Using ETFs to tap into the AI mega-trend
Jason Hnatyk: Welcome to another edition of inside investing. I'm your host Jason Hnatyk. We've got another great episode in store for everyone here today where we're going to unpack one of the hot items in not only in the investing landscape but also at in the culture at large. Now BlackRock is making a bold call on artificial intelligence. The world's largest asset manager calls it a mega trend that will continue to drive markets for years to come one that would re: one one that could revolutionize our lives and our investment portfolios. Here in this episode we'll hear where the pocket of opportunity may lie and how investors can gain exposure to it. Joining us
here is Jay Jacobs he's the US head of Thematics and Active Equity ETFs at BlackRock. Jay it's great to have you here with us we appreciate you taking the time. Jay Jacobs: Jason it's a pleasure to be here thanks for having me. Jason Hnatyk: Anytime, uh now can you share a little bit about your own uh individual background in researching major themes in the markets and particularly artificial intelligence and how we might be able to develop them into kind of investing products. Jay Jacobs: Absolutely so I've been uh investing thematically for over 10 years which is really trying to look at what are some long-term structural trends that are disrupting major swaths of the global economy. Uh what's great about doing this for 10 years is we've already seen some of these themes really go from really nacent ideas you know think about where electric vehicles were 10 years ago even where artificial intelligence was 10 years ago, cyber security so many themes have evolved so significantly over the last 10 years and yet are still really early in their adoption curves. These trends take decades to fully play
out we like to look at them in the in this uh kind of s-shaped adoption pattern where they start out very slow usually the technology is not that great it's expensive people don't know about it but slowly that technology gets better and it starts to accelerate and it gets cheaper. And more and more people start to know about it and all of a sudden it becomes a really massive world changing theme. And we think we're really at the precipice of several of those themes impacting the world over the next several years. So I'm very excited uh to to chat with you today about AI as
one of those big opportunities but also just excited generally about the opportunity in the markets today because I think the market hasn't fully uh appreciated some of the disruption that's about to happen over the next 10 years or so. Jason Hnatyk: Yeah I'm excited as well just thinking about just the drop in time that AI's been around and kind of how much it's already started to change and evolve so really look into learning from you over the next uh next little while here in our discussion. Uh let's focus here on a little bit on on BlackRock itself. In uh you've called AI one of the five kind of investment mega trends to watch uh what's the investment thesis around AI and why does your team feel like it might be such a game changer? Jay Jacobs: Taking a step back we introduced our mega forces framework in June of last year. And the idea was really what are the five major forces that are going to impact
everyone over the coming decade. And it doesn't matter if you are invested in ETFs or stocks or bonds or real estate or other alternatives the reality is that these mega forces are happen are having a economy and society-wide impact on the world and so those five mega forces uh include AI and digital disruption. It includes a transition to a net zero economy demographic divergence that we're seeing around the world between aging populations the the aging and developed markets versus several youthful populations. It's about geopolitical fragmentation in some of the
rewiring supply chains we have around the world in kind of a post covid environment. And then finally it's about the future of finance and how finance is evolving and and different accessibility and roles that institutions are playing in finance. So AI is one of our top five mega forces in the world and it's very clear why. This is a massive theme with a huge addressable market it's not just about technology. This is about how does AI impact healthcare how does it impact uh legal services, consulting services programming autonomous vehicles automation in factories. Really every sector is going to be impacted by artificial intelligence in the coming years that gives us a lot of confidence that this is not a fad this is not something that is going to come and go in a few years. This is the beginning of a very long-term powerful trend in fact the way
that we really think about AI is this isn't one technology this is a platform technology that many things will be built on on top of to uh provide a little bit of an analogy around that you can think of another platform technology being 4G cellular connectivity before you had 4G you had 3G and it really allowed you to text people. It was a not very powerful technology it was very hard to browse the internet you certainly couldn't do things like watching videos but when 4G technology came out it let you be on social media on your phone do e-commerce on your phone, FaceTime your friends, you name it. You had such a powerful connectivity through your mobile phone that all these other applications and businesses started getting built on smartphones in creating really an entire economy around it. Artificial intelligence is like that it is a big unlock for other technologies to thrive whether it is how we interact with uh with different services how we interact with the internet how other technologies are going to be built using artificial intelligence going forward but this is a platform for a lot of economic growth and again that is really what gives us a lot of confidence that this is a mega force and not a fad. Jason Hnatyk: I think AI could even be kind of like an amplifier of the other four mega trends that you've uh that you've outlined in your in your theory as well so. Jay Jacobs: Oh I think you're you're exactly right I
mean one of the most interesting uh uh kind of relationships between these two mega forces is aging populations and artificial intelligence. So in aging populations what you're seeing is in developed markets uh you're you're seeing a slowing of uh the workforce in the last 10 years in the United States about 20 million jobs were created. In the next 10 years we think it'll only be about four or five million and that's because the United States is getting older there's less youth to replace people who are retiring out of the workforce. So what that means is if you want to
maintain economic growth you have to do more with less or you have to do more with the same amount of people that comes down to productivity. How do you empower people to do more how do you give them superpowers in the workplace. A lot of that's going to be driven by artificial intelligence if you can remove some of the more mundane tasks that you do day to day because it could be automated or you can have a virtual assistant, uh that frees you up to do more meaningful work and suddenly with you know fewer people or the same amount of people you can really continue to have economic growth so I think there's really interesting uh uh relationship between people getting older around the world and the importance of artificial intelligence driving economic growth. Jason Hnatyk: All right I'm
