Global markets: Hedge funds have reversed their buying spree in global artificial intelligence (AI) stocks, wiping out all their net purchases made since the beginning of 2025 in just five trading sessions. The sharp pullback has raised questions about whether the AI investment boom is losing momentum or whether investors are simply booking profits after a prolonged rally.
Data from Goldman Sachs’ prime brokerage desk showed that hedge funds were net sellers of global AI stocks for five consecutive sessions. The selling erased the net buying accumulated since the start of the year, signalling a significant change in investor positioning.
The retreat was part of a broader reduction in technology exposure. Hedge funds sold around $20 billion worth of technology stocks over two weeks, according to data reported by Reuters on October 26. The scale of the selling was close to the $23 billion recorded during a major unwinding in July.
The pressure was also reflected in the performance of stocks favoured by hedge funds. Goldman Sachs’ basket of hedge fund-favoured stocks fell 4.2% on October 22, marking its steepest single-day decline since 2011.
The developments have intensified scrutiny of the AI trade, which has driven substantial gains in technology stocks over the past two years. Investors are increasingly questioning whether the enormous spending on AI infrastructure will generate returns sufficient to justify the valuations of companies involved in the sector.
Hedge funds erase their AI stock purchases
The speed of the reversal has been one of the most notable features of the latest market movement.
Goldman Sachs’ prime brokerage data showed that hedge funds sold global AI stocks for five consecutive trading sessions. This wiped out all the net buying they had accumulated since the beginning of 2025.
However, the selling pattern suggests that funds were primarily reducing existing investments rather than aggressively betting against the sector.
The pace of selling from existing holdings was 3.5 times that of fresh short positions. This indicates that hedge funds were more focused on unwinding positions and lowering their exposure than opening new bets against AI-linked companies.
The distinction is important. Selling existing shares can reflect profit-taking, portfolio adjustments or a reduction in risk. It does not necessarily mean investors expect AI companies to experience a prolonged decline.
Nevertheless, the scale of the reduction suggests that funds are reassessing how much exposure they want to maintain in a market that has attracted substantial institutional investment.
Technology stocks face a wider sell-off
The retreat extended beyond companies directly associated with AI.
Goldman Sachs’ prime brokerage desk recorded approximately $20 billion in technology stock sales over two weeks, Reuters reported on October 26. The figure was close to the $23 billion recorded during a major technology stock unwinding in July.
Another indication of the changing investor appetite came from Morgan Stanley data cited by Steno Research. Hedge funds reduced their overall exposure by around 7.8%, representing the largest cut in the data series dating back to 2016.
The pressure was particularly visible among companies supplying the hardware and components needed for the AI infrastructure expansion.
Samsung Electronics, SK Hynix, Micron and SanDisk collectively lost more than $700 billion in market value in less than two weeks, according to the figures cited in the report.
Chip stocks declined 15.6% in October, while the S&P 500 fell by less than 1% over the same period. The difference highlights how sharply the technology and semiconductor segments were affected compared with the broader US equity market.
The sell-off also demonstrates that companies benefiting from the expansion of AI infrastructure are not insulated from changes in investor sentiment. Even businesses positioned to supply essential components can face significant share price pressure when market participants reassess valuations and future earnings.
Why investors are questioning AI spending
A central concern behind the latest pullback is whether technology companies will earn enough from AI to justify the enormous amounts they are committing to its development.
The AI investment cycle has been supported by expectations of sustained demand for advanced semiconductors, data centres, networking equipment and computing infrastructure. Major technology companies have announced substantial spending plans to expand their capacity and develop AI-related products and services.
However, infrastructure investment requires considerable capital. Investors increasingly want evidence that this expenditure will translate into sustainable revenue growth, stronger profit margins and healthy cash flows.
The concern is not necessarily that demand for AI technology will disappear. Instead, the question is whether the financial returns will match the scale of investment.
Reuters reported on October 26, citing a Goldman Sachs note, that concerns about returns from AI spending were central to the latest market volatility.
