Transcription
Welcome to Alternative Realities, a podcast about the people, strategies, and stories driving the world of alternative investments. I'm Aaron Mulvihill, and today we're kicking off part one of our crossover on AI and investing. This episode focuses on private markets.
My guest today is Stephanie Aliaga, Global Market Strategist on the Market Insights team at J.P. Morgan Asset Management. In addition to her time researching the global economy, Stephanie has spent extensive time researching and talking about artificial intelligence, providing clients and investors with timely insights and perspectives about both the public and the private markets.
On today's episode, we will discuss which sectors are most at risk from AI disruption, why the private markets have been home to many of the leading AI companies, and some of the innovations she is seeing across the AI space. It's a fascinating discussion, so let's jump right into it. Stephanie, can you give our audience some background on yourself, your career, and how you carved out this niche within AI?
Absolutely. Well, first off, thanks for having me. I, like yourself, am a Global Market Strategist on the Market Insights team. I started off as a Research Analyst. I was working really closely with Dr. David Kelly, and I was his, kind of like, macro research analyst, and that's my background. I was a macro person through and through. That was sort of my identity until a few months after (Unclear) launched. I was at a conference in D.C. hearing from a lot of different policymakers. And it was, it was not a big theme driving the markets just yet, but it was something that was murmuring, bubbling up in conversations. And I walked away from a range of, you know, meetings with different policymakers and think tanks, realizing we need to get a handle on what this technology is. If it's as transformative as some people think it's going to be, it's going to have real implications for what we think about the economy, about inflation, about productivity, and of course, markets. And that kind of started a long, kind of, research agenda around AI. I wrote a 20-page paper around what it can mean for productivity. It was the biggest lift I had done since college. This is pre ChatGPT really, like, I was not using it for research. Yeah, really wrote it. Like, now we're going to say it's like the old days. But anyway, fast forward to today. I'm, I'm a Market Strategist. A lot of my job is like traveling around the country, meeting with our clients, talking about, you know, trying to synthesize what is happening in the world of AI, what this means for investors, what this means for markets more broadly, and how do we, kind of, position portfolios and separate the signal from the noise. And it's been a lot of fun. You know, this is perhaps the most important theme of our generation, of our lifetime. And following it has definitely kept me busy, that's for sure, but also consistently learning.
Wonderful. And it's been great working with you on, on that theme. So there's been a lot of talk about disruption from AI, software selloff, and so on. What are the areas or sectors you think are most at risk, and how should investors approach that?
So this keyword disruption, it's like the, perhaps the word for this next chapter of AI. And it's kind of funny. You know, in this AI wave, we've gone through so many different cycles of sentiment around AI. Like, one moment we're talking about an AI bubble. Everything's overhyped, and another moment we're talking about AI reaching singularity, artificial general intelligence. And, you know, I actually think both conversations have been really productive. But first, on the, on the latter, I guess, and this, you know, markets really opening their eyes to this potential for AI to be a pretty disruptive force across industries. What does that mean? Well, really the catalyst for this has been the advancements in coding agents. You know, LLMs are phenomenal. They're great. They've gone a lot smarter. But we've now gotten, we made significant strides in useful AI, AI models that can actually do tasks that make up our workdays. And can actually do them autonomously and do them quite well. And so recent developments have really been this, kind of, like ChatGPT moment for our generative AI. And I think the question is, what does that mean for incumbents across the economy? You know, software, for instance, was about, kind of, the first real big culprit, maybe canary in the coal mine for this risk. And, you know, in some ways it's okay. Anyone can vibe code an application. Now applications used to be pretty hard to create, and where all the money was made, but now the barriers to entry have moved lower. And then also the people you're selling the services to, the applications, like maybe it won't be manned by a teams of tens and tens of hundreds of employees, but something far smaller than that. And so really what I think markets are in the process of, and I think we're in the early innings here, is a repricing of those expectations. If now all of the sudden the annual recurring revenue that you would typically expect from some of these companies to last for years and years, if you forecast out two or three years, you have more uncertainty to what that looks like. And if that's the case, then you should probably pay a lower valuation today for some of those companies. So we're in the midst of that process. That doesn't mean we won't need any of these software incumbents. It doesn't mean that any of these companies will fail, but it does mean we're entering a much more discerning phase where markets are going to need proof of either revenue acceleration, despite of AI, revenue acceleration with good AI products, and some proof that AI disruption, wherever it may come from, is, ultimately either helping or hurting some of these companies. And it's disruption balanced with a lot of excitement as well about where these agents and AI models can, can take us.
