AI ETFs: should you buy one?

Specialist exchange-traded funds (ETFs) are a simple way to access interesting themes through a single investment. With artificial intelligence (AI) being one of the hottest investment trends at the moment, is now the right time to buy an AI ETF?

AI has taken the world by storm, reaching a tipping point when ChatGPT launched in November 2022. AI quickly went from being something we’d heard about in science fiction to something many of us were using in our everyday lives. The top picks for DIY investors have been dominated by big tech stocks like the ‘Magnificent Seven’, with AI powering much of their performance.

But with global capital expenditure soaring – and recent estimates suggesting that will accelerate – can the exceptional rate of returns continue?

Try 6 free issues of MoneyWeek today

Get unparalleled financial insight, analysis and expert opinion you can profit from.

Start your trial

Sign up for MoneyWeek’s free twice-daily newsletter.

Join more than 165,000 subscribers and keep yourself informed with latest financial news, insights and expert analysis to help you understand what really matters when it comes to your finances.

Join more than 165,000 subscribers and keep yourself informed with latest financial news, insights and expert analysis to help you understand what really matters when it comes to your finances.

Sign up

What is an AI ETF?

As the AI landscape is moving so rapidly, going all-in on a single company may feel too risky. Instead, an AI ETF offers the ease of a single investment but with broader reach.

ETFs are funds that are traded on a stock exchange, pooling money to buy a basket of stocks. AI ETFs hold stocks that are related to AI.

An AI ETF may invest across various related sectors, such as machine learning, cloud computing, semiconductors, robotics, automation or data infrastructure. Some will take a wider view, perhaps investing across US big tech while others may be more niche.

ETFs can be active or passively managed. If your AI ETF is passive, it could track one of several indices, such as the NASDAQ CTA Artificial Intelligence, Indxx AI & Big Data, ROBO Global Robotics and Automation or STOXX AI Global Artificial Intelligence indices.

What types of AI ETFs are available?

Currently, UK investors can choose from 17 AI ETFs, which track 16 indices, according to JustETF.

The cheapest is the Franklin AI, Metaverse and Blockchain UCITS ETF (LON:METU), which has a total expense ratio of 0.3%. At the other end of the cost spectrum is the ARK Artificial Intelligence & Robotics UCITS ETF (LON:ARCI), which costs 0.75% a year.

The largest by far, with over £6.8 billion in assets, is Xtrackers Artificial Intelligence & Big Data UCITS ETF (LON:XAIX).

With thematic investing, it’s important to look at the purity of where the underlying companies are getting their revenues, says Lynn Hutchinson, head of ETF and index solutions at wealth management business Raymond James.

She says many of the large funds are tilted towards big tech companies, which may not always be present in the theme you’re trying to target.

At the broader end, she highlights L&G Artificial Intelligence (LON:AIAG) and Global X Artificial Intelligence & Technology ETF (LON:AIQU). But she also points out that some more focused ETFs can offer AI exposure via the theme’s subsectors: examples include the VanEck Semiconductor (LON:SMGB) and HANetf’s Defiance AI & Power Infrastructure ETFs (LON:AIPG).

What should I consider when comparing AI ETFs?

Thematic investing can offer easy access to big, long-term global growth trends. But ETF provider WisdomTree stresses it’s important to understand the nuances.

Pierre Debru, head of research, Europe at WisdomTree explains that two ETFs may invest in the same theme but hold very different baskets of underlying stocks, and can therefore deliver very different returns.

The firm recently carried out research on dispersion of returns in thematic ETFs. It found that even with a well-defined theme such as semiconductors, over 12 months there was around a 20 percentage point difference between the best and worst performers.

Debru explains: “[The divergence is because] there is no single, recognised benchmark for a theme like AI or quantum computing. When a manager invests in a US small cap, it is clear what the investment universe is and what the benchmark is. As a result, managers make very different stock selection and weighting choices and outcomes vary accordingly.”

With many AI companies being growth stocks and possibly early in their lifestage, it’s important to be aware of volatility.

Concentration in some products, because of the dominance of some of the larger position sizes, is another risk investors should take into account, says Raymond James’s Hutchinson.

