Welcome trillions.
I'm Joel Whatever and I'm Eric Belchunas.
Eric, maybe you've noticed the media tech landscapes are obsessed with artificial intelligence right now you think, yeah, yeah, you used GPT yet.
Indirectly, but I totally know about it. It's impressive, although you know, I'm still trying to make sense of it all. I will say that it does seem like every two years something just hits the zeitgeist with the Wall Street Hype Machine, blockchain, and then it was like ESG. I just feel like now AI is like front and center. This is the new new thing. Anything with that associated with it is going to have success. Is it a bubble maybe? Is it just something to satiate the neat
marketers or is it truly the next one? Or is it truly the next big thing? I will say I was at the SEC's first ever Investment Division of Investment Management conference in DC two fridays ago. Gensler spoke at this conference and he said AI is going to be bigger than the Internet, and they're looking into how to regulate it and whatnot. So that was an eye opener because I do find sometimes you don't know at first whether something's just this huge hype marketing thing or it's
really worthy of all that attention. I'm not sold totally, but certainly in the ETF world, there's a ton of attention that's coming about AI, and we're going to continue to see ETFs that have AI in the name doing a variety of things, and so we should cover it for sure.
Yeah, and we didn't ready to talk about this for a second, but I think we've got two perfect guests to kind of walk us through it. One's going to be Rebecca Sen at Bloomberg Intelligence, who's been watching this space closely. And then we've got Dave masa chief strategy officer at round Hill Investments, which just launched an ETF called Chat this time on Trillions AI Mania. Dave, Rebecca, Welcome to Trillions.
Thank you for having me.
It's pleasure to be here.
Rebecca. I want to start with you. You, like I said, you've been watching the space closely. There's there's two distinct ways that we can talk about this, so let's just be clear about that. How do you break it down?
So if we look at all of the ETFs that haven't mention of AI Autonomous robotics, there's really two ways that we classified it. The first are ETFs that track AI companies, so that's the like of ourc roundhill ball there ETF that just launched that day we'll talk about. But then the second category are really ETFs that utilize AI. So these are ETFs that had an artificial intelligence that's telling them which stocks to buy and sell, when to buy and sell, and the likes of those would be fourth.
In Korea, they have a core ETFs and they just launched one last week. But then the other one is also AI EQ, which is tracking the IBM supercomputer and they analyze thousands of data points all day long, twenty four to seven, and they say that they could do the work of a thousand research analysts. So there's really two categories that we found.
So yeah, one is a thematic bet on the sector, right, and the other is AI is coming for all of our jobs and it's just going to take over in either mint money or drive it portfolios into the ground or somewhere between.
Yeah, So let's just chew through the second version and then we'll get to Dave and the thematic play. The first version is AI powered ETF, So these are ETFs that use AI to invest. Now, I consider this just sort of like an evolution of smart beta smart beta because it's using algorithms and numbers and stats. It's crunching numbers to try to figure out a way to get alpha. That's smart beta is sort of what that is already, and there's been AI for a couple of years. It's
not totally brand new. AIEQ, as Rebecca mentioned, was the one that was using IBM's Watson supercomputer. We recently looked at this one. The returns aren't great. It's lagging the S and P by quite a bit. I dug into why, and my conclusion was it's just trades too much. It's turnover is really high. It's going through stocks left and right. And this brings up the bigger problem with AI powered ETFs, which is you can't really solve the problem that regular
managers have, which is costs impeding onto your return. So AI is gonna, in my opinion, have to figure out a way to maybe limit costs, limit trading, and limit the fee in order to have outperformance. So it's not like AI is like somehow discovered some holy grail. It's gonna face the same challenges that regular managers have and that smart beta has, both of which have learned to get cheap and limit turnover, and they've started to succeed
because they've done that. So I think AI is early a lot of ETFs that have come out using AI or pretty high fees, the turnovers high. I think over time we might see one or two succeed if they can get some performance. But I'm a little more bearish on this side of the fence versus the thematic place.
But Rebecca, how could this version of AI powered ETFs? How could this evolve more going forward?
