First Trust ROI Podcast

Ep 37 | Ryan Issakainen | Ripple Effects & Second Order Investment Opportunities Arising from AI | ROI Podcast

• First Trust Portfolios

Share episode

Companies are investing billions to build out artificial intelligence capabilities.  Ryan shares First Trust’s perspective on various investment themes that may benefit—either directly or indirectly—in the years ahead.

----------------------------------------------------------------------------------------------
Subscribe Here to the ROI Podcast & other First Trust Market News
Website: First Trust Portfolios
Connect with us on LinkedIn: First Trust LinkedIn
Follow us on X: First Trust on X
Subscribe to the First Trust YouTube Channel
Subscribe to the ROI Podcast YouTube Channel

Ryan:

Hi, welcome to this episode of the First Trust ROI podcast. I'm Ryan Isakainen, etf strategist at First Trust and host of the ROI podcast. This is our first podcast episode of 2025. We had a wonderful time talking with a really insightful group of guests in 2024, many of whom will join us again in 2025. And we also look forward to joining some new guests, or some new guests joining us in the year ahead.

Ryan:

Starting with our second episode of 2025, where I'll be joined by Mandeep Singh from Bloomberg Intelligence. He's a senior technology analyst. We're going to talk about where technology is going to be taking us over the next several years. We're going to spend a little bit of time digging into artificial intelligence, what the state of that is today and what we can expect going forward. What we can expect going forward. For today's episode, I wanted to share some of our thoughts on the years ahead 2025 and beyond when it comes to some of the trends and investing themes that we're really paying close attention to at the outset of this year. And, of course, I'm sure my conversation for our next episode with Mandeep Singh will help to augment some of those thoughts, because you certainly can't talk about investing themes without touching on the impact of artificial intelligence.

Ryan:

This is already having a big impact on the way that companies and organizations allocate capital. We look at the value of certain companies that provide the guts, the semiconductors, the infrastructure related to artificial intelligence, and those companies have soared over the last couple of years. They've been driven by massive growth in revenue and, of course, that shows up in the capital spending of some of the largest companies in the world. For example, we look at Amazon, microsoft, google and Meta four of the largest companies in the world. For example, we look at Amazon, microsoft, google and Meta four of the largest companies in the world and in 2023, together they spent something like $150 billion in capital expenditures. A big chunk of that went towards their efforts in artificial intelligence. Well, last year, in 2024, that's forecast to have grown to about $220 billion. 2024, that's forecasts have grown to about $220 billion. And then next year, 2025, or this year, over the next year, we expect that to grow to something like $260 billion. That's just massive amounts of money that's being invested. You take those three years together, that's something like $630 billion invested in three years by just four companies. It's really unprecedented, and what's on the minds of many investors today is when does that growth in spend begin to taper off? Because it can't continue forever, nor should it. And then the next phase of the artificial intelligence story, if history is any guide, will be what companies are able then to turn around and take some of those investments that have been made and turn that into profits. And so that's one of the things that we want to think about.

Ryan:

Unfortunately, some of that only gets known in hindsight, ultimately, when we think back to previous big innovations. Ultimately, when we think back to previous big innovations, I think about the internet Early in my career, the winners in the development and the build-out of the internet. We really had no idea until we had the benefit of hindsight. You think of some of the biggest companies in the world Amazon, the $2 trillion behemoth that it's grown to. Well, back when the internet was at its birth, this was an online bookseller. We think about companies like Google, which is now called Alphabet. We had no idea how Google was going to monetize some of its search results early on, but now it's a $2 plus trillion market cap. Think of a company like Apple. Apple's a $3.5 trillion market cap company. Well, apple's iPhone and their dominance of that space and that ecosystem. It just wouldn't have happened without the benefits of the internet and ultimately we think that something similar will happen down the road with artificial intelligence. Unfortunately, it's really difficult to say what those ultimate winners and losers will be when we think about artificial intelligence going forward. We'll pay close attention as the winners and losers emerge, but there are certain industries where we think you can already begin to see some of the benefits, some of the investments that have been made in artificial intelligence starting to pay off. One that I have been paying close attention to is the biotechnology industry. It's a business where it takes roughly a decade to develop a new drug, something like a 90% failure rate. Investment totals something like $2 billion to bring a new drug to market. It's a really difficult business and it's one that's ripe to be disrupted and improved the productivity and efficiency by innovations related to artificial intelligence.