Can you give us some insight in I think we're talking this morning about recording this episode, and it'll be released in a couple of weeks. And you were saying, well, it's probably going to be outdated by then, moving so fast, but just high level, what's your take on where the models are today in terms of advancement, and is this a winner-take-all (Unclear)?
Yeah, it's a really important question. So I think we've used a variety of different benchmarks to gauge the intelligence of these models. And, you know, benchmarks have their challenges because if you train a model on a benchmark, they're really good at learning how to ace an exam. Right. But what I found to be really, really helpful is research done by this group called META, which basically takes a basket of all of these tasks that current human engineers, software developers, and so forth do. And they train or they, they, they give AI agents that call upon all these different models, those tasks. And what they find is the length of tasks that these AI models can autonomously complete has been doubling every seven months. We're in this like exponential progress here of AI models completing, executing more complex workflows. And when I first got my hands on this around the holidays, I spent some time and I was like, okay, I'm going to, I'm going to sort of vibe coding, you know? And I have avoided every coding class that they tried to throw at me. And coding at first, kind of like, threw me off as, I don't know how to code. How am I going to be able to manage these systems? You don't need to. Yeah. And once I started playing with these things, I understood why everyone in the tech community was freaking out. Suddenly, the barriers to entry for creating things has collapsed in so many ways. And so what does that mean for what we can create, the insights, the depth of research analysis that we can do on our team, the products, the materials, the volume of work and the quality of that work that we can churn out. And we're still in such early innings of figuring that out. I think in every industry it's going to be a little different. On one hand, you have, okay, you take all the operational grunt work, and if it's something you can write in a manual, it's something an AI agent can automate. Those are real productivity gains. And then I think the big question going forward is going to be, what do we see in terms of completely new products, in terms of value-add quality improvements? What comes out of our imagination now that we have just a much greater, more diverse, set of tools that we can use?
Yeah. And what are some of the more interesting or exciting areas that you're seeing?
In that, I mean, I think like, software is obviously a very clear first place of this because now anyone can kind of vibe code an application. Yeah. And so whatever application you want to make, anyone can do them. The, most of these applications, by your imagination. Right now, most of these applications aren't that good. It's like there is a range here of quality, and there is a real value-add that an incumbent can bring. Because they've been doing this business for so long, they have the data. But I think the question is, is the moat in the data that they've garnered from, you know, their incumbency, or is the moat in asking questions and having this innovative spirit? And that's, I think, one of the big questions. But it actually reminds me, one of your earlier questions was, is this going to be winner-take-all? And what models are people using? You know, there's this race at the frontier of models that we've been tracking for the last few years. And I think what we've learned is a few things. You know, one, there is still a great importance to have the leading frontier model, but that race changes week by week, right? What's the best in class? Some weeks it's Gemini, some weeks it's Claude, some weeks it's ChatGPT. They're the ones that get the headlines, and then everybody jumps on to use those models. Exactly. You know, and what's actually pretty remarkable is we thought there'd be a lot more stickiness. And consumers, you know, you've been using (Unclear) for two years, so you're never going to go, you know, switch gears and completely change operating system. It turns out new people are getting more comfortable with trying a variety of these different tools, which is good from the worker's perspective. Yeah. But anyway, I think what this ultimately is going to look like is you may have these different models pick different segments to really specialize in. There's one that's going to be your consumer-facing AI, another that is focused far more enterprise on coding, on programming, maybe another that's, you know, phenomenal creating music or creating videos. Another one that's phenomenal at helping train robots. And that's the differentiation that we're seeing. So there's room for many to succeed. But of course, the visibility that we have right now is limited as to who's going to ultimately be the, the winner, take us to AGI or whatever may be.
Yeah. And as this race is playing out, if we kind of switch the economics of it and the impact on markets. Now the biggest question that gets asked by, by investors is, is AI a bubble, particularly looking at the private or the public markets valuations that we're seeing. What's your take on that?