So I think going to air points. If we look at all of the AI powered etf from a performance standpoint, just looking at year to date, they actually LaGG the S and P five hundred. So looking at all of the ETFs, the AI powered ETF on average return three percent versus the ETFs that track AI companies return twenty percent. And they're also more expensive. So we found that on average they cost roughly seventy five basis points versus fifty
basis points. And so in terms of where the growth is, I think with the AI powered ETF, you really need to find the right fund managers, as Eric was saying, a lot of this, A lot of them churt, so they have a higher training cost. And so even though a lot of these funds are powered by a supercomputer an AI, there's still someone tweaking the model. There's still someone tweaking the code. So I think that also impacts the performance of the fund.
Okay, but the big phenomenon that we're seeing with the chat gbts of the world, generative AI is sort of what this is called, right, how much generative AI is even in these AI powered ETFs or is there yet another chapter to what this investment future could look like.
I think this is where round Hill differentiates themselves with their new ETF, because they really are in that generative space. I think a lot of the ETFs that were launched previously are a little bit more traditional. They are quantitative based, but there is still someone in the background that's tweaking the model, that's changing it. And so I think in terms of where the future goes there as we adopt AI more and more. And I think chat GPT is
a perfect example. Since chat gbt launched in November, they got more than one hundred million users in less than two months. So to put that in context, how many users do you think it took Uber to How long do you think it took Uber to get one hundred million users a year?
Two years, three years, six years, six years, okay.
Six years. So Instagram took two years and Spotify took four years, and so chat GBT got one hundred millionars million users in just two months. And so this really shows that not only is there a hype into this, but people are really invested and interested in this. And I think as you get more and more data in the AI space, that is only going to improve and grow.
So I think if we look at specifically at ETFs, the ETFs that are powered by AI, they don't have enough data points and a lot of the technology when they first launched. I think AIEQ eric that launched a while ago, and looking at all of the robotics ETF, I think the first ETIF that launched was in two thousand and six, and so if we look at where technology is now versus in two thousand and six, it's improved a lot and it's only going to grow exponentially.
That also the biggest challenge, and what nobody can get around, is that nobody really knows the future. It's just hard to predict the future and robot Ai Smart Beta Active Manager, it's just very difficult. My guess is one of these breaks out the pressro goes with like robots Win, that gets some money, but then it underperforms and maybe sees some outflows.
And you've said you're saying that you've seen this movie before. I've seen this movie before.
So now what Rebecca talked about Chat GPT and the frenzy to get on there, I think that speaks more to thematic play. In other words, let me buy some stocks that are going to benefit from this and let me get a piece of that action. That is where the thematic side comes from. Bringing Dave, now, I think that's a perfect intro. Dave, you've launched a new ETF called Chat. You're the chief strategy officer at Roundhill. What was the thought behind this and how long have you been working on it?
Well, I think Rebecca hit the nail on the head, which was exciting for us when thinking about AI is not really just AI itself. This has been around for some time, but AI's been waiting for the killer app, and the killer app with Chat gibt and that made it so that because of its use and you biquitous nature. All you need is computer access or a smartphone and you can use Genera of AI in your daily life.
That it opened up our eyes that this is not necessarily a theme or a fad that's happening ten fifteen years in the future. This can happen today and now we're seeing CEOs or as you noted, Gary Genzer, discussing AI, general AI and material ways because that's really where the productivity goes. I think it's exciting and everyone can be focused on humanoids and robots and thinking that's AI, and
it is. But the ability to do so on sort of a micro level is what Genera of AI is doing, and that's really the idea behind creating an ETF just for that space, and chat provides exposure around thirty companies that are at the forefront of general AI, whether it's the picks and shovels of a company like Navidia, or companies that are exposed to it directly through ownership about OpenAI like Microsoft, or a handful of other names that folks may not is familiar with who are involved in the space.
How quickly were you able to move on this.
We launched the EUTF color I should say a couple of weeks ago, and so we moved really quickly, and this is an idea that we were using it. We started looking at chat EBT using it again either in our workflows or just to have fun and realize, say,
there's something here. Can we do some research to identify other companies that may be exposed to it because it is a very nascent space, And it turns out once you do that peel back the onion, even just one layer, you can see that companies sort of have been poised to do this, and now that the intention is there, we're really beginning to see more and more folks latch onto it.