excited to learn more about the case for AI as the next great investment opportunity but before we get into our conversation just a quick note to our viewers. We encourage you all to subscribe to YouTube channel it's TD Direct Investing so you don't miss an episode of inside investing. Now with that being said let's continue our conversation. All right now Jay uh your team here team there at BlackRock has written in 2023 that it was just the beginning for AI what do you mean by that? Jay Jacobs: So it's it's a little bit provocative to say it was just the beginning because in reality AI has roots going back to the 1950s when Alan Turing came up with a Turing test which was really almost a philosophical idea that at some point computers would get so powerful that they would be indistinguishable from humans. So he created this idea of a test of you have to be able to test computers to see if it's human or or or electronic. Uh but fast forward into the later 1950s General Motors came out with the first robot uh you fast forward all the way to 2011 Siri showed up on our phones 2015 you started having autopilot in in electric vehicles. Um and so there there is kind of
a long history of artificial intelligence but it's taken a long time to get from idea to concept. A major major disruption happened at the end of 2002 which was the release of ChatGPT and chatGPT is a large language model I'm sure many listeners have used it already. But the idea behind ChatGPT is it made AI accessible to large populations. You did not need to have a computer engineering degree you did not need to be a robotic designer to use artificial intelligence you just needed a computer and the ability to type because all of a sudden you can interact with ChatGPT ask it to write poems, ask it to summarize an email. You name it, the use cases were tremendous and what happened 100 million people signed up for ChatGPT in two months making it the fastest growing digital platform of all time. So clearly AI had this moment not just in advancements in technology actually
a little bit of a side note ChatGPT was not a major technological advancement it was a major social advancement that allowed so many people to participate in this AI Revolution. And now we've seen an explosion of growth we've seen massive investment in the artificial intelligence space we've seen competitors to ChatGPT come out we've seen new editions of ChatGPT coming out like GPT-4 and app 5 um and this is really setting off a cascading event uh series of events that is making artificial intelligence really have this uh this renaissance. What I think will be the major shift from 2023 in the first part of 2024 to the next year or so is this is going from proof of concept to a commercial product ChatGPT was really a proof of concept that people would want to interact with AI that people would figure out how to use AI but it really hadn't been monetized yet. What we are starting to see now is that companies are integrating ChatGPT into their processes they're licensing these platforms they're applying it to their proprietary data they're developing products that consumers can use that have ai embedded into them think about travel websites that can build itineraries for you using artificial intelligence. This is a major step forward because it means it's
not so much about how good is the technology that that's that already has happened. The technology is great it's about how did these companies make money leveraging AI and I think that chart that you're showing on the screen here is really important there's really two sides to this one side is you can use artificial intelligence to get more efficient that means you can reduce costs or you don't need to hire as many people because AI can be that assistant. And another piece of it is that artificial intelligence can make your product better. It'll get people to spend more money on your product going back to that travel website that builds an itinerary for you, if it makes it easier to plan a trip, more people are going to take trips. If it makes it easier to book reservations at great restaurants people will go to more restaurants. So the idea that AI not only can be an efficiency uh technology but can also be a revenue driving product is really important as we move from this proof of concept stage that we've been in, to a commercial stage where AI should really generate terrific revenues going forward. Jason Hnatyk: All right let's continue to pull on that
thread we talked about monetization that was a huge step for X you know formerly known as Twitter as well as kind of the meta companies. Uh what's the next phase in the monetization and development for AI systems, how quickly might might that occur in your in your opinion? Jay Jacobs: Well really the next step is um reaching kind of multi multiple modalities of artificial intelligence so you know a lot of a lot of what people have probably used if they use ChatGPT is text to text. You type in I'd like to write a poem about artificial intelligence and ChatGPT will come back with a beautiful poem. Um when you can start mixing multiple modalities when you can start asking AI when you can feed a picture into AI and say describe this picture to me using text or if you can type in in a text chat bot I'd like to develop a picture that looks like myself in a webinar um that starts to introduce new modalities. Now that already exists to some extent but the more that you
can transition between text to image to video to maybe even 3D images. All of a sudden this becomes an even more powerful technology and that's where a lot of the innovation is happening today. It's around how do you kind of create a more generalized large language model where it doesn't matter how you interact with it or what medium you're interacting with, it in it can play in all these different fields. It's an incredibly powerful technology that we're really starting to see uh massive development in. Jason Hnatyk: I know your team at Blockrock has discussed the importance of what we're calling the tech stack when evaluating AI opportunities. can you help us understand what that means? Jay Jacobs: Absolutely so I think one of the important things about artificial intelligence is we we I'm guilty of this I call it a technology. The reality is artificial intelligence is a
collection of technologies. In some ways it's these large language models that we've been talking about like ChatGPT in other ways it's automation technology that's powering robots to do things on factory floors or in warehouse distribution centers. Uh it includes machine learning that can train new algorithms so there's a lot of different iterations of artificial intelligence and one of the ways to make sense of all of this and to think through who participates on artificial intelligence and most important to investors who can win from the growth of artificial intelligence, is to look at it as a tech stack the multiple technologies that are aligning to create these artificial intelligence products. At the bottom of the stack you see uh more hardware this is really where semiconductors live. So it really doesn't matter what kind of AI you're running whether it's a large language model whether that's a ChatGPT model whether it's a cloud, whether it's uh you know from anthropic it doesn't matter. They all need really powerful semiconductors to be able to train these models to be able to do really powerful things with artificial intelligence. Um
