Two ETFs for Artificial Intelligence and Robotics
The best way for investors to get exposure to any given investment theme is by finding an Exchange Traded Fund (ETF) that tracks said theme. The problem with many of these ETFs is that they aren’t exactly focused on pure plays. The creators of the ETF want a broad enough selection of stocks within the ETF so investors can enjoy the benefits of diversification. Of course, that means getting a bit creative as to what might be included. One recent example we came across was Morgan Stanley’s list of “space stocks” that included Facebook and Apple, neither of which are doing much in space at the moment. However, these things can often become a self-fulfilling prophecy. If enough people invest in an ETF, then suddenly it starts to perform well and gains credibility simply from the large number of Assets Under Management (AUM).
One ETF that has been performing quite well lately is the Robo Global Robotics & Automation Index ETF (ROBO) which we highlighted before and which now has $2.3 billion assets under management. To put that number in perspective, we can take a look at the top-100 ETFs by AUM and see that the smallest ETF in that list (Vanguard Healthcare ETF) has more than $7 billion in AUM. In other words, just to crack into the top 100 most popular ETFs, ROBO still needs to triple the number of assets they are managing. If they did manage to do that, they’d also be raking in nearly $22 million in management fees. That’s because ROBO ETF charges a .95% management fee which would be considered well above average. (The average ETF carries an expense ratio of 0.44% according to Morningstar via the WSJ.)
When it comes to ROBO’s performance, we get mixed reviews depending on the time frame. If you bought shares of ROBO on the day they first started trading (10/21/2013 ), you would have made about +64% on your money over a 53-month time frame while paying some pretty hefty management fees every year (some of which they use to pay 5 PHDs to scout out robot stocks). Below you can see the performance of ROBO (blue line) compared to the most popular NASDAQ ETF out there, QQQ (red line):
If instead of buying ROBO you bought a NASDAQ tracker ETF like QQQ, you would have made around +87% on your money over the same time frame while paying just .20% in fees. That’s because QQQ is a tracker ETF, meaning it tracks the NASDAQ index automagically without needing the input of 5 well-paid PHDs.
Of course, you can play around with shorter time frames and then ROBO starts to outperform QQQ but we’re not interested in short term speculation, just long-term performance. We’re also very interested in investing in robotics and artificial intelligence, so our attention was drawn towards two ETFs out there that claim to provide exposure to robotics and artificial intelligence:
- Global X Robotics & Artificial Intelligence ETF (BOTZ)
The First Trust ETF, ROBT, just emerged on the scene last month so let’s tackle that one first.
The methodology for ROBT starts off like an MBA textbook with company categorizations like “enablers, engager, and enhancers” which are described as follows:
- Enablers – Companies that develop the building block components for robotics or AI, such as advanced machinery, autonomous systems/self-driving vehicles, semiconductors, and databases used for machine learning.
- Engagers – Companies that design, create, integrate, or deliver robotics and/or AI in the form of products, software, or systems.
- Enhancers – Companies that provide their own value-added services within the AI and robotics ecosystem, but which are not core to their product or service offering.
There’s probably a pretty cool Venn Diagram somewhere out there that helps explain this classification even better. ROBT then applies certain filters which should be standard for all ETFs (we’re looking at you HMMJ) and restricts the stocks they select to be above $250 million in market cap, and to have sufficient liquidity and free float. This handy flowchart explains it better than we ever could:
So what’s the end result? An ETF that charges a .65% management fee and has a rather odd mix of 98 stocks with weightings that range from the largest (Hortonworks) at 2.25% to the smallest (Tesla) at .44%. The largest holding is a company that most people may not be familiar with. Hortonworks is a $1.5 billion company that builds and sells a platform (based on something the software developers among us will know as HADOOP) that enterprises use for storing and manipulating “big data”. Let’s look at the next 10 companies in the ETF:
Half of these names we’ve come across before. Illumina is a great picks-and-shovels play on genetics, but not really AI and robotics. Nuance Communications does natural language processing using AI. Ambarella is a play on a number of disruptive technologies including autonomous driving. Cloudera is a company that we had a hard time understanding, but which does something similar to Hortonworks we mentioned earlier. FLIR builds thermal equipment which we came across when writing about how to create the perfect security drone. The other half of these names we’re not so familiar with the exception of Blackberry which somehow made the list.
