AI-run stock funds beat the market but hold just $30 million

Key Points

1. Fewer than six ETFs let AI models handle research and stock picking entirely, Reuters reported.
2. Those funds hold about $30 million of the ETF industry's nearly $17 trillion in assets.
3. Early returns beat the S&P 500, but track records span only months and fees run high.

The latest:

A small group of exchange traded funds that hand research and stock selection entirely to artificial intelligence models are outperforming the S&P 500 in their first months of trading, Reuters reported. Fewer than half a dozen such funds exist, managing roughly $30 million against an ETF industry holding close to $17 trillion. Their managers argue performance will eventually pull in investors.

Details:

• The scale: Fewer than six ETFs currently rely solely on AI bots to control everything from research to stock selection, according to Reuters. Their combined assets of about $30 million represent a fraction of a percent of the roughly $17 trillion managed across the wider ETF industry.
• The pitch: Tal Schwartz, founder and CEO of Ai Funds, said investors have until now chosen between owning an index passively or paying a human to pick stocks, and he wants a third option in which AI does all of it. He said machines process more data faster and strip out human emotion and cognitive bias.
• The returns: Schwartz’s High Conviction US Equity AI-Managed ETF has gained 2.7% since its August 21 launch, against 1.1% for the S&P 500 over the same stretch, according to Reuters. The fund has drawn under $20 million in assets, per data from ETF.com.
• The outliers: Two funds launched in February by Israeli asset manager FinQ have far outpaced the S&P 500’s 11.0% gain since then. Its dollar-neutral US Large-Cap AI-managed ETF is up 25.35%, while the unhedged version has risen 23.98% since inception.
• The newcomers: The Pacer S&P 500 3AI Top 100 ETF and the Pacer S&P World 3AI Top 100 ETF are up 10.34% and 10.33% respectively since launching in early May, compared with a 5.8% gain for the S&P 500 across that period.
• The skeptics: Jeff Judge, a financial adviser at Chesapeake Financial Planners in Maryland, said advisers are comfortable with AI that informs a decision and uneasy with AI that makes it, adding that his clients are not asking for these products. Advisers and analysts cited short track records and above-average fees as limits on demand.
• The stress test: Todd Sohn, ETF analyst at Strategas, said proof the approach works will come down to which managers can thoughtfully handle risk and survive a falling market. None of the fully AI-run funds cited has yet traded through a sustained downturn.
• The milder version: AI-enhanced ETFs, which use the technology for quantitative analysis, stock screening or risk monitoring rather than final decisions, are the more common industry approach, insiders told Reuters.
• Pictet’s record: David Wright, head of quantitative investments at Pictet Asset Management in Geneva, said the firm has trained models using AI for seven or eight years and run live capital with them for about four. Its AI Enhanced US Equity ETF, launched in February, has gathered $163 million and gained 14.8% against 11.4% for the S&P 500.
• The wider adoption: A Mercer survey earlier this year found 91% of asset managers planned to expand their use of AI-based portfolio management tools during the year, with firms still working out which tasks can safely be handed to machines.

Between the lines:

The asset gap tells the real story: Pictet’s AI-assisted fund alone holds more than five times the combined assets of every fully autonomous AI fund cited. Investors are buying AI as an input, not as a decision-maker. Sohn’s point sharpens that — every return figure here was earned in a rising market, and none of these models has been tested by a falling one.

What’s next

Watch whether the FinQ and Pacer funds hold their leads through a market decline, and whether asset inflows follow the performance claims. Mercer’s next survey of asset manager AI adoption is the other signal.