Artificial intelligence (AI), the concept that a machine-created brain could outperform the human brain, has fascinated and terrified us humans since the days of Jules Verne.  Real life examples such as Deep Blue defeating Chess Grandmaster Gary Kasparov and IBM Watson defeating Ken Jennings at Jeopardy intrigues us even more.  Machine learning techniques, expert algorithms and artificial intelligence implementations continue to evolve and be useful.

The latest sensation is ChatGPT.  Trained by AI and machine learning, the system is designed to provide information and answer questions through a conversational interface. The AI is trained on a huge sample of text taken from the internet. The new AI was created with a focus on ease of use. “The dialogue format makes it possible for ChatGPT to answer follow-up questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests,” its creator, OpenAI,  said at its November launch.

Early users have described the technology as an alternative to Google because it is capable of providing descriptions, answers and solutions to complex questions including ways to write code and solve layout problems and optimization queries. Real-world applications could include generating content for websites, answering customer inquiries, providing recommendations, or creating automated chatbots.

This excitement has revitalized interest in ETFs utilizing AI. There are basically two types of US ETFs involved in AI technologies. The first grouping is thematic: These are ETFs that invest in companies that produce artificial intelligence products and robotics.   

The second actually use artificial intelligence for stock selection. 

ETFs in the thematic category include:

BOTZ, Global X Robotics & Artificial Intelligence Thematic ETF (largest in AUM);

KOMP, SPDR S&P Kensho New Economies Composite ETF (fastest growing);

ARKQ, ARK Autonomous Technology & Robotics ETF (largest active);

IRBO, iShares Robotics and Artificial Intelligence Multisector ETF (lowest fee);

ROBO, Robo Global Robotics & Automation Index;(longest track record);

ROBT, First Trust Nasdaq Artificial Intelligence and Robotics ETF(equally weighted);

LOUP, Innovator Loup Frontier Tech ETF (smallest Avg. Market Cap.);

THNQ, Robo Global Artificial Intelligence ETF (purest AI-only play);

LRNZ, TrueShares Technology, AI & Deep Learning ETF (active)

WTAI. WisdomTree Artificial Intelligence ETF (equally weighted, active-selection).

First Trust, Global X and Wisdom Tree have been among the early leaders in producing thematic ETFs since the beginning of their existence.  iShares and SPDR are more apt to take successful themes in those ETFs and offer a lower-cast and generally broader product on that theme.  ARK Investment management is one of the best-known providers of active ETFs in the emerging technology space. TrueShares specializes in actively managed funds selecting managers with expertise in a specific area.  Finally, ROBO Global is a much smaller company dedicated to innovative technologies and also the first ETF Advisor in this space with ROBO for which they combined with longtime “white label” ETF sponsor, Exchange Traded Concepts (ETC). Eight of the 10 have $20 million or more under management.  LRNZ has about $15 million and WTAI is under $5 Million thus far.

As a lifetime quant who has designed and managed expert-system, algorithmic and risk-optimized portfolios, I am personally even more fascinated by portfolios and ETFs actually managed using artificial intelligence.  The concept isn’t new. For example, Fidelity Investments ran one in the 1990s but obviously the technology was much more primitive and had much less data available than today as the Internet was in its infancy.  There are also AI platform providers such as Boost.ai and ALLINDEX that facilitate AI-created portfolios that can be used for tax-managed direct indexing.  Given this background, I’m surprised not to see more AI-managed ETFs. 

Here is what we found. The leader by far in Assets Under Management (AUM) in the space with $104 Million is AIEQ, AI Powered Equity ETF. The next largest with just $11 Million is AMOM, the QRAFT AI-Enhanced U.S. Large Cap Momentum ETF. This is the largest of the three ETFs by QRAFT, a fledging issuer exclusively in the AI-managed space,

The only other ETFs in the space have scant assets thus far, including:

QRFT, QRAFT AI-Enhanced U.S. Large Cap ETF

NVQ, QRAFT AI-Enhanced U.S. Next Value ETF,

DIP, BTD Capital Fund

AIVL, WisdomTree U.S. AI Enhanced Value Fund.

Of this group, only Wisdom Tree is an established sponsor of non-AI ETFs.  In fairness, DIP and AIVL have just been introduced in 2023.  Interestingly, the last three superimpose a value objective onto their AI systems. The result may have the machine-learning AI engine trying to reconcile various secondary objectives with its prime objective of maximizing alpha.  At any rate, it is obvious that thus far, large investors prefer thematic ETFs that invest in companies involved in artificial intelligence and/or robotics over AI-managed portfolio ETFs.

For a comparative analysis of these ETFs, we’ve selected the top three from the thematic category: BOTZ, KOMP and ARKQ; two ETFs from the AI-managed category, AIEQ and AMOM, are also included. 

Brief Profiles

(sourced by ETF.com):

BOTZ provides cross- sector exposure to companies working on the development and production of robotics and artificial intelligence through a market- cap selected and weighted index. Eligible companies are listed in developed countries and must earn a significant portion of their revenue from, or have a stated business purpose in, the field of robotics or artificial intelligence. This field includes varied applications from the development of drones to healthcare robots and predictive analytics software. Viewed through traditional sector classification systems, BOTZ leans heavily towards industrials and technology. The index is reconstituted and rebalanced annually.

KOMP consists of companies from developed and emerging markets that are transforming the economy. he fund’s underlying index is a composite of various S&P New Economy sub-indexes, each constructed around a theme such as autonomous vehicles, 3D printing, genetic engineering, nanotechnology. The sub-indexes themselves select constituents through a computer algorithm That scans company filings for relevant terms, selects holdings, and classifies them as “core” or “non-core.” KOMP’s underlying composite index weights each sub-index by a proprietary risk/return ratio. The index rebalances on a semi-annual basis.