Dassault Systemes is a $35 billion European company that builds 3D design software and product lifecycle management (PLM) software. Teradata Corporation is a $4.8 billion provider of database-related products (so in the same set as Cloudera and Hortenworks). Technology One Limited is an Australian enterprise software development company. Lastly, there is ServiceNow, a $28 billion cloud computing company that does cloud stuff.
It’s worth noting that the one company which is probably benefiting the most from the rush to buy AI hardware like GPUs is way down on the list. Nvidia (NVDA) is only given a .68% weighting in ROBT. There’s also a smattering of the usual suspects like Apple, Facebook, Softbank, Google, Sony, and Samsung along with a number of Japanese robot stocks of the type that ROBO is more heavily weighted towards. If you’re wondering who actually builds the index for ROBT, then that would be Nasdaq. While the ETF is branded as a Final Trust investment product, the underlying index that powers the ETF is the “Nasdaq CTA Artificial Intelligence and Robotics Index”. It’s only rebalanced twice a year, so it won’t be changing often. It’s also nothing like our next ETF which is said to provide the exact same type of exposure.
Global X Robotics & Artificial Intelligence ETF (BOTZ)
BOTZ has been around a bit longer than ROBT having first started trading on 9/12/2016. It has quickly attracted more than $2.4 billion in assets which could be related to their much more reasonable management fee of .65%. The BOTZ ETF is based on the underlying “Indxx Global Robotics & Artificial Intelligence Thematic Index” which is significantly more concentrated than ROBT with just 29 stocks. While the top-10 stocks in ROBT accounted for just over 21% of the ETF, for BOTZ we can see that the top-10 holdings make up nearly 70% of the entire portfolio:
This ETF is much more concentrated in robotics with 6 of the top 10 holdings consisting of Japanese robot stocks (highlighted in yellow). We also see Nvidia having the heaviest weighting in the entire portfolio, though this is likely the consequence of Nvidia shares having risen an astounding +285% since first being added to this ETF. That brings us to performance. BOTZ has returned +64% since it first started trading in September of 2016. Below you can see the performance of BOTZ (blue line) compared to a Nasdaq tracker ETF called QQQ (red line):
As you can see in the above chart, BOTZ has managed to outperform QQQ since inception by about +20% so far.
The key takeaway here is that there are wide and varying opinions about what stocks out there will give retail investors exposure to “artificial intelligence stocks” and “robot stocks.” Speaking of robot stocks, wouldn’t it be nice to see how many stocks are common among the three robotics ETFs we just talked about – BOTZ, ROBT, and ROBO? Remember how we were talking about cool Venn Diagrams earlier? Here is one that shows the common stocks among all three ETFs as of the day we published this article based on the latest info available from each ETF provider:
Turns out that there are only 18 stocks found in all three ETFs which means investors will get three completely different market exposures from these three ETFs though they all claim to be offering exposure to the same themes. This seems about right since none of the pundits ever seem to agree upon which stocks provide exposure to new technologies for retail investors. Many take the easy way out and just say buy shares of Google and you’re sorted for all of them.
Conclusion
We can probably expect that soon there will be smart AI algorithms scanning all online collateral and SEC filings for each company out there to figure out who the real “enablers, engager, and enhancers” are in the AI space. Until then, you may want to hop on for the ride and pick up some shares in some or all of these ETFs and follow the money. Or you may want to go right to the source and buy some Japanese robot stocks on the Japanese stock exchanges.
One little-known Norwegian company has developed a cubic warehouse automation solution that's taking off like wildfire, but only some investors can buy the stock. Become a Nanalyze Premium annual subscriber and see all the tech stocks we like and the ones we're holding ourselves.