So how do you figure out who's a real deal player here that has unique exposure? Like an Nvidia comes to mind, right, where you've got a chip maker who is all about the creating the chip four AI applications, and Microsoft, which is a backro of chat GPT right, But then like there's a kind of a falloff. I mean, I'll keep Google sort of in that camp or Alphabet in that camp, maybe because we know that they'll have
something significant with bard. But after that, like, how do you figure out who's legit versus just generating more hype from the hype machine.
Yeah, this is really important because I think what we're gonna find is, and we saw this with crypto and blockchain, there's got to be a handful of companies that perhaps changed their name we all know one from Long Island famously, or companies that are using it but not really exposed to it. Perhaps maybe some consumer names come to mind, a Wendy's or Pepsi. So what we're doing is two things, and I call it a kind of a talking to talk and walking the walk approach. You need both here
to identify those companies that are truly exposed. So first and foremost we're transcript Score and so here we're looking at a proprietary keyword analysis of public documents, whether they are company filings, transcripts, presentations, or press releases to see if that company has references to various AI and related technology terms. And I you know, you'd imagine a company
like Navidia jumps to the top of that list. But then importantly we also have what we call a sector score, and that's to do some real quantitative research to understand is a company actually spending or has direct revenue exposed to generate AI and the related technologies, and so companies that are doing more or have higher portion of revenue coming from that, or are spending on it for the future from a true R and D standpoint, are going
to get higher score. So you pair those together, you know, subject to some standard market cap and liquidity requirements, you end up with a portfolio relatively concentrated in the grand scheme of things that has exposure to microcaps, small caps, mid caps, and of course some of the megacap names that are powering this forward. So investors can expect to see around twenty five to thirty securities going forward. It's a global portfolio, a lot of US exposure, but a
lot of China exposure in the portfolio as well. And this is going to adapt through time, which is one of the reasons why we're excited that this is actually an active approach.
How do you figure out the waitings?
The waitings are really i'd say a derivation of both the transcript score and the sector score. So subject so companies that have are talking about it a lot and are actually spending on it, we're going to get higher rates. Hence the video being the top holding, and then and then so on. So if you look at our top ten holdings, uh, you know, it's names like Navidia, Microsoft, C three, AI, Alphabet and then AMD another semiconductor which is making some inroads with g GPUs and the chips needed.
For general AI.
It's interesting.
I was at the inside ETFs conference and I looked at and I was on an AI panel, and I can't say I'm a total expert, but I looked at some of the ETFs and I brought shot up and my one critique of it. This is before Navidia's earnings. I looked at and I said, you know, I don't love theme ETFs that have big cap names at the top because I already own those in my low cost
beta core. I don't need to be redundant. Then navidia earnings came out and I felt like maybe they were justified a little more because Navidia is the top holding, and they felt a lot of that juice. But typically I find DAVE in a fledgling area where the theme might not be totally ripe enough for a full ETF equal weighting or some kind of modified market cap waiting to give you more exposure to the small and mid cap space could be better because it helps differentiate the ETF more from BETA.
Well, I think you.
Raised a good point, and we could have this conversation about a variety of different ETF areas. It's whether it's factor based, smart, beta, ESG is a whole other can of worms and stomats.
Right.
The way I think about it for this space is it would be I think disingenuous not to have the exposure to a company like the Video or Microsoft or Alphabet in a material way in this portfolio, and because they're exposed to it, they get a higher weight. But you know, again in the top ten holding there's names like iFly Tech, since Time Group, even C three AI to some extent, a company like More you know, or
a risk in networks that are not megacap growth. Right, So even among the thirty names, we have a wide representation of companies across the market cap spectrum. For US, it's really are you truly exposed to genera of AI or not? And that's what we're going to hang our hat on. And of course this market has been rewarding that that may not always be the case going forward.
But if we can provide that exposure again in a concentrated way to companies truly exposed to jener of AI, then we feel like we're doing.
Our job here.