you know on top of that you look at more of the infrastructure layer uh this is um you know where is all that semiconductor power being housed, how is it being supported, it needs cooling, it needs electricity, it needs um security most importantly it needs data data is the fuel that powers a lot of these um artificial intelligence programs and so all of this kind of exists in an infrastructure layer that's incredibly important to artificial intelligence. And then on top of all of that is really the application layer which is how is artificial intelligence being used what products is it being infused into how are we finding new use cases for artificial intelligence going forward. So this tech stack really is designed to pull apart the different parts of the value chain of artificial intelligence and to make more sense of what's really a collection of very powerful technologies from an investor perspective because different parts of this tech stack are going to benefit at different part of the market cycle and at different parts of the adoption of artificial intelligence going forward. Jason Hnatyk: All right we've got a well-timed audience question from
William that's a great follow-up to your answer there Jay. He asks data processing and robotics are two branches of AI they might be some of the more obvious branches. Can you maybe get into what might be you know something that might not be as quite so obvious that's relying on AI kind of one that comes to mind for me is maybe like uh like utilities and and things like that, power generation because there's there's such you know for uh you know hydro and things like that but love to hear your your answer on that. Jay Jacobs: Yeah I think you're exactly right I mean when you look
at the value chain of artificial intelligence and what is required to run it uh electricity is becoming a really big uh challenge. Um so a lot of these chips that are used for artificial intelligence are four times as power dense as previous generations meaning you need to multiply your power consumption by four to be able to run the latest GPU's and so power is a really important part of the value chain as well. But if you're thinking about what are the use cases of AI it's really limitless. I mean one of the opportunities that we see as as really one of the biggest areas
of disruption for AI is in the healthcare space. Um this takes a lot of different forms so one form is how do you manage hospitals more effectively hospitals are uh are kind of a symphony of chaos right. You have you have emergency rooms that are staffed with doctors and nurses. Uh you have all these medical devices that are that are moving around and different patients that are moving around. And a lot of just managing inventory and managing people can make hospitals so much more efficient it's a really challenging problem for humans but something that by capturing data via artificial intelligence. Whether it's you know cameras in an emergency room tracking where people are what being used suddenly you can make much stronger inferences about maybe we need this number of doctors on a Monday versus this number of doctors on a Tuesday. Or we need this number of
stethoscopes you know you name it. You can you can kind of go wild thinking about it but you can get a lot more resource efficient in how you run a hospital. The second piece which is absolutely in the healthcare space but is very different is the development of pharmaceutical drugs. Uh right now the average drug takes about 10 years to develop and two billion dollars of investment. We believe
artificial intelligence can cut both of those in half, getting new drugs to the market faster and a lot less uh with a lot less expense because in a lot of ways artificial intelligence is really just a better way to research these things. You have all these different drug compounds that can interact with different ailments in in people's bodies and with different genetic compositions it is a very hard math problem that AI is very well suited to solve. And if AI can do that better than humans or at least assist humans in that research you can remove a lot of that time and cost from the process so this could have tremendous impact on our everyday lives if people start investing very heavily in artificial intelligence from a healthcare perspective. Jason Hnatyk: Yeah it's interesting to think about healthcare. So you think of it so disconnected from AI but there can opportunity to have a lot of cost savings and bring a lot of efficiencies and kind of coming where we've just come from over the last number of years you know being prepared for future events and and uh you know future enhancements and the aging population as was one of your mega trends I think this kind of all kind of ties nicely together. All right great so let's continue here with with some questions here. Uh so where does your team see the biggest opportunities to invest in the,
rather the AI ecosystem here at this point today? Jay Jacobs: At this point today the most interesting area is where the revenues are flowing today and that is at that hardware level we we are seeing it with semiconductors uh predominantly people that are manufacturing GPU's or graphics processing units. Um this is the common denominator as I was saying before if you are building an AI model you need to train it using powerful graphics processing units. But while that's gotten a lot of attention today we expect to see a broadening across the semiconductor space over the next several months because yes you need those GPU's you need these really high-end powerful power hungry chips to train AI but when you have these models that are up and running they run a little bit differently going forward. It's kind of like you when you learn calculus or you learn some really difficult math problem for the first time takes a lot of brain power you're sitting there you're trying to think through it you're getting tired you need to have a break and a snack and come back learn it a second time. But once you get the hang of something, you're just running problems over and over again and you're just getting very efficient at it. So when we're in this training
stage of AI it really relies on on graphics processing units but when you're powering AI to just start running that train that model on an inference basis going forward. Uh you can use more generic chips off the shelf uh it puts more emphasis on just data collection. You've trained the model but now you just want more data to run through it so that's about memory chips uh power management trips to get more power efficient as well. So right now if you look what's happened in the market a lot of the focus has just been on a handful of GPU manufacturers um but we see a much uh bigger opportunity broadening out across the semiconductor space as AI becomes a little bit more of a of a daily process rather than this kind of transformational um training model. I think on top of that you can also look at the infrastructure layer as well as we as we see more uh artificial intelligence demand we're going to have to build more data centers.