Ryan:

For example, you may have seen or read about over the last couple of years, there's this machine learning program from an organization known as DeepMind. They developed this program called the AlphaFold project and there's been various iterations of that and ultimately it solved a multi-decade problem known as the protein folding problem. Ultimately, it's figuring out how amino acids form into three-dimensional shapes for proteins. Proteins are the building blocks of anything biological, and that is just a new set of information. Essentially, deepmind was able to, using machine learning, was able to predict the shape of almost all the proteins known to science, and so now researchers have this library that they didn't have before, where they can look at a disease and say I want to develop a cure for this targeted disease and I have a much better idea of some of the aspects of how it, how the biology, works within the human body, and that's a huge step forward. It doesn't. It doesn't ultimately create a cure. But what's? What's really exciting now, as you read about some of the uh, some of the ways that companies are implementing uh, artificial intelligence, is they're able to compound that knowledge by taking something like the, the alpha fold library of proteins, and they're able to utilize artificial intelligence to develop new molecules to, first off, understand the disease but then figure out how new molecules whether that's sort of the small molecule treatments of the past or antibody treatments or proteins how they'll then be able to bind to the cells causing disease. They're able to effectively sort through the various candidates that they'll introduce into drug trials to hopefully pick those that'll be most successful and they're also able to optimize for those that'll be most easily manufactured, and this is something that we did not have the capability of until very, very recently. And ultimately, I believe that this is one of the industries that we'll see over the next several years, introducing new cures, new disease, new ways to treat diseases that we didn't have before as a result of artificial intelligence. There'll be other industries as well that we think will be disrupted by the inclusion, the utilization of these tools related to AI.

Ryan:

Apart from linkages to AI, though, I look at biotech, by the way, over the next several years and it seems possible to me that you know one Thank you has been a lack of M&A deals and industry consolidation going forward under the new administration. We think is probably going to benefit those companies. And, of course, you look at the incoming administration. They've got Vivek Ramaswamy, who has the ear of the president. What did he do before? Well, he was a biotech entrepreneur, and I think maybe that goes to offset some of the other folks that have the presidents, the incoming presidents ear that may be a little bit more cautious about drug development. So the biotech industry is one that we're watching closely and we think may, over the next several years, benefit from the development of artificial intelligence. But there are other ways, other investment themes that we think that are maybe a little bit more off the radar, that are also going to benefit from what I call ripple effects of AI or some of the second order impacts of this new technology. It's always great to participate in things that are in the news that everyone's talking about, but I think there may be more opportunities in areas that are tied to that theme, but maybe not as directly as you would think.

Ryan:

And one area that is becoming more popular to talk about that we at First Trust have been very excited about for a long time is the related power demand that is growing because of artificial intelligence. Growth in the expected demand for electricity that sort of what we would expect between now and 2029, over the next five years is expected to triple what the expected growth rate was just a couple of years ago. In other words, when we look at 2022 estimates from FERC that's, the Federal Energy Regulatory Commission well, their five-year growth rate projected when they made those forecasts back in 2022 over this next five years, was about 3.6% growth. It doesn't seem like that much and, of course, if you look back over the last couple of decades, the growth in electricity demand has not been very strong. Things have gotten more efficient. There hasn't been a huge growth in demand for electricity. Well, they just released their 2024 estimates for that same five-year period and the increased expected growth rate went to about 12.5%. So from 3.6% to 12.5% and again, that may seem small, but it's triple the rate of growth that was expected just a few years ago, and I certainly think that we could see further demand growth expectations when you think about things like data centers, which AI data centers are a key vector of demand growth for electricity going forward, in addition to some of the others that we can talk about. But there was a report from Deloitte that was published near the end of last year that said that data center electricity demand growth is expected globally to top 1,000 terawatt hours by 2030. And that's more than double what the sort of consumption has been last year, and so that's a tremendous amount. To put that into context when we think about how much electricity the nation of Japan consumed last year or in 2023 is the most recent statistic we have. It was about 900 terawatt hours of power, so we're talking about more being consumed by data centers than all of the nation of Japan. It's a tremendous amount of demand for electricity growth. And, of course, that doesn't even speak to trends related to cryptocurrency mining, trends related to electric vehicles, trends related to the sort of onshoring of manufacturing and build out of new factories. We expect the demand for electricity to soar over the next several years as a result of some of these new vectors of demand.