So I guess going off that Dave, there are seventy ETFs that has AI mentioned in their description and roughly sixteen billion in assets. We expect that by twenty thirty there's going to be one hundred and fifty ETFs with the mention of AI. Given how crazy and everyone's interest in this area with thirty five billion, what do you tell investors that say, you know, how do I pick which ETFs? A lot of them do have AI in
their wording. There's a lot of hype in there. And to Eric's point, you know, just looking at NASDAQ one hundred and sm P five hundred, they have all the six names alphabet Navidia, Apple, Microsoft, and so what would you say to investors?
Yeah, so it's funny. I did actually a research paper that was published at this point, probably ten years ago about givin ndtfs, and at the time it was crazy because it had first marked that there was one hundred divid endtfs. So now as you all know, there's hundreds more. And if you look at the performance dispersion of them. They were massive, right, and at the end of the day, a big difference was whether they give it in growth, which is more quality, or give it in yield there's
more value. And I think the same can be said about the Mattos that when I in this market, it's very easy to put a buzzy name in an ETF name and see if that gets people's attention. I think, particularly post COVID experience and some of the selloff, investors have become more discriminating about the funds that they own and they understand that, you know, like our friend Todd
Rosenblue says, you need to know what you own. And some people just want to buy either the largest or most well known, but it might not be the exposure you want. So I always advocate take a look at one what is the process to get the names in there? If it's an index process, take a quick look and understand the index methodology. You don't need to be an expert in any of this space, but can you understand what they're trying to do and then do the whole
things generally reflect that? Right, So if we're in our case saying we're going to provide exposure to companies at the forefront of GENERATIVII, and a name like Navidia isn't a top holding, I'd probably question that. But also if there's names like sense Time Group, which is a leader in China in creating a computer focused AI marketplace, they have their own large language model, then that makes sense too.
So this is going to be a space where I think there's there's, as noted, a ton of continued investor interest. There could be right or wrong reasons to buying any of these ETFs, but for us, when it comes to what is really focused on, generative AI chat stands alone.
Okay, Dave, I'm curious. You've got this thematic option, You've laid out the case for it. At what point what would it take to put that engine that we talked about earlier in the show, the AI powered ETF engine, into the thematic thing, and then you have an AIM powered AI thematic ETF. What would that take? Yeah?
So, look, look, we are huge believers in the transformation that jenerava I can bring both to everything from enterprise software where our director of research publisher report focused on an estimate of our tam of over one hundred and twenty billion in a ten years time, and also the consumer applications. But the idea that AI can be used to identify the company systematically at this point in time, I questioned that to me, I would agree at the intro to the show that this may just be the
next evolution of quantitative investing. When I was a assistant quant portfolio manager in the mid two thousands, we had three components to our SoC selection model and we actually outperformed before the global financial crisis consistently, and one of those was priced book. It was so simple, but it worked, and then guess what happens? It gets arbitraged away. I think with AI, it's all dependent AI powered ets, AI
powered investment processes. It all matters at this stage of development of is it being powered by humans appropriately?
Now?
Over time, as general of AI continues to improve, as large language models and other use cases begin to become more real time, we have to remember a lot of the genera of AI potential right now is still having that kind of backward looking learning. It's we are just at the cusp of it being applied across a wide
range of industries. Then maybe I'd have more confidence in those particular approaches, But for now, our quantitative process, both sort of on the transcript side and then on the sector analysis side, gives me more confidence that will be identifying the names that will continue to be exposed to general of AI in the future.
Is there any busy work that you can unleash AI on to improve your daily life?
Dave, Uh, Well, one thing that you know, I think people are experimenting with, and we've heard some stories about about ai US is going right and also going wrong. You know, there's there's things that I use to help me craft uh to almost serve as an editor on a daily basis, for for whether it's a blog that I'm writing or just other research that I'm doing to
power that. So we're using it. In fact, if you go on the research section of Roundhill Investments website, we will we will note where where articles are being helped to be written by chat GBT. Now, I think the generative text is where all the easy attention is being paid. But I think in the short term we're going to see people experimenting with image generation, uh, sound generation and things of that nature. So again we are just at
the cusp of that. But yeah, busy work is something that we're looking to offload pretty frequently, uh to to help to help guide us and be that assistant for us.
