Key Takeaways
- Major institutional investors have been offloading semiconductor positions following a substantial rally, new Goldman Sachs prime brokerage data reveals.
- The AI semiconductor basket tracked by Goldman has delivered over 50% outperformance versus the S&P 500 in 2025.
- Chip stocks emerged as the most net-sold US subsector during the past month.
- Despite semiconductor selling, hedge fund exposure to AI-related equities remains at historically elevated levels.
- Institutional short interest in broad market indices and ETFs has reached its highest point in ten years.
Major institutional investors have been strategically trimming their chip stock holdings following an impressive rally, fresh data from Goldman Sachs’ prime brokerage division indicates. Industry analysts are characterizing these moves as tactical profit-taking rather than a fundamental retreat from the artificial intelligence investment thesis.
The AI semiconductor basket monitored by Goldman has delivered outperformance exceeding 50% compared to the S&P 500 throughout 2025. The broader S&P 500 benchmark itself climbed more than 18% from late March through a recent three-session decline. Such accelerated appreciation typically prompts institutional investors to secure gains.
During the most recent four-week period, semiconductors and related equipment manufacturers emerged as the most aggressively net-sold US subsector across Goldman’s prime brokerage client base. The selling primarily consisted of reducing existing long holdings rather than initiating fresh short positions against chip companies.
The semiconductor industry has now shifted into net-sold territory on a year-to-date basis. This marks a reversal from the first quarter when institutional capital was flowing into the sector.
South Korea’s Kospi benchmark, frequently monitored as a barometer for global AI infrastructure demand, momentarily surpassed the 8,000 threshold for the first time in mid-May. The index pushed its year-to-date advance past 80% before experiencing a notable correction.
AI Investment Thesis Remains Strong
Notwithstanding the semiconductor selling activity, Goldman’s tracking data demonstrates that aggregate hedge fund positioning in US artificial intelligence equities within its technology, media, and telecommunications universe continues hovering near all-time peaks.
According to Vincent Lin and Goldman’s prime services team, institutional clients are “consolidating and managing their semiconductor exposure within their overall portfolios, rather than signaling a paradigm shift away from the AI theme.”
The interpretation suggests these transactions reflect profit realization on successful positions rather than abandoning conviction in AI’s long-term investment potential.
Index Hedging Reaches Decade High
Concurrently with chip stock liquidation, hedge funds have been establishing short positions in broad equity benchmarks and exchange-traded products. These hedging positions have now climbed to their most elevated level in a full decade.
Institutional managers typically deploy this hedging approach to mitigate portfolio exposure to general market volatility while maintaining conviction positions in individual securities.
Aggregate gross leverage across the hedge fund universe has climbed to a fresh five-year maximum this month. Meanwhile, net leverage metrics have remained comparatively steady.
Goldman’s analysis emphasized that this positioning profile contrasts sharply with the bullish sentiment currently prevalent among retail market participants. The investment bank characterized the divergence as incompatible with widespread market euphoria.
The S&P 500 was changing hands near 7,410 at the time of publication, registering a decline of approximately 0.31% intraday.
Collectively, the positioning data indicates institutional investors are implementing careful risk management protocols while preserving their fundamental AI investment allocations.
The post Hedge Funds Dump Semiconductor Stocks While Maintaining AI Positions, Goldman Sachs Reports appeared first on Blockonomi.






Be the first to comment