AI-managed investing: Which smid-cap funds have outperformed the Russell 2000 YTD?

The market for AI systems used in asset management grew from $5.39 billion in 2025 to $7.1 billion in 2026 and could reach $21.82 billion by 2030, implying a compound annual growth rate of 32.4%, according to Research and Markets.

In 2025, 95% of fund managers surveyed used generative AI in their work versus 86% in 2023, according to data from the Alternative Investment Management Association. Demand is also coming from clients: 60% of institutional investors said they would be more likely to invest in a fund that allocates a meaningful portion of its budget to generative AI research and implementation.

The first ETF with a portfolio constructed by an AI model was launched in October 2017: the Amplify AI Powered Equity ETF (AIEQ). Developed by EquBot, the system runs on the IBM Watson platform and processes data every day on around 6,000 U.S. companies, including regulatory filings, news, analyst reports, and social media posts. Since its inception, the fund has generated a cumulative return of around 130%.

In the almost decade since, the selection of AI-managed funds has expanded significantly. Examples include those offered by WisdomTree and South Korea’s QRAFT. VanEck offers the index-tracking BUZZ fund, which uses online investor sentiment analysis. In February, Pictet listed its PQUS fund, in which securities are selected by a machine-learning model. The model was trained on around 400 company characteristics covering 15 years of market history and is retrained every quarter. However, the model does not make every decision independently: fund managers set the portfolio constraints themselves, for example, by factor, sector, and country. The model is responsible only for selecting individual securities within those parameters.

To assess how successfully AI invests, Oninvest selected 19 vehicles that use AI and machine learning to select stocks, forecast returns, analyze financial and textual data, and determine the weights of securities in their portfolios.

They can broadly be divided into two groups: 11 smid-cap funds and eight broad-market strategies that are not restricted by company size. The first group includes only one ETF, with the rest being mutual funds, while the second consists entirely of ETFs.

Year to date through August 4, the BlackRock Advantage SMID Cap Fund delivered the strongest performance in the sample, gaining 25.3%. It was followed by the BlackRock Advantage Small Cap Core Fund, up 24.3%. The Russell 2000 gained 22.4% over the same period. Only those two funds beat the benchmark. The remaining strategies lagged the index, including the Federated Hermes MDT Small Cap Value Fund, Voya MI Dynamic SMID Cap Fund, Federated Hermes MDT Small Cap Core ETF, and Federated Hermes MDT Mid Cap Growth Fund.

By comparison, the performance picture was reversed among AI strategies focused on large caps: five of the eight outperformed the S&P 500, which has returned 13.02% year to date. The strongest performers were the QRAFT AI-Enhanced U.S. Large Cap Momentum ETF, up 21.29%, and WisdomTree International AI Enhanced Value Fund, up 19.47% (fund returns include reinvested distributions).

Below we take a look at some of the most interesting funds: one from each major provider in the smid-cap segment and, for comparison, one ETF from the broad-market group.

Federated Hermes MDT Small Cap Core ETF (FSCC; YTD gain: 20.8%)

This is the only full-fledged ETF in our smid-cap group. Federated Hermes launched it in July 2024 alongside three other quantitative strategies, bringing the MDT approach, which the company has been developing for more than 30 years, into an ETF structure.

The model ranks stocks using several algorithms that assess companies across different parameters. Morningstar analyst Drew Carter, reviewing a mutual fund based on the same MDT approach, rated its investment process “above average.” In his view, the strategy combines valuation and quality signals with technical indicators in a single, flexible machine-learning system. However, he rated the portfolio-management team only “average.”

As of July 31, the ETF had assets of $309.4 million, while its net expense ratio of 0.36% is the lowest in our group. The portfolio is well diversified and has no pronounced sector tilt. Its largest holdings include American Healthcare REIT, which owns senior housing and medical centers; data-center fuel-cell manufacturer Bloom Energy; home-care services provider BrightSpring Health Services; and electrical switchgear manufacturer Powell Industries.

