sources
AI-generated, AI-reviewed, under COINOTAG editorial oversight.