Absolutely. And I say that conversation goes to public and private markets. I think in public markets, at least, we have a lot more data to analyze and historical data. And I will say, you know, it's, it's really hard to make the bubble argument, especially today. You know, valuations are elevated, but they don't look outright frothy, or bubbly, I should say. Last year, the market for PE only grew 3% despite an 18% gain in the markets this year. Markets have been sideways so far. The PEG ratio, which is essentially, you know, the price that you're willing to pay for a long-run earnings growth. That is half of what it was roughly then, relative to the dot-com era. Because they're generating real earnings. Because they're generating real earnings. So that's the key thing here. The other thing to keep in mind here is that every bubble that we've had historically has been heavily reliant on external capital. Leverage has been ultimately the culprit of every bubble that creates that burst. But what's different in today's environment is, of course, it's much more cash. Finance is informed by robust demand. That doesn't mean that markets aren't vulnerable here. I think we may not be levered to debt, but we're levered to expectations. And because so much is riding on—I am in investors portfolios in broad markets. Any change in the earnings expectations here and the timeline that these data centers are going to get online or to get power onboard for them, any change in how expensive this is going to be, because memory prices have surged and capacity sold out for 2026. Any of these changes in the math of how this rollout comes out can have reverberations across portfolios. So I think the bubble conversation has been really productive in the sense that it is really identified for investors, the risk in portfolios from concentration of anything that's wrong here. It's not about leverage. It's really about ensuring you have some balance to weather some of these surprises that are undoubtedly going to come in the world of AI.
You mentioned private companies are not immune from this kind of valuation questions, bubble questions. And a lot of the leading companies today within AI, the model builders and so on, are still private. Why do you think these companies are staying private? They've not IPOed yet. We might see IPOs this year. But what's your take on, on, on that?
Yeah. It's such a unique feature of today's AI wave that the AI-native companies that are growing at rates you've never seen before, and or growing to valuations you've really never seen before, are growing in private markets. And it goes back to the, I guess, the philosophy of why do companies go public at all? And I think, you know, one of the big draws has obviously been the, you know, ability to access, you know, far greater kind of capital raises. And in today's day and age, these companies don't really need to go to public markets to access capital. You can reach $380 billion, as was just reached recently, by Anthropic, or $500 billion from OpenAI in private markets. And so if you don't need to go shopping in public markets for capital, you know, okay, you might want to stay private. And what are the other reasons why you might want to stay private? Well, what are these companies trying to do? They are in this race for dominance in AI, and it's actually quite attractive to have to keep some cards close to the chest. And you don't have that luxury as a public market company as much, particularly when it comes to strategic thinking, long-term planning, building out data centers for five, ten years into the future because you see where demand is going. So that's a big part of it, too. Now, these companies may go public. They may also not. And this year or next year, I think when and if they ultimately do, it'll be a huge event for the markets. But the key thing that I, I try to remind investors about this whole phenomenon is that let's say you take those three largest private market companies. If they were public today, they would be in at least the top 15 companies in the S&P 500. And they're not even public yet. So I think the question for investors is, okay, if there's a lot of growth in the pipeline around AI, and it's not just about the model labs, these behemoths, but all of the other startups that are specialized in AI robotics, AI applications across different industries. A lot of that growth is happening outside of the purview of public markets. And so you need to consider that in your portfolios when you're shopping in small-cap public equities. And then if you have access to some of these companies or to investment funds that are vehicles that can invest in these companies, I mean, you might want to think about, you know, that growth exposure, because when we put together a Long-Term Capital Market Assumptions, private equity stands out across all of our other major asset classes as most likely to achieve, pretty strong, elevated long-run forward expected returns.
Indeed. Yeah. I think if a company like Nvidia, for example, that IPO fairly early in its in its timeline, you know, when it was starting to be interesting but was more of a gaming company and then yeah, an investor is exposed to that. They've enjoyed very impressive growth. But you know, today these companies just are not accessible at all in the public markets. So what does that mean for, for portfolios. And how do investors either lean into AI or diversify away from that. If they're trying to get, get away from that, that theme and the volatility is coming with it.
