In case you haven’t heard, artificial intelligence is kind of becoming a thing. Since the introduction of ChatGPT to the world a little over a year ago, it feels like AI is now integrating with almost every aspect of our lives.

Not long ago, it felt like going into a store kiosk, selecting a few personal preferences and having it spit out a list of items that would be perfect for you was the cutting edge of technology. Today, you can use AI to generate a semester’s worth of college essays with a few clicks, create you own AI girlfriend or see what any Disney character would look like in real life. Clearly, the opportunities are endless!

They’re also endless for investors!

If you already own a broad tech ETF, such as the Technology Select Sector SPDR ETF (XLK) or the Invesco QQQ ETF (QQQ), you’re already invested in many of the companies spending billions of dollars on artificial intelligence development. But you’re not getting pure play AI exposure. Sure, Microsoft, Alphabet and other mega-cap companies will probably end up controlling the space eventually, but it’s still a relatively small piece of their overall business model.

If you want concentrated exposure, you need an ETF that looks only at companies that derive a major portion of their revenues from AI.

Unfortunately, you’re not really going to find that in the ETF space. It’s the same situation we saw in the early stages of blockchain - a lot of companies were developing it, but few were really focused on it. That’s changed over time as now there are many more companies that could be described as pure blockchain plays, but it takes some time to get there. In the AI space right now, there are a lot of companies involved in developing their technology, but few whose sole focus is AI.

If you want to try to invest in AI, there are a couple of ETFs out there that do a reasonable, if not good, job of providing more targeted exposure. As I see it, there are four major thematic ETFs focused on AI companies at the moment. One of the names I won’t be talking about here is the ETFMG AI Powered Equity ETF (AIEQ). While it has “AI” in the name, it uses it in the fund’s research and selection process. It doesn’t necessarily invest in AI companies.

Here are the ones that do.

Roundhill Generative AI ETF (CHAT)

CHAT is the newest entrant in the arena having launched only back in May of this year. Coming in with an expense ratio of 0.75% makes it the priciest fund on this list, but it’s also the only one that’s actively-managed. In a space evolving as rapidly as AI, I think it makes sense to choose an active strategy that can be nimble & flexible and have the ability to change as conditions change. Index funds may not be able to recompose themselves but for a few times a year and that can be a real disadvantage.

CHAT’s strategy looks at two proprietary components - a transcript score and a sector score. The transcript score uses a keyword analysis from company presentations, earnings releases, etc. in order to assess commitment to the AI space. The sector score aims to quantify the expected financial performance and benefit from AI technology. Companies with greater revenue generation, investment and/or exposure to generative AI receive higher aggregate scores and greater weight in the portfolio.

The issue I have with CHAT is that the mega-caps are rising to the top of the portfolio, making it far less of a pure play AI exposure. The top holdings include NVIDIA, Alphabet, Microsoft, Adobe, AMD, Facebook, Amazon and Salesforce. These companies may ultimately become the biggest eventual winners in the AI race, but, for now, this just ends up looking like a broad tech sector ETF.

Global X Artificial Intelligence & Technology ETF (AIQ)

AIQ takes the approach of identifying potential AI-involved companies and splitting them into two buckets. The first is AI developers and servicers, which develop AI and use AI in their own products or provide artificial intelligence capabilities to their customers as a service. The second is hardware and computing companies, which include the manufacturers of things, such as semiconductors, or those that develop quantum computing technology. The top 60 companies from the first bucket and the top 25 companies from the second will form the final index (within certain market cap and liquidity constraints). Components are weighted using a modified market cap weighted methodology.

It’s that last piece that, like CHAT, tilts the portfolio heavily towards large- and mega-cap companies. Its top holdings include Intel, Facebook, Microsoft, Netflix, Alphabet, Salesforce, Amazon, IBM and Adobe, so you can see that it has pretty much the same issues that CHAT does. It has a bit more international exposure, which is a good thing since a lot of the technological innovation within AI is going to be coming from overseas, but its 85% weighting to large-caps means you’re getting very little concentrated AI investments.

WisdomTree Artificial Intelligence & Innovation ETF (WTAI)

WTAI’s index methodology breaks down companies into three major categories - software, semiconductors and other hardware. In order to qualify as being “involved in artificial intelligence activities or innovation”, natural language processing is also used looking through press releases, 10k’s, earnings report transcripts and news releases among other sources. Qualifying components then go through a modified equal-weighting process with roughly 80 names making the cut for the final portfolio.

WTAI comes with several advantages relative to the other ETFs already mentioned. Its expense ratio of just 0.45% makes it easily the cheapest fund out of this group. Its equal-weighting methodology makes the portfolio much less top-heavy than its peers (it only has about 17% of assets committed to the top 10 holdings whereas CHAT has roughly 50%). WTAI is also more diverse across market caps. Right now, it has a 50/22/22/6 split between large-, mid-, small- and micro-caps. You still get names, such as Microsoft, Alphabet and Qualcomm in the top 10, but the equal-weight strategy minimizes their weightings and allows smaller and more pure play names to get more attention.

ROBO Global Artificial Intelligence ETF (THNQ)

THNQ takes more of a numbers-based approach to filling out its portfolio. According to the prospectus, each eligible company is analyzed and then given a “THNQ score” ranging from 1 to 100 that is determined based on the levels of revenue the company receives from AI activities, the amount of investment the firm makes in artificial intelligence and the company’s technology and market leadership in the artificial intelligence universe. The fund will typically be comprised of between 50-100 securities and weightings will be made according to the THNQ score.

The revenue generation and investment qualifications make sense in sectors like this because it represents bottom line business being done in this space. I also like that the portfolio is weighted these factors as opposed to cap-weighting or even equal-weighting. The former just overweights the giants like almost every other large-cap ETF in the world, while the latter can give high exposure and low exposure companies the same relative impact within the fund. THNQ also does better at spreading its investments across all market caps.

Conclusion

The inclination may be to choose AIQ among this group of four simply because it’s easily the largest of the bunch. I think WTAI, however, is the better choice as an investment. I wish that the selection process did more than focus on natural language processing, but the end product is a solid AI-focused portfolio that invests in companies of all sizes and has a bit more pure play exposure than the others. It’s also the cheapest of the bunch, which may be less of a factor considering you’re investing in a rapidly evolving and volatile sector, but take what you can get!

Over time, I expect the ETF choices in this space to get better. It’s going to take some time for more pure play companies to develop in the same way that they did in the blockchain industry. When that happens, investors should have the choice between the likely eventual winners (the mega-caps) or the pure plays that represent more direct exposure to AI and have the potential to get acquired by competitors. Either way, it’s a great time for innovation and AI deserves at least a piece of your overall portfolio right now.