Ryan:

Now, there's implications of that. I know who actually benefits if there's a lot more electricity demand. Well, first off, we think that power grids around the world are going to need massive amounts of investment. In fact, bloomberg estimates that investments in power grids between now and 2050 are going to be somewhere around between $14 trillion and $23 trillion, and I don't care if you're on the low end of that estimate. That is an enormous amount of capital spending that needs to be made in order to meet the demand growth for electricity over the next several decades. So, companies that are providing the electrical components, the poles, the wires, the transformers, the other parts that you need to modernize a power grid to make it so electrons flow in two different directions because, let's face it, you've also got new sources of power production that have come online, the largest of which is solar and second is wind. Both are intermittent. They're more distributed. So you need to modernize the power grid in light of all of these areas and, as a result of that, we think you're going to need to make massive investments, and those are some of the sorts of companies that may benefit.

Ryan:

Now, the second implication of that, in our opinion, is that power production has to also increase, so it's not just transporting power. You have to actually increase the amount of power that is being produced, and in order to do that, you have to actually increase the production from sources that haven't been planned. If we look, for example, at what has been planned over the next several years, the vast majority of new power production in the US over the next five years is expected to come from solar power. That's in order to meet mandates. A lot of the power producing companies they're happy to produce power via solar Wind comes in second. Natural gas comes in. Third, there's a massive amount of those that are expected to come online.

Ryan:

However, there's a problem with that, and the problem is that if you're running a data center, you need to have access to power 24 hours a day, seven days a week, 365 days a year. You can't have intermittency when it comes to running your data center, and so you can make massive investments in batteries. But that adds cost, and so one of the trends that we've seen more and more of, especially in the fourth quarter of this year, is deals that were announced by some of the technology companies related to nuclear power, and this is a big development, because not too long ago, nuclear power was not something that these companies were embracing. Constellation made an announcement that they were going to restart a reactor at Three Mile Island to provide nuclear power. But guess what? Even that deal, they estimate that you won't have it online, if everything goes well, until about 2028 at the earliest, and of course, that's three years from now. And so we think about the more immediate demand for power, because these companies are building out data centers as quickly as they can, and I'm not sure that even something like Constellation's selling of nuclear power by 2028 is going to satisfy some of that demand. But the reality is there's very few of those sorts of deals that are available to increase power production.

Ryan:

Now you see other deals announced in agreements that have been made between big technology companies and nuclear companies for things like small modular reactors, but, again, those are not commercially available yet and they probably won't be based on most estimates until at least 2030, and, in some cases, pushing back several years beyond that. And so, even if you're planning to adopt nuclear power using some of those small modular reactors, you're probably not going to have them in place in time for the investments that need to be made over the next five or so years. And so, if renewables are too intermittent to satisfy that demand, and nuclear, well, it takes too long and you really don't have the ability to scale it up quickly. Well, what does that leave? Well, you could use, perhaps, coal, but that's too dirty, or you could use natural gas. And one of the things that we expect over the next several years is you're going to hear more and more deals, like was announced recently by Meta and Entergy, where Meta says that they're going to build a $10 billion AI data center in Louisiana and, as a result of that, entergy is going to build three new natural gas turbine power plants to supply the power. Because, well, natural gas is cleaner than coal, it doesn't have the intermittency of wind and solar, and it's something that you can build and scale up quickly, unlike nuclear, and so, at least for the next several years, we think that some of those deals will be announced and with them will be long-term contracts, which will benefit some of the natural gas companies, especially the companies that are transporting natural gas. We also think that some of the utility companies related to this build-out of the power grid and power production as well, will benefit, because, of course, if you're dealing with a regulated utility, you and I end up bearing the cost of those investments. In other words, they're able to pass along in many cases the cost of the investments that they're making. They're actually allowed to charge a rate of return based on those capital investments, and that's good if you were a utility company. So that's another area where we think may benefit as sort of a second order effect in AI.

Ryan:

Another area that we're thinking about is this sort of theme that we saw emerge last year related to reshoring, in other words, bringing manufacturing from overseas back into the US, and we look at construction spending on things like data centers. Well, it's at an all-time high in the US Over the last three years. We're looking at a growth rate of about 44% per year we look at manufacturing or construction spending on manufacturing facilities related to computer and electronics. Well, that involves semiconductors, and if you've seen in the news some of the announcements made by big semiconductor companies that are building new plants, well, the rate of growth over the last three years for that has been about 82% per year. So massive, massive amounts of growth, a strong likelihood that some of these trends continue going forward.

