Hedge funds have been selling US technology stocks at the fastest clip in more than a decade, cutting long positions and adding short bets across nearly every subsector of the trade that defined the bull market. Goldman Sachs Prime Services data, reported by Bloomberg, shows the retreat has been sustained, broad-based, and accelerating.

When the most leveraged, most informed cohort in equity markets abandons the sector that drove the entire rally, it tells you something about crowding risk, valuation fatigue, and where capital may flow next. For metals investors, the signal is less about semiconductors than about what happens when the dominant momentum trade breaks down.

The Numbers Behind the Exit

Goldman’s prime brokerage desk, led by Vincent Lin, tracks positioning across its hedge fund client base. The data shows funds were net sellers of US tech in six of the past eight weeks, with a cumulative reduction in market value of roughly 10%. That makes it the largest retreat from the sector since Goldman began tracking the series more than a decade ago.

The S&P 500 Information Technology Index has dropped approximately 10% since early June. Technology was the worst-performing sector last week and, by Goldman’s measure, “by far the most net sold US sector.” Most subsectors saw outflows. Technology hardware, storage, and peripherals led the selling, followed by IT services. Semiconductors and software were net sold too, though by smaller amounts.

The Goldman desk put it plainly:

Related Posts
Gold Climbs Over 10% in August as Investors Embrace Complex Options for Upside Exposure

“Amid continued volatility and sharp selloff across the semis/memory/AI infrastructure complex, the persistence and magnitude of selling since early June point to significant length reduction by tech investors, and some signs of capitulation are starting to emerge.”

That word, capitulation, is worth pausing on. Prime brokerage desks do not use it casually. It implies that some funds are no longer trimming around the edges but liquidating positions under stress or conviction loss.

AI Valuations Under Pressure

The proximate cause, as described by Goldman’s strategists, is growing skepticism that AI-fueled valuations are sustainable. Investors are questioning whether the largest technology companies will sustain their current pace of AI infrastructure spending. The concern is not that AI is fake, but that the capital expenditure cycle may be peaking before the revenue cycle catches up.

Goldman strategist Ben Snider, leading a team that has tracked what they call the “AI infrastructure Momentum trade,” offered a blunt assessment: “History, positioning, and lack of a favorable catalyst point to continued near-term challenges for the AI infrastructure Momentum trade despite solid fundamentals.”

Snider isn’t saying the fundamentals are bad so much as pointing out that the trade is crowded, the momentum has broken, and there is no obvious catalyst to restart it. That distinction matters. Crowded trades with good fundamentals can still inflict serious damage when the exit gets narrow.

We have seen versions of this dynamic before in the AI bottleneck trade, where capital piled into increasingly specific components of the supply chain. The narrower the trade, the more violent the unwind.

A Pattern of Hedge Fund Pain

This is not the first time Goldman’s data has flagged hedge fund distress tied to tech crowding. Earlier in 2025, Newsmax reported that hedge fund stock pickers gave up roughly half their average yearly gains in a single day during a tech-driven selloff. Multi-strategy funds lost money in 18 out of 29 trading days over one stretch, one of the worst performance streaks Goldman had ever recorded for that fund type.

The current episode appears different in scale. A two-month, record-setting drawdown in positioning is not a one-day shock. It is a sustained re-evaluation. Funds are not just getting stopped out. They are choosing to leave.

Where is the money going? Goldman notes investors rotating into other sectors, with consumer stocks mentioned specifically. The details beyond that are thin. But the direction matters: capital is moving away from the growth-and-momentum complex and toward something less volatile, less crowded, or both.

What This Means for Gold and Hard Assets

On the surface, a hedge fund rotation out of tech and into consumer stocks has nothing to do with gold. But the second-order effects are worth thinking through.

First, when the dominant equity momentum trade breaks down, it tends to raise volatility across asset classes. Funds that were long tech on leverage need to de-risk elsewhere. That can create temporary selling pressure in commodities and metals. But it also creates the conditions that drive safe-haven demand once the dust settles.

Second, the AI capex cycle has been one of the primary arguments for continued earnings growth in US equities. If that cycle is peaking or pausing, the earnings outlook for the index narrows. A narrower earnings base makes the broader market more fragile, and fragile equity markets have historically been constructive for gold.

Third, consider what Goldman’s own institutional posture tells us. The same bank that lifted its S&P 500 target is now reporting that its own hedge fund clients are fleeing the sector that justified that target. The tension between the sell-side forecast and the buy-side behavior is instructive. When the people with the most at stake disagree with the people writing the reports, follow the money.

Crowding Risk and Capital Preservation

The broader lesson here is about crowding. The AI trade attracted an extraordinary concentration of capital into a small number of names and subsectors. That concentration generated enormous returns on the way in. It is now generating pain on the way out.

Gold, by contrast, does not have a crowding problem. Central bank buying has been persistent. Physical demand has been steady. The metal does not depend on a single earnings narrative or a single capex cycle. That structural difference is exactly what capital-preservation investors should be thinking about when they see a record-setting unwind in the most popular equity trade of the cycle.

As Ken Griffin noted at a recent Goldman symposium, skepticism about AI’s near-term investability is growing among the most sophisticated allocators. When that skepticism translates into actual selling at the pace Goldman is now reporting, it stops being a contrarian opinion and starts being a positioning fact.

The Rotation Question

One open question is whether the capital leaving tech finds a home in risk assets at all, or whether some of it migrates toward real assets, Treasuries, or cash. Goldman’s data mentions consumer stocks as a destination, but the full picture is unclear. The prime brokerage data captures only part of the flow.

If the rotation stays within equities, gold may not benefit directly. But if the tech unwind triggers broader de-risking, or if it coincides with a deterioration in the macro outlook, the case for metals strengthens. The setup is conditional, not guaranteed.

What is not conditional is the signal itself. The fastest, most sustained hedge fund exit from US tech in over a decade is a data point that deserves weight. It tells you that the most aggressive, most levered participants in the market have decided the risk-reward has shifted. They may be early. They may be wrong. But they are acting, and acting at scale.

The broader institutional reshuffling visible across markets, from capital migration away from traditional exchange structures to warnings about hidden credit risk in private markets, suggests a system where the old playbook is under revision. Tech was the consensus. Consensus trades do not unwind quietly.

Reading the Signal

For gold and silver investors, the takeaway is not that tech selling automatically means metals buying, since markets are messier than that. What matters is that the trade which absorbed the most capital, the most leverage, and the most conviction over the past two years is now in retreat. That changes the distribution of risk across the entire portfolio.

When the most crowded trade in the world starts to break, the assets that benefit are the ones that were never crowded in the first place. Gold has been many things over the past few years. Crowded has not been one of them.