AI Model Driving Top Ranked Funds Warn Against AI Stocks

For years, well before artificial intelligence became a dominant force on Wall Street, Daniel Mahr has been leveraging machine-learning to generate profits in the stock market. However, his AI-driven model is now consistently warning against investing heavily in the very companies leading the AI boom.

‍

Mahr’s proprietary trading system, which powers his Federated Hermes fund, has maintained a bearish stance on Nvidia Corp. and other major tech giants since 2023. That remains the case even after these stocks experienced a staggering $2 trillion decline since February, which has drawn in investors looking for bargains at lower valuations.

‍

Unlike many in the industry, Mahr has achieved market-beating returns despite being underweight in Big Tech—an impressive feat given AI-driven investment strategies often yield mixed results.

‍

“The recent drop hasn’t made them cheap; they’re still relatively expensive,” said the 43-year-old fund manager. “One of the things we generally don’t like about these companies is their volatility, and they certainly haven’t become any less volatile in recent weeks.”

‍

Mahr’s cautious stance on the so-called Magnificent Seven—Apple Inc., Microsoft Corp., Amazon.com Inc., Meta Platforms Inc., Alphabet Inc., Tesla Inc., and Nvidia—hasn’t hindered his success. His $1.6 billion Federated Hermes MDT All Cap Core Fund (QIACX) has delivered an annualized return of 26% over the past five years, outperforming the Russell 3000 Index by nearly five percentage points and surpassing 98% of its peers. Even in the challenging tariff-driven market environment, the fund’s 1.4% year-to-date decline is smaller than the broader market’s losses.

‍

Based in Boston, Mahr is at the forefront of an investment approach that increasingly relies on AI for data analysis, portfolio allocation, and risk management. Some hedge funds have even integrated AI chatbots into their research processes.

‍

Mahr’s other mutual funds also utilize machine-learning. His $2.8 billion mid-cap growth fund (FGSIX) and $2.3 billion large-cap growth fund (QILGX) have each delivered annualized returns of at least 24% over the past five years, consistently outperforming their respective benchmarks. In contrast, the Eurekahedge AI Hedge Fund Index, which tracks AI-driven strategies, has risen by only about 4% annually in the same period.

‍

His funds take a sector-neutral approach, favoring certain stocks within an industry while maintaining a corresponding underweight in others. While QIACX has a smaller overall exposure to the Magnificent Seven compared to its benchmark, it currently leans toward lesser-known tech firms such as Qualcomm Inc. and Fortinet Inc. Additionally, Mahr’s model has recently flagged the industrial sector as an attractive investment opportunity.

‍

Mahr’s journey began in 2002 when he joined MDT Advisers as a quantitative equity analyst after earning both a bachelor’s and a master’s degree in computer science from Harvard University. He moved to Federated Hermes in 2006 following its acquisition of MDT.

‍

What differentiates his investment model is its use of decision-tree analysis, a supervised learning algorithm that examines historical data to predict stock performance. Each branch in the decision tree represents a factor that helps determine whether a company is worth investing in. This method mimics human decision-making processes.

‍

Two decades ago, Mahr’s model relied on just six factors. Today, it incorporates 16, including fundamental metrics like valuation and external financing, as well as technical indicators such as stock momentum. A recent addition is the “economic moat” factor, which assesses a company’s competitive advantage.

‍

Different industries and companies are evaluated based on distinct sets of factors, allowing the model to adapt quickly to new information and generate a diversified portfolio. “The goal is consistent performance rather than swinging for home runs,” Mahr explained. He also noted that many AI-powered investment strategies tend to apply a rigid formula across all stocks, which he believes is ineffective.

‍

Decision-tree analysis remains relatively uncommon in the equity mutual fund industry, according to Jack Shannon, a principal at Morningstar Inc. who specializes in equity strategies. However, he acknowledges that AI-driven investing has limitations.

‍

“It ultimately relies on backward-looking data,” Shannon said. “The best quants recognize the risks associated with their models. Those who succeed are the ones willing to adapt and remain humble about the limits of their approach.”

‍

Amid the uncertainty of Donald Trump’s economic policies—including trade and government spending—investors are adjusting their strategies to anticipate new market winners and losers. However, Mahr avoids making bold, high-risk bets that depend on unpredictable outcomes.

‍

“We’re not trying to position the portfolio for a specific tariff outcome,” he said. “We want to hold investments that will perform well whether the situation turns out to be inconsequential or escalates into a full-blown trade war. We don’t want a binary event to determine the fate of our portfolio.”

‍