There's going to have to be more services in those data centers providing the power the cooling the security to all those chips and all that data and so this is a really important layer as well which is already starting to see a lot of revenue growth uh as we see a demand picking up. So the the today opportunity the most immediate opportunity is really at that hardware level for semiconductors and digital infrastructure. Jason Hnatyk: And you've talked about the need for kind of investment dollars coming into the AI space. How does the, but that's going to require kind of a commitment from like the population at large how do we see you like popular culture weighing in on the AI space like it's it's been in you know movies for as long as we can remember you know Terminator 2 you know you you mentioned you know Alan Turing with the imitation game even X Machina might come to mind. How do, how does popular culture weigh in on the investing side of from an
AI an AI business case? Jay Jacobs: There's been a fascination with AI for several decades uh that's that that's certainly the case I think um you know part of the amazing aspect of ChatGPT though was that it didn't just exist in stories anymore this became a product that we could use in our everyday lives and I give some credit to Siri and Alexa and I give some credit to you know uh advanced driving tools that that can give you you know autopilot in certain cars. Um all of that is making us more comfortable with artificial intelligence every single day. Um so I think pop culture uh you know is it was kind of a a leading indicator here of of where uh you know where we were going to see more AI in our everyday lives but it's not until you really get to roll up your sleeves and use AI that you start to understand it. Now this does impact what's happening in in the everyday world
though a really good example here is when ChatGPT first came out New York City public schools within a couple of months actually banned the use of ChatGPT they said we don't want our students using it this technology is going to result in people writing essays and you know cheating on essays and not really being able to learn how to write and do these other things. A few months later they came back and they reversed that policy because they realized if you're going to be a worker in the 21st century you need to use how to you you need to know how to leverage this tool. It's like a pencil it's like a calculator you can't go through your life not knowing how to use this tool that doesn't mean it replaces the ability to write. it doesn't mean a calculator replaces the ability to learn math and understand it. But you need to learn how to leverage these tools to be uh you know at the cutting edge and so I think that um a lot of the fear around artificial intelligence has resulted in some kind of snap reactions that actually are not the right way to react to it. We
should be viewing this as a tool we should get we should be getting all generations but especially the next generation very familiar with it because this is going to be on their desktop for their entire careers. Jason Hnatyk: All right so we've talked about seeing a lot of immediate investment opportunity. You mentioned GPU's on the hardware side of things in the immediate term what about a longer term scale um such as AI interference becomes uh or inference rather becomes more widespread. Jay Jacobs: I think a really interesting area for long-term opportunities is going to be in the software space and I think this is where it requires a additional level of nuance because some companies will benefit whereas some companies will be disrupted by artificial intelligence. Um why
is this the case well artificial intelligence and these large language models like ChatGPT they're not just good at writing poems they're really good at coding because coding is a language and you can train uh large language models on any language not just English but other coding languages and it has really become a superpower for software engineers. That there are things that software engineers can do with three people that used to take 10 people to do and so you can have a really rapid software innovation now using not that many engineers going forward. And that means that if you were a software company that really had a moat around you because nobody else had just programmed that software yet that moat might not be as powerful anymore someone that could have been five years behind you is suddenly a year behind you or two years behind you and that that really presents a big risk to some software companies. On the other hand because you can do more with less in a in a space where software engineers are very well paid and there's very low unemployment frankly there's shortages in software engineering right now. Um you can do more with less and you can innovate more quickly and develop new products that can generate more revenue so you can see how there's kind of this this split where some software companies I think are going to leverage this to get more efficient and develop more products more quickly other software companies are going to realize that whatever software they developed in the past is actually not as distinct or valuable as they thought. And this is where I think really um being kind of active in how you think about a specific
space like software is important to differentiate between those winners and losers. Jason Hnatyk: All right with that being said there's bound to be a lot of disruption in the software marketplace as AI becomes more widespread how does BlackRock think about picking winners in this space? Jay Jacobs: Yeah so it really comes down to who is going to leverage this technology best to provide a better product or to get more efficient you know we talked a little bit about that example of of travel companies where if you can get people to book more trips because you have artificial intelligence helping them develop an itinerary that's a great revenue driver going forward, and I think if we look at across a lot of software companies and a lot of internet companies will realize that some you know small advancements in artificial intelligence can have a huge impact. Um you know another area where this could be really impactful is in the streaming space. So you know when you log into your streaming service of choice it is recommending the next TV show or next movie for you to watch. I don't know about you, the hit rate for me tends to be pretty low I don't know if I just have unique tastes but I feel like those recommendation engines are not really dialed into my preferences yet but as artificial intelligence gets better it'll better understand me through my data and consumption and recommend better movies and better TV shows or even better yet. It'll inform those streaming services of what TV shows and movies they should be developing and maybe even because of the multiple modalities of large language models it can rapidly create new TV shows that are very specific to certain niche audiences. So suddenly it's not um you know if you go back in the history
