Wall Street Investment Trends Shift Dramatically! Tech Bulls Stage a "Mass Exodus" Amid Surging Long-term Bond Yields, Banks Compete with Gold and Copper Miners for Market Leadership

According to The Intellectual Finance APP, the global technology stocks centered on the theme of AI computing power trading are experiencing a typical "discount rate surge shock + contraction in financial market risk appetite," driven by the rapid increase in debt issuance by AI-related technology companies, mounting debates about an "AI credit bubble burst," and the rise in U.S. Treasury yields for maturities of 10 years or more.

Goldman Sachs statistics show that the Information Technology sector (mainly comprising technology stocks) was the largest net sold sector in the U.S. stock market last week, marking the biggest reduction in total exposure in more than two years. More importantly, almost all of the sell-off was due to institutional long positions being quickly trimmed rather than an acceleration of short positions, indicating that Wall Street institutions are proactively realizing profits, reducing the AI concentration in their portfolios, and reallocating toward classic defensive sectors to hedge against surging yields.

This shift in positioning is closer to a phase-out of leverage and overcrowded trades in AI-related stocks, rather than a complete rejection of the long-term demand logic behind AI computing power and software applications. However, if long-term U.S. Treasury yields and credit spreads for technology firms continue to accelerate upward, AI concept stocks that rely on external financing, future large cash flows, long capital return cycles, and high sensitivity to denominator risk-free rates will remain under pressure. On the other hand, leaders in the AI computing power supply chain with strong free cash flow, clear order visibility, and tough-to-replace technology barriers—as well as key cloud AI inference power suppliers—could be the first to receive capital reallocation after institutional overcrowding and de-leveraging clear out, shifting the market from a universal AI computing boom to fundamental-based stock selection.

Not a Short Attack, but a Long Retreat! AI Conviction Faces Bond Market Yield Test, Tech Stock Is the Largest Net Sold Sector

Goldman Sachs data show that last week the Information Technology sector was the investment theme with the largest net selling in the U.S. equity market, far outstripping other segments, and this large-scale selloff was almost entirely driven by the long side.

Goldman Sachs senior analyst Robby Stankard noted that the sector's total allocation exposure saw the largest percentage decline in more than two years, highlighting that investors have sharply reduced their risk exposure to this sector.

The firm's compiled chart indicates that the Information Technology sector's total exposure is about 20% of the total U.S. market exposure, significantly lower than the year’s earlier peak of nearly 24%. The net exposure metric is about 19%; during the recent selloff, this figure dropped from a peak of about 26% in 2026 to about 16%.

Previously, both metrics had been rising sharply in the first half of this year. Total exposure rose from about 18% at the start of 2026 to nearly 24%, while net exposure spiked from about 17% to 26%, before seeing a major reversal following the Korean stock market pullback in June and major global semiconductor selloff in July.

The Goldman Sachs analyst team led by Stankard stated that the latest round of selloffs was almost entirely powered by long liquidations, and not by an increase in short positions, showing that investors are reducing bullish exposure in popular AI computing-related tech sectors.

Goldman Sachs’ team noted in their research report that the market’s investment orientation has changed, but more accurately, it’s not a “complete abandonment of the AI computing theme,” but a shift from “single high-beta AI hardware trades” to “de-leveraging/de-crowding, balanced positioning toward high cash flow and long-term low-concentration stocks, and cross-asset diffusion.”

Goldman Sachs data show that, after global hedge funds continuously bought the dip since late July, they suddenly flipped to net selling U.S. stocks, with the Net Leverage metric dropping to a one-year low of 48.3%. Coupled with Information Technology becoming the largest net sold sector—and the selling almost all from long reductions—the total tech sector exposure has fallen from this year’s nearly 24% peak to about 20%, and net exposure once from about 26% to 16%. This means institutions are compressing crowded positions and portfolio volatility, rather than massively adding new shorts to bet on an AI fundamental collapse.

Surge in Long-Dated Treasury Yields and Bond Issuance Reshaping Wall Street Strategy and Investment Direction

Goldman analysts state that the key factors triggering this style rotation are the dual and persistent rises in “long-term discount rates + AI financing costs”: 30-year U.S. Treasury yields touched around 5.33%, and the 10-year neared 4.7%, significantly increasing the discount rate on future cash flows for technology stocks, data center financing costs, and the relative attraction of Treasuries compared to equities.

Since 2026, AI-related bond issuance has totaled about $220 billion, far surpassing the $12.5 billion in the previous year; the credit spread for tech bonds has widened to about 89 basis points, roughly 9 basis points above the broad investment-grade market, indicating that investors are now demanding a higher “AI financing risk premium.” However, AI bond issuance is mainly an amplifier of long-term term premiums, not the sole driver of yield increases—factors like the U.S. fiscal deficit, inflation resiliency, and Fed credibility remain deeper underlying forces.

The so-called “AI momentum decay” referenced in Goldman’s report mainly means a rotation out of momentum factors at the price level and does not equate to a peak in AI orders, Token demand, or computing capex: Software has become the heaviest weight in three-month momentum longs, while semiconductors and AI composite sectors have slid to the short side, reflecting a shift away from “which AI value chain players are garnering the largest AI capex” toward “who can continuously convert computing power and Tokens into robust revenue, EPS, and free cash flow.” Goldman still expects corporate AI computing infrastructure deployment to accelerate, with increasingly complex and large inference workloads tightening AI cloud resource constraints for hyperscalers.

Therefore, the AI computing theme trade has not completely ended, but said goodbye to indiscriminate multiple expansion: Capital-rich platform cloud leaders, AI application software, and storage-chip giants and data center infrastructure core suppliers where share prices significantly lag EPS trajectories still have opportunities; while new cloud firms highly dependent on debt, project financing, and constant refinancing face even stricter balance sheet scrutiny.

At the cross-asset level, Goldman’s analyst team says capital is apparently rapidly building out a classic “barbell” portfolio of “high-quality AI cash flow + banks + hard assets”: European and Japanese banks are benefiting from steeper yield curves, “higher for longer” rate expectations, and improvements in net interest income (NII); spot gold commodities and gold mining companies are used as hedges against fiscal dilution and a weakening U.S. dollar; copper mining stocks, meanwhile, benefit from constraints in physical supply for grid, data center, and energy infrastructure at the core of geopolitical rivalries between global powers East and West.

Goldman adds that investors should not treat all real assets as winners—utilities and real estate are still being sold off due to their bond-like attributes. If long-end term premiums, credit spreads, and AI-related financing supply continue to rise significantly, European and Japanese bank stocks, gold, copper miners, and AI application platforms with current cash flows and high-quality fundamentals, plus long-term low-concentration holdings, will be more likely to take over market leadership.