As of April 15, 2026, there were 47 exchange-traded funds with names that featured the terms “AI” or “artificial intelligence,” or to which Morningstar had assigned its “Artificial Intelligence” thematic tag.
Some of those ETFs are plays on the large-language models themselves, others on the infrastructure, hardware, and data that make the chatbots and agents go. Still other ETFs leverage AI to pick securities. Here’s the list of the 47 AI-related ETFs I assembled:
The average AI-themed ETF is around three years old, the oldest being the Global X Robotics & Artificial Intelligence ETF (BOTZ; 9/12/2016 inception date), the youngest the AI Data Center REIT & Infrastructure ETF (RCKZ; 4/14/2026). Most AI-themed ETFs have come along since 2022.
Even though they’re fairly new, these ETFs have been popular with investors. Collectively, they recently held around $25 billion in net assets, thanks to strong demand: As a group, the ETFs hauled in more than $10.0 billion in net flows over the past 12 months, $7.7 billion over the second half of last year alone. All told, they’ve gathered a cumulative $19 billion of net inflows since the first debuted in 2016.
Performance: It’s Complicated
How have these ETFs performed? It’s been mixed. On the plus side, the average AI-themed ETF has topped the S&P since 2016. But that hasn’t been good enough to beat the Nasdaq 100 Index over that span.
That average return also masks significant variation in the performance of the underlying AI-themed ETFs. To illustrate, here is the difference in rolling one-year returns between AI ETFs at the 25th and 75th percentiles.
There was a 17-percentage-point difference between the haves and have-nots in the average one-year period, a gap big enough to drive a truck through. That dispersion is about 3 times wider than the return difference you’d find among large-blend funds and even exceeds that of funds assigned to Morningstar’s “technology” fund peer group.
Parsing ‘AI-Themed’
What’s given rise to these wide performance disparities? Though most of these ETFs fly under the same AI banner or invest in resources and capabilities that are central or at least adjacent to the technology, they aren’t monolithic.
For instance, take the largest AI-themed ETF by net assets, Global X Artificial Intelligence & Technology (AIQ). That ETF spreads its assets across more than 80 stocks and thus doesn’t concentrate too much in its top names. Contrast that with TrueShares Tech, AI, and Deep Learning ETF, which holds only 20 or so stocks and, as such, loads up on its top picks, which have routinely soaked up two-thirds or more of the portfolio.
Those differences appear to explain the large gap between the two ETFs’ returns since August 2022, which has seen the Global X ETF gain around 20% per year versus only 9% annually for the TrueShares ETF.
The Global X ETF hit pay dirt with names like Netflix, SK Hynix, and Samsung, none of which the TrueShares ETF held during this period, and avoided trouble in reeling stocks like Figma, Relay Therapeutics, and Prime Medicine, all of which figured into the TrueShares portfolio at various points along the way.
The point isn’t to dwell on what might have gone right or wrong for a pair of ETFs that share a similar name but rather to note these differences matter. And judging from the correlation of each AI-themed ETF’s returns to the return of the average fund, it appears those differences are fairly widespread.
TakeawaysYou Don’t Need This
I probably should have led with this, but I don’t see a compelling reason for anyone to go out of their way to invest in a thematic strategy like AI ETFs. And the reason is simple: If you’re investing in a diversified portfolio like an index fund, you’re probably already participating in that theme, and all others for that matter. Anything beyond that is likely to be redundant and pile on extra risk you don’t need.
Careful What You Wish For
To those who would say, “Yeah, but I want to dial up my exposure to the theme,” I would ask: What makes you think it’s not already priced in? A rule of thumb I use is to assume that by the time an emerging technology, concept, or theme has reached my attention, it’s thoroughly mainstream. Kind of the investing equivalent of the Groucho Marx quip, “I refuse to join any club that would have me as a member.”
I’ve already written about the yawning gap between the return of the average dollar invested in thematic ETFs and those ETFs’ aggregate returns. What caused such a wide gap to open between the two? Inopportunely timed purchases and sales by investors, who too often bought high and sold low.
End of Story?
Themes are all about a story that’s still in the process of unfolding. In the case of AI, it’s a mass-consumer adoption and build-out story in which the technology is said to herald a new era of unheard-of innovation and productivity. That’s stirring, for sure. But have you worked out how, when, or how that story might end? You can’t reasonably be expected to have figured that out—the heads of the frontier AI labs aren’t even sure how it’ll end. Which makes it all the more important to have evaluated what fundamental change in the theme-based strategy you’re considering would trigger you to cut or sell the stake outright.
Know What You’re Buying
That’s a good segue to this—you’re not buying a narrative, you’re buying a stake in businesses that will generate cash flows in the future. The timing and magnitude of those cash flows, we don’t know. The price other investors will be willing to pay for that stream of cash flows is even less certain. Also, that basket of stocks could change in the future depending on how the theme-based strategy concerned works. The price that the portfolio fetches today reflects the market’s collective assessment of those stocks’ worth. What is your view? Why do you expect those stocks to appreciate? What’s the fundamental explanation for that appreciation? What does the market have wrong?
Get Your Hands Dirty
Even if you get all the way to the end—you’re undeterred by the portfolio overlap, you think you got in early, you’ve worked out a scenario in which you’ll sell, you’ve formed a sturdy fundamental thesis for why it’s a worthwhile investment, you just think I need to lighten up—you’ve still got to drill down to figure out what makes a strategy tick and how that differs from competing options. As this article underscores, those details—what qualifies as “AI,” how many investments the ETF holds, how it weights those stocks, how long it holds them, when and why it might sell them, what it might replace them with—can matter a great deal. But given the sheer multitude of these vehicles and the various ways they operate, you might need a large language model to keep it all straight.
Switched On
Here are other things I’m reading, writing, or listening to:
Remembering Mark MobiusMorningstar’s annual “diversification landscape” paper is out“The Accidental Star Manager”The Both (Self-titled)I guess I’m what you call an “Ella fella”Don’t Be a Stranger
I love hearing from you. Have some feedback? An angle for an article? Email me at jeffrey.ptak@morningstar.com. If you’re so inclined, you can also follow me on Twitter/X at @syouth1, and I do some odds-and-ends writing on a Substack called Basis Pointing.