of of television you know you used to have these broadcast television shows that were trying to appeal to really broad audiences in the future just could get more and more specific television shows or movies for very niche audiences and that'll be the first thing that's recommended to you in a streaming service. So anywhere where big data exists and anywhere where that data is proprietary to that platform there's tremendous value to be unlocked via artificial intelligence. So that's really one of the keys that we look for is who owns proprietary data that can be used to make their product better or more powerful. It's streaming services it's some software companies it's some of those healthcare companies that we are talking about that have tons of data that they haven't fully leveraged yet and we think that AI is really going to help monetize what is really kind of a scarce and unique and valuable resource and data. Jason Hnatyk: All right now what could the disruption from AI mean for revenue and earnings growth for companies that are able to develop and effectively leverage the technology. Jay Jacobs: It really it it cuts both ways right so if you're in a growth market you really want to be growing topline revenue and I think a lot of companies are already looking into this how do use artificial intelligence to build better products that we can charge more for uh you know think about your desktop we're seeing artificial intelligence already brought into um search engines being brought into word processors being brought into presentation deck managers. All
of those things that's not just happening through the goodness of of technology's heart here. They want to make a better product that they can charge more for so I think that we're going to absolutely see topline revenue growth through the integration of AI into existing products. Um on the other hand uh if you have kind of nonlinear scaling with product development through artificial intelligence as a resource I think this is really key as well. But that was a word jump but what I
mean by that is if you want to double the amount of products you develop as a tech company but you don't want to double the amount of software engineers, artificial intelligence can make you a lot more efficient so that's the other way that I think revenue and profitability are connected in artificial intelligence here which is you have a lot of growth companies that want to develop more products they want to develop those products faster. They want to get them to market artificial intelligence is going to help enable that and do it in a way that is really uh able to control costs and be more efficient as well. So I think there's just tremendous opportunities not just at the revenue level for companies but an overall profitability as well. Jason Hnatyk: All right and as a follow up to that do you think we'll have an ecosystem of a bunch of different smaller a AI platforms or a few larger platforms that that dominate the space kind of like like guess getting to the point do you think there's going to be a lot of M&A activity uh in the in the AI offering as we move forward as well? Joel Jacobs: The short answer is is sort of yes to all of those questions. I think at one end of the spectrum what we've seen in so many uh technologies in the past is winner take all or winner take most markets. You know how many search engines do we use, how how
many smartphone manufacturers exist how many social media platforms? There's really kind of a handful in each of those different buckets so we would expect for these large language models these big platforms uh that are developing artificial intelligence it'll probably come down to just a handful that really are the market share leaders. However beyond that when we go back to the tech stack there's so many different ways to apply those large language models or those different AI platforms uh to unique use cases um and this is where I think there's going to be smaller companies that really start to thrive. What if you're a company that solely focuses on how to use ChatGPT to make better legal contracts cheaper. So you have a specialty in the legal field you have a few lawyers on staff you know the clientele of who wants what type of legal documents and so you're using ChatGPT as the platform behind this but maybe you're infusing it with your own data or your own inputs to be able to craft really good legal documents and sell that into a very specific market. So you know I don't know that's that's probably a smaller midcap company you could think about an iteration of that that's just focused on the healthcare space. You could think about iterations of that that are just focused on the software engineering space. So I do
think we're going to see an economy of smaller players that are taking these large platforms and applying them to really specific parts of the economy. And then ultimately do these small players get bought up I think that's also a possibility as well. Um but this is the likely uh path forward for artificial intelligence it's not that everyone's going to understand how to use this off the shelf on day one. You need these smaller players that can take these AI platforms and apply them to really specific use cases. Jason Hnatyk: All right so want to be an equal opportunity here we've talked about a lot of the benefits as well as some of the investment opportunities now let's talk about some of the key risks to AI's widespread adoption and the development for investors. What do you think about that Jay? Jay Jacobs: Like any new technology there's absolutely just risks in technological adoption.