This has shifted attention from how much companies are spending to how effectively they can monetise their investments. Building data centres and purchasing advanced chips can increase capacity, but the resulting services must generate sufficient revenue to justify the cost.
If companies struggle to convert rising AI demand into profitable business, investors may become less willing to support the premium valuations attached to their shares.
Meta’s share price highlights investor concerns
Meta’s market performance provided another example of how investors are reacting to the cost of AI expansion.
The company’s shares fell around 10% in the week after it reported strong earnings but indicated that it would need to increase capital spending to support its AI investments.
The decline came despite the positive earnings performance. It suggested that investors were looking beyond the company’s immediate financial results and examining the funding requirements of its future technology plans.
Meta’s experience reflects a broader challenge facing major technology companies. Strong demand for AI services and infrastructure does not automatically guarantee that large-scale investment will produce attractive returns.
The amount of capital required to build and maintain AI infrastructure is also becoming an important consideration. If companies need additional funding to sustain their investment plans, investors may question the impact on future profitability and cash generation.
The market reaction indicates that investors are becoming more selective about AI-related spending. Companies may increasingly need to demonstrate clear commercial benefits rather than relying solely on expectations of future growth.
Are high valuations adding to the pressure?
Valuations are another factor contributing to the volatility.
AI-linked stocks have attracted significant investor interest over the past two years. Expectations of rapid earnings growth have supported premium valuations for companies associated with AI software, semiconductors and infrastructure.
Such valuations can become vulnerable when investors begin questioning the assumptions behind future growth.
When a large number of institutional investors hold similar positions, even a moderate change in sentiment can lead to substantial selling. If funds attempt to reduce exposure at the same time, the resulting pressure can accelerate a market correction.
The latest Goldman Sachs data points to this possibility. Since selling from existing holdings was substantially higher than fresh short-selling activity, the immediate trend appears to be one of position reduction rather than a widespread move to bet against AI companies.
For investors, this raises the risk of crowded trades. A stock can remain fundamentally attractive while experiencing sharp declines if too many market participants attempt to exit their positions simultaneously.
Hedge funds have not abandoned technology stocks
Despite the scale of the recent selling, the available data does not suggest that hedge funds have completely withdrawn from technology.
Technology stocks still accounted for 23.5% of Goldman Sachs’ global prime brokerage book, according to figures cited by The Economic Times. This allocation remained in the 90th percentile of the past five years.
The figure suggests that hedge funds continue to maintain substantial exposure to the technology sector, even after the latest reduction.
This distinction matters when interpreting the sell-off. A sharp decline in buying or a wave of selling does not automatically signal that investors have lost confidence in the long-term potential of AI.
Hedge funds may be taking profits after a strong rally, reducing concentrated positions or adjusting portfolios in response to changing market conditions. They may also be reassessing the balance between expected returns and the risks associated with high valuations.
The continued presence of technology stocks as a significant part of Goldman Sachs’ prime brokerage book indicates that the sector remains important to institutional portfolios.
What could determine the next phase of the AI trade?
The next phase will depend heavily on whether companies can demonstrate that their AI investments are producing sustainable financial returns.
Investors will be watching revenue growth from AI products, demand for computing infrastructure, operating margins and cash flows. They will also assess whether capital expenditure plans are delivering measurable business benefits.
If AI-related revenue grows sufficiently to support the scale of investment, confidence in the sector could stabilise. Companies able to demonstrate strong returns may continue to attract institutional interest.
However, if spending rises faster than revenue and profits, investors could reassess the valuations of companies involved in the AI ecosystem. Such a shift could put further pressure on semiconductor manufacturers, data centre suppliers and major technology firms.
The latest figures show that hedge funds have made a substantial change to their AI stock positions, but they do not establish that the broader AI investment opportunity has ended.
For now, the market appears to be moving from enthusiasm over the scale of AI investment towards a more demanding assessment of its commercial value. The central question is no longer simply how much companies are spending on artificial intelligence, but whether that spending can deliver the returns investors expect.