AI-Powered ETFs Aren’t Living Up to Investor Expectations

The sales pitch for AI-powered ETFs was that algorithms could not only pick stocks more efficiently and logically than humans, but also less expensively. So far, most haven’t lived up to investors’ expectations.

AI-driven strategies use algorithms to research and gather data on stocks and investments to select a fund’s underlying holdings. Developers of such ETFs were betting that AI could build a valuable niche by processing information faster than any human analyst, spotting trends before the rest of Wall Street. Instead, they’ve underperformed or closed altogether, said Bryan Armour, Morningstar director of ETF and passive strategies.

“They are trying to make it work,” Armour said. “There have been a few attempts, but so far, it hasn’t been very successful.”

Closing Time

Of the 12 AI-powered ETFs that Morningstar began tracking around 2020, seven have closed due to low assets under management, according to Armour. A key reason is that the mistakes often made by active managers or new investors, such as overtrading, were also being made by the AI models. (The Amplify AI-Powered Equity ETF, AIEQ, which closed last year, had a 2,700% turnover ratio in its last year of operation, according to Armour.) The ones left standing have high fees — the Qraft AI-Enhanced US Large Cap Momentum ETF (AMOM), for example, has an expense ratio of 0.75% — preventing them from taking off, Armour added. “Of the ones that are still alive… all are pretty expensive,” he said. “You would think with AI picking it, it would be a cheaper substitute for an active manager. That hasn’t been the case.”

Other closed AI-driven ETFs include:

• The BTD Capital Fund (DIP), which used AI to invest in US equities and closed last year.
• The Optimize AI Smart Sentiment Event-Driven ETF (OAIE), which used AI to monitor option activity, closed in early 2024 after less than two years on the market.

Feel the BUZZ

One exception is VanEck’s Social Sentiment ETF (BUZZ), which has climbed 45% so far this year. The fund uses AI to sift through millions of blog posts, news articles and forum comments, measuring whether feelings about a stock are positive or negative. From there, a rules-based system selects the 75 most bullish large-cap names, capping each at a 3% weight and rebalancing monthly. Armour said BUZZ arose out of the memestock craze of 2021, when investors were trying to pick the next GameStop.

“At a high level, we think of the index as a way to measure the investor sentiment of stocks as a factor,” said VanEck product manager Coulter Regal. “There’s all types of factors — momentum, value, growth. But one thing that wasn’t really out there, or at least wasn’t measurable until recently, was sentiment.”

Where To Go From Here? AI may be more useful as a tool to gauge investor sentiment rather than as a standalone strategy. And investors are unlikely to want to give up the decision-making know-how of a flesh-and-blood fund manager, Armour said. “At least for a while, people would probably prefer having a human making the end decisions,” he added.