Uh we've seen other periods of AI's development where it seems like there's a huge unlock and then it slows down you know think about when Siri and Alexa came out really splashy really exciting and then you know we realized the use cases may have been more limited so it took a little more time for those technologies to develop. AI will not be you know one way in its accelerated growth. I think we're going to see kind of fits and starts, but a more specific uh risk to artificial intelligence I think is cyber security. So at its heart artificial intelligence is making data more
valuable it is taking data that is unstructured and making it structured. It is taking data that hasn't been analyzed and turning it into valuable outputs you know think about you know what TV show I'm going to watch next it's making that data more valuable if it can be more predictive of what TV show I want to watch. The problem with more valuable data is more people want to steal that data or disrupt the processing of that data and that's where I think cyber security is going to be really important this isn't just about what TV shows I watch if you think about proprietary data across the economy. Healthcare data for example um if that becomes more valuable people will want it
for various different reasons. And so I think uh you know at the top of the show we were talking a little bit about themes that are happening in um uh kind of tangentially to each other that aging populations and artificial intelligence are actually kind of intrinsically related I think the growth of cyber security and artificial intelligence are intrinsically related as well. Because as data gets more valuable through AI cyber security has to keep up to play defense against bad actors in the space. Jason Hnatyk: All right now BlackRock is one of the ETF providers that offers
thematic ETFs focused on AI, why do you feel investors may want to consider gaining exposure to AI through these products? Jay Jacobs: One of the challenges with AI is that uh people know of maybe a couple of names you know I know there's been a lot of focus on the Mag 7 or maybe realistically what's the Mag 4 this year of companies that develop semiconductors and companies that develop some of these large language models but the reality is that's just the tip of the iceberg there's dozens if not well over a hundred companies around the world today that are playing a really important part of the AI and Robotics ecosystem. So a thematic ETF is really designed from the bottom up to capture the entire opportunity set of a specific theme. Not just even looking within one sector or even just within one geography but looking around the world at the leaders in a specific theme like artificial intelligence. So this gives investors a lot more diversification. Uh it doesn't matter so much just how the top names perform but how does the entire basket perform uh and also it makes sure that you're planting a lot of seeds. You know we don't necessarily know who will be the leading AI company 10 years from now uh but we think it's very likely if you're casting a wider net across several dozen names around the world today that those that the likely AI leaders will be a part of that ecosystem. So that diversification getting value to getting exposure
to the envire entire value chain not just a few household names is really important for thematic investors to make sure they participate in the theme and manage their risk. Jason Hnatyk: That seems all to me all the more important if you're casting that broad net you were talking about the adoption of ChatGPT of the what was it just a few months in terms of 100 million subscribers so there could be somebody else out there that's going to rise to prominence in a very short period of time. So yeah good uh good good thought. Um all right so if an investor wanted to gain exposure to kind of the AI mega trend using ETFs how might they want to think about selecting funds based on the way the industry is developing? Jay Jacobs: So there's different ways to think about it. So we you know we talked about thematic ETFs just now which are really designed to capture an entire value chain and really for a lot of investors that's going to be a great way to play something like artificial intelligence don't try to get too picky or too granular own the entire value chain and benefit from a rising tide within artificial intelligence. However we we do see some investors want to get more
granular looking specifically at different parts of that value chain or different parts of that tech stack where they see opportunity you know we talked about it earlier semiconductors really play a really important part of this ecosystem and are generating revenue today in artificial intelligence which not all companies are. Um that can be a really specific layer that people want to get exposure to and they might use a sub sector ETF to get semiconductor exposure or to get digital infrastructure exposure you can get just exposure to the you know handful of companies that are participating in that layer I think one of the important things for investors to think about though is not just kind of what they're buying but where does it fit in the context of a diversified portfolio. AI is a tech heavy theme. We're talking about semiconductors we're talking about digital infrastructure we're talking about large language models yes there are some names in healthcare and other sectors but I would say you know you know largely speaking over 50% if not even close to 80% of the names really in the artificial intelligence space are some form of technology company so it makes a lot of sense if an investor wants to get exposure to AI you might want to reduce your tech sector exposure or sell down an exposure within your portfolio that's already tech heavy and replace it with an AI specific thematic ETF. That's one way to get exposure. The other way is to think about the core of your portfolio where you have you know really broad exposures across the United States, Canada, globally you can shrink that core a little bit maybe 80% of your portfolio is core holdings and you carve out a thematic basket outside of that which you're going to recognize is this is for long-term investing this is going to be more volatile because it's not as widespread as the core. This is kind of a subset of companies uh and I'm going to pick a few themes to diversify
across and really have kind of this long-term growth generator for my portfolio. And that's where AI could sit it could sit alongside other themes we talked about demographic divergence we talked about transition to net zero economy you could pick a few themes in that bucket and try to diversify so you're not all in on just one disruptive trend. Jason Hnatyk: All right let's pivot to a question from the audience this one's coming to us from Bruce. Uh Bruce says there's so much investment piling into AI stocks do you see a risk of investors overpaying for technology that can't meet sky high expectations? Jay Jacobs: I think the risk we see in the market today is too much concentration in a handful of stocks and this is a big difference between when you look at kind of the Mag 7 when you look at the entire value chain of ETFs you know the Mag 7 a lot of those companies have had a very significant runup in the last few years because of AI exuberance but as I mentioned there's this long tale of dozens of companies in the AI space that haven't benefited whether they are other semiconductor manufacturers, other digital infrastructure companies other AI application developers these data owners that own really valuable proprietary data. They have not experienced this bubble as much if at all in fact some of these companies underperformed the broad market benchmarks last year even despite AI exuberance. And that's because simply the market hasn't fully realized that these are artificial intelligence companies so um there there's there's a risk of being too concentrated in these popular names but if you look more broadly across the value chain that uh that diversification uh that breadth of opportunity uh that that risk is certainly lessened. So you know for for investors look you know listening to here I would say don't