BlackRock Advantage SMID Cap Fund (MASPX; YTD gain: 25.3%)

The fund is the sample’s performance leader and the only one to significantly outperform its segment: it returned 25.3% year to date versus 22.36% for the Russell 2000. Formally, it is an institutional-class mutual fund with an expense ratio of 0.48%, and its benchmark is the Russell 2500.

In a prospectus filed with the U.S. Securities and Exchange Commission, BlackRock explicitly states that it uses machine learning and AI, including large language models and sentiment analysis, to forecast returns. BlackRock describes the strategy as a low-cost portfolio with a tech-driven approach to stock selection.

Its bets on individual stocks are modest: no holding accounts for more than 1% of assets. The largest positions include Texas bank Cullen/Frost Bankers; heating, ventilation, and cooling systems contractor Comfort Systems USA; aircraft-bearing manufacturer RBC Bearings; and specialty-alloys producer Carpenter Technology.

WisdomTree International AI Enhanced Value Fund (AIVI; YTD gain: 19.47%)

From the broad-market group, we selected the WisdomTree International ETF – the only fund in our sample that invests outside of the U.S. In 2025, it delivered the best performance of all 19 funds studied by Oninvest, returning 38.69%. It has gained 19.47% since the start of 2026. Within its group, it trailed only the QRAFT AI-Enhanced U.S. Large Cap Momentum ETF, which gained 21.29%.

In January 2022, WisdomTree restructured two of its dividend funds into strategies based on Equity Machine Intelligence, the model also used by the Voya MI Dynamic SMID Cap Fund. WisdomTree International AI uses it to find undervalued stocks in developed markets outside of the U.S. WisdomTree describes the model as an “analyst” with access to 20 years of market data and around 10,000 variables for each company.

The portfolio is dominated by large European and Asian companies, including Singaporean bank United Overseas Bank, France’s TotalEnergies, Sweden’s Swedbank, British American Tobacco, and National Australia Bank. The expense ratio is 0.58%.

WisdomTree International AI’s U.S. counterpart gained only 9.72% in 2025. This illustrates how a strategy’s performance is determined not only by the model but also by the market in which it operates.

The strong returns delivered by AI-managed smid-cap funds in 2026 primarily reflect gains in the segment itself rather than superior stock selection. Only two of the nine funds beat the Russell 2000, while five of the eight large-cap strategies outperformed the S&P 500.

However, the potential for successful stock selection is indeed greater among small caps. In December 2025, Goldman Sachs noted that “dispersion within the Russell 2000 index indicates particularly fertile ground for alpha generation,” with the index’s return dispersion more than twice that of the S&P 500. This gives fund managers more opportunities to identify stocks capable of generating excess returns relative to the index.

It is also important to understand what these products are not. Fully autonomous funds are virtually nonexistent: according to a 2026 Mercer survey, only around 5% of asset managers give models autonomous or semiautonomous authority over investment decisions.

WisdomTree’s selection process passes through five stages of human review, while BlackRock’s AI model is integrated into the portfolio-optimization process but does not replace it. In its third-quarter 2025 letter, the Voya team acknowledged that the fund’s underperformance versus the Russell 2000 was driven specifically by stock selection.

These funds should be viewed primarily as active quantitative strategies rather than as bets on AI itself. In its January outlook for 2026, Two Sigma noted that success “will depend as much on research discipline and organizational capabilities as on raw technological advancement.”

For investors, there is a simple rule to follow: look first at the market, the expense ratio, and the portfolio’s composition, and only then at the model selecting the stocks. The average annual expense ratio for U.S. equity mutual funds is 0.4%, while the average for index ETFs is 0.14%. Against that backdrop, paying 0.70-0.86% a year, as some Voya and Federated Hermes funds charge, for below-benchmark performance is hard to justify.

This text is for informational purposes only and does not constitute personalized investment advice.