Yeah. This requires a really, I think, a really careful look across the AI value chain. Because in my client conversations this year, I do feel a lot more anxiety around the concentration around these AI companies, around the risk of disruption. And I think the question is, okay, what can I have in my portfolio to help weather some of these risks? And I think to that end, we look to things like fixed income. You can look to things like infrastructure, right. An asset class that, it can help play defense in portfolios because you have stable income generation, but is also helping you on both sides on, on the offense side because it's exposed to all of this, buildout when it comes to AI. Right. There, is on the receiving end of all this AI CapEx. And then beyond that, I think investors should be prepared for the downside, but also the upside. There was this viral piece that (Unclear) had, had headwinds around. Something big is happening. And I think really the reason why this resonated so much is it helped make clear for a large subset of people, the upside risk, of these tools are just getting so good so quickly. We need to get a handle of what this looks like. And so there's a real opportunity cost for investors to just kind of sit out of this race. But I think the question is, how are you invested? How are you distributing your risk? AI technologies may be transformative. That can have really disruptive implications for leadership across industries and sectors. So how do I think about this in terms of putting together a portfolio? You obviously have your U.S. equity sleeve. And chances are that sleeve has grown pretty meaningfully after the last three years of AI-fueled blockbuster gains. We want to think about having some diversification there, right? Reducing your concentration risk, leaning into other parts of the market that are benefiting from the you know, all this AI CapEx. Beyond that, look, internationally, the AI value chain and particularly even the hardware ecosystem is increasingly global and interconnected. There's opportunities there in alternatives in private markets. A phenomenal way to help, diversify portfolios, right. Lower correlations to the S&P 500, depending on what asset class you're looking at, but also different nuanced ways to gain access and exposure to these themes. We've talked about the opportunities in some of these smaller growing, AI-native companies and some of the bigger ones, too, but also infrastructure, real estate, trans...transport. So it's really, you know, we go back to the basics here and it's the boring takeaway. But diversification we think is really, really key. And being mindful of how you're exposing yourself to the upside opportunity to I just as much as the downside.
Wrapping up with a lighter question, I know you're a big user of AI in your daily life. How are you using it today? And how would you recommend somebody who's maybe at an earlier stage of getting into AI and incorporating it in their life?
So my favorite questions, and one that I always include in my presentations around AI, because I think it is so key. This technology is the future. However you slice and dice it. All of us owe it to ourselves, to our employers, to our future prospects to get on the learning curve. Get on the learning curve, lean into it and be creative. You know, I don't think AI is going to automate creativity. I think it really empowers it. If you are don't know where to start, sit down and ask AI what is so phenomenal about this tool? Unlike anything that's come behind it, is how low the barriers to entry are in terms of skills. The first thing I did was talk to it all about the work that I do, the goals that I have, some projects that I've been wanting to dive into, but I just haven't had the time. I haven't had the resources. And now all of a sudden, I find myself chief orchestrator for all of these different workflows. I've, I've coded my own agents, I'm creating my own, like, dashboards to help me be smarter, more agile. It's helping me think more critically about what I'm writing. And I'm still barely grazing the surface. So I think the first thing that we can all do to ourselves is really, like, carve out time to spend with these models. The other thing I'll say, use a premium model. The gap between what you're getting at the free tier versus what is offered at the frontier is increasingly wide and significant. And if you don't want to pay the money, just think about what you're spending on streaming platforms, Netflix or whatever it may be. Really. You know, what's more productive? Let's keep evolving. You, you know, be smarter in today's day and age. So that's what I'll say. And then the other thing is, don't be afraid of the term coding. You know that you don't actually need to know a lick about code to vibe code. Hence the, you know, the great term that Andrej Karpathy created. So that's what I would say. And then finally, you know, the other thing around anxiety that I'm hearing in my conversations this year, is anxiety that AI is going to come for your job. And I think the more productive way to approach that topic is not to just fear AI is going to come and automate your job description. Ask yourself what you can do to outgrow it. And I'll leave you with that.
Wonderful. Thank you very much, Stephanie, for coming on Alternative Realities. That was Stephanie Aliaga, Global Market Strategist at J.P. Morgan Asset Management. To all our viewers and listeners, thank you very much for listening to the Alternative Reality podcast. Don't miss part two of this discussion where Stephanie will join Meera Pandit on Insights Now to discuss AI in public markets. You can download our recently released first Quarter 2026 Guide to Alternatives on the J.P. Morgan Asset Management website at jpmorgan.com/gta. And if you've not already done so, please subscribe to Alternative Realities on your favorite podcasting platform.