The attractiveness of bonds is rising; if the yield on 10-year U.S. Treasuries rises to 6%, 37% of respondents will consider switching from stocks to bonds.

Written by: Bu Shuqing, Wall Street Journal

The allocation enthusiasm of global fund managers for tech stocks has reached its highest level since the rebound from the COVID-19 pandemic in 2020. Meanwhile, the market consensus on AI trading is quietly diverging, with semiconductor stocks taking the place of the most crowded trading targets.

According to the latest monthly fund manager survey by Bank of America, the proportion of surveyed fund managers overweighting the technology sector jumped from 30% last month to 37%, the highest since August 2020. At the same time, about 45% of respondents believe that an AI bubble will emerge by 2027, while 42% hold the opposite view, indicating a clear market divergence.

At the time this survey result was released, tech stocks were under pressure. On Tuesday, South Korean memory chip giants Samsung Electronics and SK Hynix fell significantly, dragging Asian stock benchmarks down to their lowest levels since August, and futures contracts tracking the U.S. Nasdaq 100 index dropped 0.6% on the same day.

The proportion of overweight tech stocks reaches a six-year high

Bank of America's survey shows that global investors chasing the AI-driven stock market rally have pushed the proportion of tech stocks overweight to a six-year peak. Artificial intelligence has been the hottest trading theme in the stock market in recent years, driving the S&P 500 index to accumulate nearly 120% growth over a four-year bull market.

However, as investors begin to question whether the high capital expenditures related to AI technologies can deliver corresponding returns, some enthusiasm surrounding AI is waning. In this survey, respondents were nearly evenly split on their views regarding the AI bubble - 45% believe it will form by 2027, while 42% disagree, showing a substantial disagreement in the market's assessment of AI's long-term value.

This survey was conducted from October 2 to 8, with a total of 161 participants answering, managing assets totaling $453 billion.

Semiconductors replace AI as the market's most crowded trade

Although AI remains the central narrative driving the overall rise in tech stock allocations, investors' focus has shifted at the trading level. The Bank of America survey in October indicated that respondents ranked global semiconductor stocks as the current market's most crowded trade, with AI concept stocks no longer dominating.

This shift has already manifested in market conditions. On Tuesday in Asian markets, the significant decline of Samsung Electronics and SK Hynix directly pulled down regional stock indices, highlighting the volatility risk inherent in the concentrated holdings within the semiconductor sector. When crowded trades encounter a reversal of sentiment, the adjustments of related assets tend to be more severe.

The attractiveness of bonds is rising, and Treasury yields become the key threshold

As stock allocations heat up, negative sentiment in the bond market is also receding. The survey indicates that investors' views on bonds have dropped to the lowest level of pessimism since May 2025.

Specifically, about 37% of respondents stated that if the yield on 10-year U.S. Treasuries rises to 6%, it would prompt them to shift funds from stocks to bonds; another approximately 19% of respondents indicated that they are already making such allocation shifts. As of Tuesday morning, the yield on 10-year U.S. Treasuries was reported at 5.19%, still some distance from the trigger threshold mentioned above, but the market's sensitivity to interest rate trends is increasing.