So this brings up a good point with ETF research. I always tell my team put as much as your voice in your writing as possible, get as much human in those words, because some of this stuff is going to be automated if you are dull, you know, And I do think of this one. You know, the Hollywood writers are on strike. There was this one billboard from this woman who who was at the strike and it said chat GPT never had childhood trauma. And I do think that nothing will replace the human at the end
of the day for certain tasks. But again, dull, repetitive tasks, I just see just automated. We do it at Bloomberg on several types of data stories. There's a lot of automated stories already, but I think research and you're editing BusinessWeek over there, how much of your let's say, ten years from now, how much of the copies can.
Be written by AI? No comment, no comment, I've stumped roll. Yeah, No, I mean, look like we have there's a long ways to go, Like is AI going to conduct interviews no, So I think there's a long way for a lot of it, and I think that you're right, there's a lot of things that are rote that can be disrupted. But I think, you know, the question for humans is like where you add value, and I think there's a lot of things that humans will still add value to.
So we interviewed Kathy with and asked her thoughts on AI, and she said, it's going to add any It's going to be the area that adds the most value in the tech sector. And I think when we look at AA, there's really two areas that we can look at. Most people probably look at it from a software and hardware perspective, but her view is that if we take Tesla for instance, the autonomous driving of getting from point A to point B safely is another way that AI is going to
have a huge influence in our daily lives. And so I think it's interesting as we look at a lot of these companies to see how they evolve. You know, we don't know what they're going to do in five, ten, twenty years time, and so there's a lot of growth potential.
Dave, So if Kathy would has been there's nobody more bullishean tech in the future and AI than Kathy? Could you imagine a world where Chat invests in ARC or is that not enough of a pure play?
Wow, that is an interesting way to think about it. I think my short answer would be never say never. But I don't think as of now, that's really the exposure that we're or intention that we're we're looking to have right is And also people I think are looking to Chat to serve that direct exposure to jenitor of AI and kind of for now having another fund look to do that, it may not make as much sense.
Yeah, that's a little like just bad form if you're an ETF.
It's funny.
There was a cannabis mutual fund that used HMMJ for the longest time. It had like ten percent. There was an African mutual fund that used AFK. It's happened sometimes here and there, But I generally ETFs don't want to use other ETFs unless they're an ETF vtfs.
Yeah, okay, okay, Dave, first time on trillions. Welcome by the way again. Uh, there's a question that we often ask first timers, and I'm gonna ask it of you now because you've got an epic ticker, Chat is a great one. What is your favorite ETF ticker other than your own?
So Chat is a pretty good one. The Roundhell Investment team has a great number of.
Storied history of great tickers.
We'd don't forget about the twenty five million dollar bad Boy Meta's that's probably the greatest ticker of all time, simply because of the price tag.
But you can't use any of those.
Yeah, you'd have to keep me around all day because I could go to my time at Direction and some of the great tickers there. Gosh, drip, it's really hard to pick a few.
Just pick one in Asia it's a random number exactly.
By the way, Asia is so boring, Like in the China ETFs, the tickers are six numbers. It's like five five O three three one, And it's like, come on, the AI.
Tells you what your ticker will be.
I'm old school, and I'm so I'm gonna go with something like move, you know from from our friends over at van Ax. That was to me one of like the original cool tickers. I think it still plays a role. So I always have a soft spot.
For that one wholesome, wholesome pick that's probably the most popular pick moo tan.
You know it's funny. That's the thing I like about it.
Like you know the other It's likable and so I like verbs. Yeah, that's why chat's good. Chat's also verban nown You got a two for there.
Well, I'm trying to make it an adjective and an adverb, so give us, give us a few months.
Yeah, all right, Dave Rebecca, thanks so much for joining us on rallians.
Thank you, thanks for having us, Thanks for listening to Trillions.
Until next time. You can find us on the Bloomberg terminal, Bloomberg dot com, Apple Podcasts, Spotify, or wherever else you'd like to listen. We'd love to hear from you. We're on Twitter, I'm at Joel Webber Show. He's at Eric Baltuna's. This episode of Trillions was produced by Magnus Hendrickson. Bye