THNQ: AI Not As Hot A Sector As You May Think (NYSEARCA:THNQ)

Stock markets are finally making a dip in April and it could present a good opportunity to buy the sectors and stocks previously thought to be too stretched or expensive. AI is an obvious sector to target as it is a recent “megatrend” and delivered huge gains through stocks like Nvidia (NVDA). However, as this article explores, the AI craze is not universally profitable and we have to be very careful where we invest. Robo Global® Artificial Intelligence ETF (NYSEARCA:THNQ) illustrates this point – it is an ETF focused on the hottest sector around but manages to underperform many other funds.

THNQ – Diversity at the Expense of Performance

THNQ is a passively managed ETF formed on May 8, 2020. It tracks the ROBO Global Artificial Intelligence Index, which includes companies generating significant revenue from the AI market. Stocks are selected and weighted on a rule based system outlined in the prospectus. The process is lengthy, but I have tried to break it down below –

Index components are selected from a proprietary database of Artificial Intelligence Companies that are organized into two general categories – Infrastructure or Applications & Services – and further divided into sub-sectors…

Each category’s representation in the Index varies. Each eligible company is individually 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 artificial intelligence activities, levels of investment the firm makes in artificial intelligence, and the company’s technology and market leadership in the artificial intelligence universe. Companies whose THNQ Score is greater than or equal to 50 and that meet the market capitalization and liquidity requirements described below are eligible for inclusion in the Index.

The Index is comprised of a minimum of 50 constituents and a maximum of 100 constituents.

There market capitalization and liquidity rules are designed to exclude illiquid stocks with a market cap below $200M.

The selection process results in a well diversified fund which currently has 65 holdings. These are weighted almost equally and the top 10 holdings make up only 22% of the portfolio. Nvidia (NVDA) and Microsoft (MSFT) are the two top holdings, but with 2.57% and 2.49% weightings, there is not the usual concentration seen in many other funds.

This, however, is part of the problem when it comes to performance. Mega cap stocks like NVDA and MSFT are responsible for the bulk of the rally in technology, and even the broader markets. A smaller weighting in these stocks leads to poorer performance, especially when they are supplemented with underperforming small cap stocks.

For instance, Ginkgo Bioworks Holdings (DNA) has a weighting of 1.39% and its chart looks like this –

That’s a -91% return since inception. Then there’s Ambarella (AMBA) with a 1.74% weighting and a chart like this –

We don’t know when and at what price the fund bought these stocks, but their inclusion is a worry, especially given their weightings. DNA and AMBA combined have a weighting of 3.13%, much more than NVDA at 2.59%.

As we saw in the selection process, the focus is on the proportion of revenue derived from AI, but not on any measures of quality such as revenue growth. As such, the portfolio is an interesting mix of stocks, many of which deserve a closer look, but as a whole underperform.

THNQ Performance

Here is the data on THNQ’s performance from the fund factsheet –

The returns are decent, but considering THNQ is focused on the hottest sector around, underperforming the broader market (SPY) (QQQ) is disappointing.

Part of the problem is THNQ is prone to deeper corrections. It actually outperformed into the 2021 top, but then fell much further during the 2022 bear market. It also outperformed during the rally off the October ’23 low but has corrected more again. Gains tend to get eroded on any broad market sell off.

Compared to some of its peers in the AI space, THNQ compares reasonably well. AIQ is the largest and best performing fund over the last year, but is quite closely correlated to THNQ.

What the above comparison really highlights is that the AI trend is not something boosting all related companies. It is a very wide subject with a multitude of applications. For every NVDA or Super Micro Computer (SMCI) there is a DNA or an AMBA. Indeed, the best performing AI ETF from the comparison above – AIQ – correlates almost exactly to the Technology Select Sector SPDR® Fund ETF (XLK) over the past year. This suggests there is no alpha in AI specifically and mega cap technology stocks are delivering the bulk of the returns.

For this reason, it is probably best to construct your own small portfolio focused on AI and not rely on a diversified ETF such as THNQ.

Other Considerations

THNQ is a small fund with only $168M in AUM and Average Daily Dollar Volume of $1.37M. This may be off putting for some large investors.

Another consideration is that only 79% of its portfolio is in North American stocks. 8% is in Asia, including China.

Lastly, the fund’s expense ratio is 0.68%, which is on the high side for a passively managed ETF.

Conclusions

While we regularly hear about AI and the huge gains in stocks such as NVDA, the sector as a whole is not as hot as might be expected. THNQ is diversified and weighted relatively equally but delivers no alpha compared to funds with heavier weightings in mega cap technology stocks.