just focus on those top names this is not a stock story this is an economic story and you should be getting that exposure across the value chain. Jason Hnatyk: All right you've mentioned magnificent 7 a couple times it's in the news quite a bit and the next question comes to us from Ashley she's got a very pointed comment about uh about the Mag 7 the tech stocks have more money than they know what to do with. Won't they just buy up all the AI Innovation to maintain their dominance, what's your thought on that? Jay Jacobs: They are using their money wisely in the AI space and what I mean by that is you really need a few different ingredients to succeed in AI right now. You need a lot of AI engineers or a lot of software engineers to build these programs you need a lot of valuable data and then frankly you need a lot of money because these chips are not cheap and running these chips is not cheap. So it's no surprise when you look at the Mag 7 that a lot of these companies are large cap tech companies they have engineers they have data and they have a lot of access to capital um but again this is still the early stages of AI it's the early stages where new models are being trained um where we're still developing a lot of this technology as it broadens the opportunity set is going to go far beyond those those big names so yes I think absolutely will the Mag 7 be acquisitive in the space and and buy up some small caps I think that could absolutely happen. Does regulation come
into the space to to prevent that that's also a possibility so you know this is this is early stages and I think uh this ecosystem will evolve significantly over the next 10 years with new companies entering getting acquired, some companies failing too. This is a part of the adoption process of seeing winners thrive and and and losers not being able to make it. Let's look at any technology in the past uh let's look at ecommerce let's look at the internet back in the 1990s some of those early leaders were not necessarily the me you know the um the Mag 7 of today. Uh they've been fully replaced and again that's the importance of diversification here. Jason Hnatyk: Now when BlackRock's
thinking about uh you know a thematic AI ETF when we're thinking about its construction are we thinking about looking at market cap are we thinking about sectors specifics or maybe kind of talk us through your thought process on on that portfolio construction? Jay Jacobs: The driving principle is getting purity to the theme. We want to be owning the companies that are leading in the theme no matter where they sit. Whether they are classified as a tech company or a healthcare company whether they are in the United States, Canada, Europe, Asia um whether they're a small cap, midcap or large cap. Um we really are focused on what are the pure play companies in this space. Now that's how we select the companies that are included in these funds but also we want to make sure we're providing diversification so one of the problems with uh the Mag 7 is these companies are so big that in market cap weighted indexes that tilt their exposure to bigger companies they just dominate the entire exposure. You could have 50% of your of your portfolio is just weighted in those Mag seven names because it's market cap weighted and these are multi-trillion dollar companies. So we don't just think about what companies we buy we also think about how do we plant our seeds across that portfolio, and one of the ways we do that is really just by trying to cap exposures so no one company completely dominates a specific index. That gives us more diversification it plants
more seeds across the basket so that if one of the Mag seven disappears in 10 years which is totally possible uh it limits the damage that that puts on the portfolio. Also if the Mag 7 keeps growing we do have exposure to those names as well so it's really about balancing these tradeoffs between being overly concentrated and diversification. Jason Hnatyk: All right now I know the word bubble can sometimes be a little bit of a four-letter word in the investing space but there's been lots of hype around the market with AI and we've seen tech stocks you know skyrocket very recently as a result. How much of a risk do you think there is of a bubble being here from an AI perspective? Jay Jacobs: I think for investors taking a long-term view of AI that risk is pretty mitigated. This is a multi multi- trillion dollar economic transformation some estimates have this being 15 trillion dollars of economic disruption due to AI so whatever valuations we see today and the size of the companies involved today pales in comparison to the entire economic opportunity. Now with that
being said could we see a selloff in the next six months where some of these names come back pretty aggressively. That's always a possibility could we see volatility over the next 10 years of ups and downs absolutely that's part for the course in these themes that are still early. The market is simply still trying to price what is this theme worth. And often times we find the market underprices them that systematically people are too short-term minded and don't recognize the 10-year opportunity they're just looking at the two-year opportunity. However um uh you know we we do continue to think that over the long term despite that volatility that we are certain will be there uh this continues to represent a huge opportunity for investors. Jason Hnatyk: Now kind of focusing
back on the tech stack a little bit thinking about companies that are enabling AI or think about companies that are maybe supporting its its its use and then there's even companies that are using AI. Is there one particular area that you see the most value in? Jay Jacobs: The most value is also where there's the most uncertainty which is at that very top level of the tech stack. The applications, what industries are going to be built using artificial intelligence again let's go back to the analogy of 4G that I was mentioning earlier. The big winner of 4G technology was not necessarily the telecom companies that developed 4G. Um it was the companies that learned how to harness it and to develop new Industries a mobile social uh social media world. A mobile e-commerce world um all of these uh all these industries were completely revolutionized by 4G so I think in the in the application space AI will have a transformative impact on several different industries we can make informed uh guesses you know we've talked about several of them from from healthcare to you know consulting services and legal services and programming um but there's also companies that and industries that are just getting started using artificial intelligence and I think that's where the opportunity is but also where some of the risks are because it's still so early. Uh you know I think anyone who says they know a certainty how this is going