“Make sure you’re not overlapping with other investments you may hold. If you’ve already got Nvidia, Microsoft and Broadcom in other products, you’ll potentially be overlapping.”

Capital spend slowing down is another risk to the AI theme, though it may be far off. “I’m not sure when we might see that, but we may get some slowdown,” adds Hutchinson.

Another key consideration is how the fund is managed. The active versus passive debate is long-running and comes down to personal preference, what’s happening in markets at the time and what you’re looking for your ETF to do.

For active ETFs, it's a good idea to research the fund manager and their strategy. Active funds tend to charge higher fees than passive equivalents, given their potential to beat the market through the judicious trades of their portfolio manager. This isn’t guaranteed to happen but it can potentially lead to outsized returns.

“Active managers aren't bound by an index's rules,” says Rahul Bhushan, managing director at ARK Invest Europe. He says this allows them to focus on a high-conviction portfolio, selecting companies they believe have the strongest potential for growth within the theme.

Most of the ETFs around the AI theme are passive but they still have nuances that differentiate them. The selection and rebalancing strategy of their benchmark index can make a big difference to their returns.

The index tracking error – the difference in returns between the ETF and the underlying index – is something to keep an eye on. Once you allow for fees, if the difference between these two continues, that could be a red flag might be worth looking at in more detail, adds Hutchinson.

Total return performance over 12 months to 2 July, shown in GBP. Source: London Stock Exchange.

Let’s have a look at the different features of each ETF and what they potentially offer an investor’s portfolio.

1. INTL: targeted AI exposure

The WisdomTree Artificial Intelligence UCITS ETF (LON:INTL) tracks the NASDAQ CTA Artificial Intelligence Index.

This index assesses a broad universe of stocks' relevance to three categories (AI enablers, engagers and enhancers), and selects 15 companies (reviewed twice annually) from each based on the degree of their relevance. This gives its holdings a high level of relevance to the AI theme.

As of 2 July, INTL’s top three holdings are Advanced Micro Devices, Marvell Technology and Micron Technology – each of these is a semiconductor company, though they focus on different parts of the chip market.

2. AIAI: quarterly rebalancing

The $2.1bn L&G Artificial Intelligence UCITS ETF (LON:AIAI) tracks the ROBO Global Artificial Intelligence Index. Like its Nasdaq counterpart, this index weights constituents based on their relevance to the AI theme.

However, it rebalances quarterly, meaning that there is twice as much buying and selling of the constituent stocks going on. Lots of investors like frequent rebalancing because it effectively automatically sells high, and buys low: assuming no change in relative AI relevance, stocks whose share prices have outperformed over the quarter will be sold, and those that have underperformed will be bought.

Though not for everyone (many investors prefer strategies that let outperformers run their course longer before taking profits), this makes AIAI an interesting option for investors that see the value in a quarterly rebalancing strategy.

3. AIQU: broad diversification

Global X Artificial Intelligence UCITS ETF (LON:AIQU) tracks the Indxx Artificial Intelligence Index, uses a modified market cap-weighted approach that caps individual holdings at 3% and rebalances twice annually.

While this low cap could weigh on performance (by limiting the gains from individual outperformers), it does potentially make it appealing for investors worried about megacap concentration.

It has a very low concentration of Mag 7 stocks in its top holdings. At the time of writing, only Apple and Alphabet make its top 10, with neither sitting in its top five. SK Hynix, Micron and Advanced Micro Devices are its top three positions.

This therefore makes a solid choice for AI investors who want to diversify away from the main US megacaps.

4. ARKI: actively managed AI

The ARK Artificial Intelligence & Robotics UCITS ETF takes an actively managed approach, which over the past 12 months has seen it trail some of its passive peers we’ve outlined above.

This actively managed fund focuses on AI, automation and autonomous technology, with stock selection driven by ARK’s CIO and founder, Cathie Wood. Investors can expect early-stage and little-known companies in this ETF, such as warehouse robotics company Symbotic, which has a market cap of just $28 billion, alongside well-documented names like Tesla, which is currently the fund’s biggest holding, at around 9%.