Ryan:

And part of the reason is because, well, let's face it, the new administration probably has more protectionist policies that they're going to put in place At least they're talking about that. And of course, if you've got protectionist policies for imports, then there's an incentive to actually manufacture goods here. In the US there's rising geopolitical instability and unfortunately we see that growing by the day. That translates into higher costs for businesses and again, so that's a financial or economic incentive to build out some of their manufacturing capacity in the US. And then there's a strategic imperative for certain key industries, like semiconductors, to move manufacturing to the US. And on that last point, that's one of the big reasons that the CHIPS Act was passed in the US a couple of years ago. Of course, this is a big program that the federal government signed into law where they're going to award tens of billions of dollars to semiconductor companies.

Ryan:

Now, one of the things that I didn't see reported too much in the press in the fourth quarter that we noticed was that since election day last year so really since about November 15th there's been $30 plus billion uh conditional commitments for funding that were converted into final awards for the chips act. In other words, they were sort of uh, the commerce department said, yeah, you're probably going to get this money, but they, they finalized those deals uh in the last month and a half, um, because they want to make sure that they're in place and awarded uh before the next administration comes in. Now, setting aside whether it's good policy or not, we'll leave that to others to discuss I know the Trump administration. You know Donald Trump has said things like maybe tariffs are a better way to incentivize chip manufacturers to build out their manufacturing capacity in the US. Build out their manufacturing capacity in the US.

Ryan:

Setting that aside, this is the policy that's in place and it's $30 billion that's been awarded, but what's interesting is that's really only a small piece of what's going to be spent in the years ahead. For example, micron is a company that was given one of the largest awards over $6 billion that they were awarded. Recently. They've announced plans to spend as much as $100 billion over the next couple of decades in my hometown of Syracuse, new York, and building new semiconductor plants. Now it remains to be seen. Maybe it won't be $100 billion, maybe it will be. None of us really knows what the future holds, but that is their stated intention.

Ryan:

But here's another interesting part of that story. Micron in Syracuse made this announcement over two years ago and they have yet to break ground in those plants. In fact, they say that optimistically. The earliest that they can do that is not until November of 2025. So towards the end of this year, and who knows? That could always get pushed back once again. And why that's important to me is that's one of many companies where they're in similar situations, where some of the regulatory hurdles are in place that may actually start to get reduced in the year ahead under the new administration. But also it signals that there's a lot of money, that this spending that we've seen already is perhaps a drop in the bucket compared to what we're going to see as these large semiconductor companies continue to build out their manufacturing capacity in the US, and so this is one of these long-term secular themes that we think will benefit from the ripple effects of AI. And, of course, you say well, who benefits from building big manufacturing facilities? Well, some of those companies related to construction and engineering and providing the site work and providing the HVAC for those buildings, some of those small manufacturers should really benefit from some of the build-out in manufacturing capacity in the US, and so that's one of the areas that we're paying close attention to for 2025 and beyond, because we think this is off the beaten path. These are not themes that necessarily, people are thinking of first when they think of artificial intelligence, but we do think that these companies may benefit as we think about the potential for growth in the years ahead.

Ryan:

Now, there's no doubt to us that there will be plenty of surprises in 2025. That always happens. It's inevitable. We respond to those as we see them. We're prepared with a diversified portfolio and, fortunately, we're pleased to always work with financial professionals who have the best interests of their clients in mind and can help those clients through times that are uncertain, times that can sometimes be emotional. So, once again, we're looking forward to 2025. We're paying close attention to some of those themes that may benefit from some of these long-term secular trends that are happening in society. We're always looking for opportunities for growth in the midst of that, and I'm looking forward to 2025, especially because I have the opportunity to talk to some of the smartest people that I know, get their insights and hopefully share them with you. So once again, thank you for joining us on this episode of the First Trust ROI podcast. We look forward to speaking with you throughout 2025 and hope you have a wonderful and prosperous year. Take care, everyone. We'll see you next.

People on this episode

Head Shot

Ryan O. Issakainen, CFA | Senior Vice President | ETF Strategist

Host