to play out is is is bluffing. We have to see how this is going to play out and we have to see the new industries that are going to be built using artificial intelligence as their platform technology. Jason Hnatyk: So remaining nimble but uh kind of to to see what comes next right. Um all right great next question this one comes to the comes from the audience we have one from Stephanie here Stephanie asks what's the most effective financial metric that you use in valuating AI equities? Jay Jacobs: You know I think the most valuable metric is is really growth it because we are so early in the growth stage of these companies and and topline revenue growth specifically. We're so early in these companies
what we want to see is that they are taking this incredible technology and they are commercializing it and they are entering into a large addressable market. So if you are a um you know where we see this today is in semiconductor companies they're seeing tremendous topline revenue growth because there's so much demand for artificial intelligence they are rightfully charging a lot of money for these semiconductors because they're very valuable they're selling them as quickly as they can make them they are really driving a lot of topline revenue growth um as we look you know kind of around the different parts of the tech stack we want to see that as well. This is that shift that we were talking about from proof of concept to commercialization who is able to take this incredible technology and sell it and make money off of it. And so if we see that high topline revenue growth I think that's a really important metric that this company is doing something right about harnessing this technology now revenues in isolation aren't the answer. What you really
want is profitability but not necessarily at this stage when there's such a high growth opportunity ahead of these companies they're not trying to pay a dividend to shareholders what they're trying to do is reinvest that revenue reinvest that cash flows in their business to develop the next iteration of their semiconductor or to develop the next iteration of their software platform to sell to even more people. So that's why early on in the early stages of a theme like artificial intelligence we want to see high or increasing topline revenue growth. Earnings will come down the road and that and and five years from now when we do a webinar again on this I'll be talking about earnings but right now it's revenue growth. Jason Hnatyk: That was kind of going to be my follow up to that question. Is there, is it a five year 10 year 15 year timeframe that one can realistically expect or or look for profitability or is that is that too is it too early to tell and put an actual number on that? Jay Jacobs: Different parts of the tech stack will be at different stages I think when you look at those lower parts of the tech stack and hardware and digital infastructure I think that's you know zero to five years where we're going to see profitability. Um when you look at those applications uh when you look at the large language models you look at some of the data that could be maybe 5, 10, 15 years out but that's okay hey I mean some of the Mag 7 stocks today took a very long time to be profitable and that was by design. They wanted to fully grow as fast as they
could into as large of a market as they could before they focused on profitability. And I think in artificial intelligence you'll see a similar pattern with a lot of these companies it's going to be grow as fast as you can into as big of a market as you can and then focus on profitability five, 10, 15 years from now. Jason Hnatyk: All right great our next question comes to us from William. Uh we know AI is a very hot button issue and so is green energy. So uh William wants to understand how AI can be disruptive in the green energy space? Jay Jacobs: Well AI is both a consumer of energy as well as potentially a user case for how to get more efficient with energy so we were talking about this earlier but AI is very power hungry it requires a lot of power to run these semiconductors so one of the ways that we should be thinking about this is, how do you manage your power mix to run all these really powerful data centers going forward you know power consumption in the United States for the last several decades has basically been flat. It's been very little growth despite population growth and a lot of that has to do with a shift from uh really from manufacturing to services in the United States which is less power hungry. But all of a sudden we're seeing power demand start
to tick up again and it's because of electric vehicles, it's because of artificial intelligence, it's because we're seeing some manufacturing growth again and so we do have to think about how do we bring incremental net new power to um to the grid and how does that work with other technologies like artificial intelligence. So um absolutely there's going to be more power demand for AI but I think maybe the nature of the question is how can AI be applied to be more power efficient. Some of these things are are are really uh simple right how do you optimize buildings to use power at the right times? How do you put more sensors in buildings to turn off lights when people are not around? Uh how do you do cooling in a better way because you have better anticipation of what the weather is going to be a day before and you can start heating a large building or cooling a large building as necessary? Uh how do you control traffic patterns more intelligently because uh you're able to modify your your uh infrastructure in a smart way where it's changing the frequency of green lights versus red lights or changing the number of lanes in each direction, very, in a very smart way using artificial intelligence? So there's a lot of ways you can use AI to get more efficient not necessarily about um uh producing more power using green energy but how do you get more efficient with the power you use today using artificial intelligence.
Jason Hnatyk: All right so we've talked about one hot button issue from in terms of green energy let's move on to can be another thing that's making a lot of headlines how can AI and crypto how does how does something like that intertwine itself together? Jay Jacobs: You know there's different uh there's different ways that these themes are related related. Um you know my view is artificial intelligence is you know we've talked about this with our mega forces right we have our future finance framework which crypto and digital assets fits into whereas we have artificial intelligence and and digital uh digitalization and another part of the mega forces framework um I think they can collide anywhere where there's data these two things can collide we're absolutely seeing artificial intelligence more broadly being applied to the financial space right now. How do you go through earnings reports more efficiently um how do you make predictions more effectively of where the markets might be going based off of data and how do you consume that data faster and use artificial intelligence to make inferences from it so I would actually in a lot of ways kind of lump digital assets into a broader financial question of how artificial intelligence can make us more efficient in the financial services and BlackRock in many ways is is implementing this in in real time
2024-04-17 22:46
