If you are reading this story, my guess is you're aware of the revolution in artificial intelligence (AI), and how it's driven stocks like Nvidia (NVDA +1.80%) -- the most famous example of an AI stock -- up more than 13x in value over the last five years.
Some of you are likely thinking you have missed out on Nvidia's biggest gains by this point and are looking to buy "the next Nvidia," or something of that sort. What you may not realize is that Nvidia's initial success in selling AI semiconductor chips came about primarily because it had already become famous for selling high-performance chips for mining cryptocurrencies like Bitcoin. That fame, in turn, came about primarily because these same chips were really good at rendering graphics for video games -- which was Nvidia's original business. In other words, Nvidia's success as an AI stock was kind of a fluke.
That's a long way of saying that finding another stock that can replicate Nvidia's success won't be easy.
Image source: Getty Images.
One way to mitigate the fluke and improve your odds of finding the next Nvidia is to diversify your stock picks. Instead of trying to get lucky on the next big AI winner, you might consider buying an exchange-traded fund, or ETF, that owns shares in a good cross-section of all the best AI stocks. That way, whichever AI stock turns out to be "the next Nvidia," at least you will own a piece of it.
Think about it. Diversifying your holdings and owning a whole lot of stocks not only increases your chances of owning a piece of the winner, but it also limits your downside risk if you happen to own a few losers.
Three AI ETFs that look nothing like one another
There's just one complication with this alternative: Not all AI-centric ETFs are created equal, and some are downright bad news.
How did I learn this? Beginning with Morningstar Research, I searched for AI ETFs. I was immediately presented with five options trading in the U.S., and I began digging into the top three: Sofi Agentic AI ETF (AGIQ +1.75%), Tortoise AI Infrastructure ETF (TCAI +3.51%), and Draco Evolution AI ETF (NYSEMKT: DRAI).
When I examined the stock market returns of these three ETFs, they ranged from less than 8% over the past year for Sofi Agentic, to a more respectable but still turtle-like 12% for Tortoise AI, to the downright speedy Draco Evolution -- up 24% over the past year.
Why were there such disparate returns for three funds that are all (supposedly) invested in the same sector of the economy? As it turns out, the companies these three "artificial intelligence" ETFs have chosen to invest in are all completely different.
Sofi Agentic AI ETF
Sofi Agentic, for example, takes a "round up the usual suspects" approach to its AI investing. This ETF's top 10 holdings include a veritable who's-who list of all the top AI names from Nvidia (of course) to Tesla to Palantir Technologies. According to Morningstar data, these three red-hot AI stocks make up more than 26% of Sofi Agentic's portfolio.
And yet, somehow this ETF managed to earn its investors less than 8% -- in a stock market where a basic S&P 500 ETF like the Vanguard S&P 500 ETF would have earned you 18%!
NYSEMKT: VOO
Key Data Points
Tortoise AI Infrastructure ETF
My next step was Tortoise, and here I found an entirely different list of AI holdings. Tortoise AI focused not on the companies famed for using AI, but on those that build the infrastructure that makes the AI happen -- electrical products manufacturer Vertiv Holdings, for example, computer memory maker Seagate Technology, and computer server manufacturer Dell Technologies.
More diversified than Sofi Agentic, these three holdings combine to make up less than 15% of the stocks Tortoise invests in -- and as if to prove that diversification works, Tortoise's returns are easily outpacing those of Sofi Agentic!
Draco Evolution AI ETF
My biggest surprise from this batch of Morningstar-recommended ETFs, though, came when I examined the innards of Draco Evolution AI -- because believe it or not, according to Morningstar data, Draco Evolution hardly invests in AI at all, or at least not directly.
Instead, this ETF primarily owns shares of ... other ETFs! ProShares UltraPro QQQ, for example (a Nasdaq-focused ETF), accounts for more than 21% of Draco Evolution's holdings. Draco owns nearly as much of the ultra-bullish Direxion Daily S&P 500 Bull 3X ETF. And its third-biggest holding was an even bigger surprise: iShares iBoxx $High Yield Corp Bd ETF -- a corporate bond fund!
Combined, these three ETFs make up nearly 60% of the holdings of the Draco Evolution "AI" (quotes mine) ETF. And yet, as a result it is probably the most diversified of the three ETFs and its performance has doubled the next best performer. Diversity again shows it's a winner.
The lesson for investors
What's the lesson for investors in all of this? The first and perhaps most important lesson is to know what you're investing in, and don't judge an ETF by its name alone. Just because an ETF has "AI" in its name doesn't mean it really invests much in AI at all. And even if it does, it may not be investing in the kinds of AI stocks you think it does.
And the second lesson? Turns out, Draco Evolution taught us that one almost by accident: Sometimes, you'll make the most profit by taking on the least risk, by diversifying widely -- and not putting too much of your money in one single AI basket.
