This year has proven challenging for many active fund managers who have chosen limited exposure to artificial intelligence (AI) stocks, as those heavily invested in the sector have outperformed broad market benchmarks. According to an analysis of funds overseeing nearly $11 trillion in assets, active managers within the large-blend category began 2026 with significantly lower allocations to AI-related companies compared to the S&P 500 index. While the benchmark held approximately 36.5% in these stocks, active funds averaged 28.4%, representing an underweight of more than eight percentage points. This gap has contributed to roughly 70% of active stock mutual-fund and exchange-traded fund (ETF) managers trailing their benchmarks through October 8, according to Morningstar research.

Some portfolio managers express caution over the rapid AI sector growth, citing factors such as rising borrowing costs for neocloud companies, which provide the computing infrastructure for AI operations, and concerns that the anticipated expansion of data centers may proceed more slowly than expected. Manish Gupta, portfolio manager at First Eagle Investments, holds stakes in established technology firms including Alphabet, Meta Platforms, and Microsoft but remains wary of more speculative AI-related stocks. The First Eagle U.S. Fund and Global Fund, which Gupta manages, have returned 8.9% and 9.9% respectively so far in 2026, lagging behind their benchmarks, while the First Eagle Rising Dividend Fund has outperformed with a 15.8% gain.

Gupta defends his cautious stance, stating that his team prefers prudence over following market trends. “We are better safe than sorry,” he said, emphasizing that their concerns about the market have begun to materialize. His view contrasts with the broader market’s enthusiasm for AI, which has driven some stocks to substantial gains despite recent volatility.

Historically, active managers have struggled against benchmarks in part due to high correlations among stocks, which can limit opportunities to distinguish outperformers from laggards. However, 2026 has featured unusually high dispersion in stock returns, creating an environment theoretically favorable for stock pickers. Despite this, many managers find it difficult to identify and appropriately weight winning stocks within the increasingly crowded AI sector.

Notably, some managers who are underperforming due to limited AI exposure do not reject AI’s potential outright but argue they see better value elsewhere. Mike Nicolas, co-portfolio manager of the Oakmark Investor Fund, said his approach focuses on companies trading at attractive valuations that have been overlooked amid the AI rally. The Oakmark fund, which rose about 8% this year through mid-October, holds no shares in prominent AI or semiconductor firms such as Nvidia, AMD, or Snowflake. Instead, its largest position at the end of the third quarter was Salesforce, a software company whose shares have fallen roughly 8.5% amid investor concerns that AI could threaten its business.

As the AI trade matures, the market appears to be reaching a turning point that challenges traditional stock-picking strategies, with managers divided between embracing the sector’s rapid ascent and maintaining caution due to elevated risks and valuation uncertainties.