ARKQ has a research-driven active mandate is to identify companies that will benefit from new technologies and automation. The fund uses its own internal research and analysis in selecting companies that capitalize on disruptive innovation that enable development in the markets they operate.

AIEQ in an actively managed by the Cal-Berkeley Haas’s School of Business’s version of IBM’s Watson. It uses artificial intelligence to analyze and identify US stocks believed to have the highest probability of capital appreciation over the next 12 months, while exhibiting volatility similar to the overall US market. The fund selects 30 to 125 constituents and has no restrictions on the market cap of its included securities. The risk constraint tends to result in a holdings list that more closely resembles market-cap weighting than equal weighting. 

AMOM uses a proprietary artificial intelligence (AI) security selection process that extracts patterns from analyzing data. The AI system finds a period within those 3 to 36 months that resulted in the best performance of stocks in the past and ranks those stocks by their residual returns, which are defined as “its total return after removal of market, size and value risks factored into portfolio construction under conventional portfolio management.” AMOM uses this metric on the stock management theory which states that stocks with higher residual returns have the potential to perform better and more consistently over time than conventional momentum stocks. The fund limits the .  weighting of a single security to 10% and no more than 40% of the fund’s assets may be invested in securities with more than 5% weighting in the portfolio. 

Given the nature of the themes covered we include these benchmark ETFs: tech-heavy QQQM, Invesco’s lower-cost alternative to its QQQ; along with IVV iShares lower-cost alternative to SPDR’s SPY. Both hold exactly the same portfolios as the older ETFs but with lower fees.

_{* VE ETF Reports are not yet available for BOTZ and AMOM}

Analysis 

KOM

1. Not one of the ETFs analyzed is predicted to beat S&P 500 ETF IVV in performance during the next year. Only one, KOMP, is predicted to perform in line with the benchmark index. None of these ETFs are sources of income with yields all lower than that of the benchmark ETF.
2. Since the first three ETFs, all invested in AI and robotics firms, are focused on future growth, not current valuations, it is somewhat surprising to note that KOMP and ARKQ both have price/book ratios below that of IVV.  ValuEngine’s valuation model also finds that more than half of the stocks held by these ETFs are undervalued. Not surprising at all is that both have negative price/earnings ratios indicating that cumulative portfolio earnings-per-share is negative.  This is frequently true of pioneering growth companies in their early stages.
3. In terms of concentration risk as measure by number of holdings in the top ten, only KOMP has less than the S&P 500 with the largest 10 of its more than 500 companies comprising just 12% of the total weight of its portfolio.  ARKQ, known for its high-conviction selection of fewer positions (37) has the highest concentration risk.
4. Between AIEQ and AMOM, the two largest ETFs actually managed by AI, AIEQ is off to a better start in 2023 but AMOM much better in the down market of 2022 and in the past 3-year period.  Both have management fees of 0.75%, the same as actively managed ARKQ.  On an after-fee basis, AMOM, posted the best performance of the 5 ETFs studied and just below that of IVV.

Looking inside the top 10 holdings of these portfolios, we identified two getting our highest rating of 5 (Strong Buy).  They are Aerovironment (AVAV) and Celsius Holdings (CELH).

AVAV - AeroVironment, Inc. designs, develops, produces, delivers, and supports a portfolio of robotic systems and related services for government agencies. It operates through four segments: Unmanned Aircraft Systems (UAS), Tactical Missile System (TMS), Medium Unmanned Aircraft Systems (MUAS), and High Altitude Pseudo-Satellite Systems (HAPS). The company was incorporated in 1971 and is headquartered in Arlington, Virginia.

CELH - Celsius Holdings, Inc. develops, processes, markets, distributes, and sells functional drinks and liquid supplements in North America, Europe, Asia, and internationally. It offers various carbonated and non-carbonated functional energy drinks under the CELSIUS Originals name.  The company was formerly known as Vector Ventures, Inc. Celsius Holdings, Inc. was founded in 2004 and is headquartered in Delray Beach, Florida.

Analysis

1. Both stocks have dramatically outperformed the S&P 500 Index and the Nasdaq during the past 5 years and during recent periods as well. 
2. Both stocks show stratospheric earnings growth numbers.  These are enough to make their PEG ratios attractive even while the other numbers are very high even when compared to the Nasdaq. 
3. On a forward-looking basis, both AVAV and CELH have 1-month, 3-month and one-year price appreciation projections from VE far above those of the markets or the AI ETFs that hold them.
4. Of the two, CELH has the superior historical performance and future projections in all the categories measured here.  AVAV while still much more a growth stock than a value stock is less overvalued by most conventional measures while CELH is more undervalued by the VE valuation model than AVAV.  Both are considered undervalued relative to their prospects.

Summary

Whether all the attention refocused on artificial intelligence in recent weeks is soon-to-be-forgotten hubbub is debatable. Investor interest is not debatable with eight thematic ETFs with more than 50 million dollars under assets, including 3 with more than one billion AUM and a fourth just below that threshold.  On the other hand, investors seem unconvinced about putting dollars into AI-managed ETFs, with only AIEQ gaining real traction - after nearly 10 years in the market.  

More By This Author:

Five Strong Buy Stocks That Are Good FIts Now For Conservative Institutional-Quality Portfolios
Rotating Regimes – What To Plan For In 2023
Appreciating The Passing Of A Titan & Incorporating Alternative Data Signals