Talking Markets Podcast (Artificial Intelligence) with Mike Lippert (Baron Capital)

0:00Hi everyone, Dan Cassidy here. Welcome back to the Talking Markets podcast series here on the UBS Market Moves podcast channel. We are coming to you today from Air 1285 broadcast studio here in New York. I've actually have Brian Contreras to my left joining us from the UBS studio steam to join me in a conversation on artificial intelligence and how this technology is continuing to have such an impact and we're very fortunate to have joining us once again from our partners at Baron Capital to my right. A Mike Lipper joining us from Baron Capital. He is head of Technology Research and a portfolio manager at Baron Capital where he focuses on innovative, high growth companies across industries. Mike had joined the firm as a research analyst and has 26 years of research experience. So with that, Mike, first off, great to have you back here in the new studio. I know you joined us about a year ago at our former site, so great to have you back with us.

1:02Thanks for having me. Look forward to the conversation. Absolutely. And Brian, good to be at the table with you as well. Thank you very much. And Brian, when you were putting the questions together, we were talking about what we wanted to focus our time with Mike on. I know the impact that AI has had recently to the software sector has been top of mind for many of our listeners and clients, right? Yep. And that's really where we want to get started in this conversation. So Michael, how exactly is generative AI challenging the SaaS business model? Yeah, when you say exactly, it's not quite an easy answer. Right. I was laughing when you asked the question because we've been thinking about this for a couple of years. I think we wrote a piece, it's probably two years ago now, it is software dead. And our answer was no, but that's not the right question because we only have to invest in what we believe are the software winners, not the software losers. And there will certainly be a lot of companies that will be impacted by the changes. It's not an easy answer. And as I joke before we started, we could talk about this for an hour.

2:04But there are major changes to software because of AI. I mean, the way software develops is being developed is changing. The way software is being used, what we expect of software, who the users are, whether they're going to be humans or agents that are using the software, and also the monetization of software where SaaS typically has been monetized per user, per seed. And now we're going to be shifting towards more consumption type models. These are major changes and they're real and they shouldn't be dismissed. On the first one of how software is developed, the development of software writing the code has honestly never ever been the issue. So yes, we can write software faster today. And there's now a lot more talk about whether companies will write their own software or the so-called startups can move more quickly. But I think having startups in the space that's always been around. But what we expect of software is a major difference. When you turned off your software at night, it didn't do anything for you.

3:04Now you actually expect software to do the work before the software, human how to do the work, now the software can do the work. And that is a major change in terms of the usage of software and the users of software. I don't think the world of the future will be so-called drop down menus. Humans will either talk to their software or type to their software conversationally. And oftentimes you will have an agent actually doing the work for you. So that's a real change. And of course, the monetization is a big change. How will we monetize? And if we don't need as many workers as we did before, companies that have business models based on seeds, there's a lot of concern about that. And Wall Street, in any time you have a model transition, it's always a challenging time, certainly for stocks. And so we're now seeing a lot of companies go to what you call hybrid models. So they have subscriptions or seeds. And they're also adding in consumption. So we do think this will be very, very disruptive to software. And we can please follow up a little touch on a couple of things. And so we're trying to think of what are the characteristics of a company that will allow them to thrive in the day of AI.

4:11Listen, the first thing is not detailed. Any company in technology today, if you are a legacy company or even if you are a startup, you need to act like a startup. You need to move fast. You need to invest in your business. You need to be willing to disrupt yourself. If not, you will certainly be passed by because I've not been doing this. As you said, 26 years, I've never seen technology innovation, evolution, developments, disruptions as fast as this today. So that's the first thing. The CEO, his or her management team, they need to really, really move very fast. I'd say the most important thing about software and what differentiated is, and I'm oversimplifying here. There's lots of different layers, but it's data. Why? AI cannot do anything without data. You need to train the model based on data. And of course, when you want an answer and you're prompting AI, you want to basically be feeding it with data. So companies that could either capture or generate differentiated data.

5:15The data system is more complex. They could add organizational layers to that data. That's what a company like Palantir calls ontology. But it is making sure that you're feeding the model with the right data in terms of the question. And so we think about that a lot. So for example, I think the cybersecurity vendors, which have gotten hurt recently, they are capturing different data. They are, in a sense, generating their own data. And it's very, very definitely very, very hard for AI to challenge that. I'll just make one more point and you could please follow up. You know, when they talk about system of record companies, which is more often in the application space than what you might call the system space. There's no doubt if you are just a system of record. And the data that you capture is not all that differentiated. It's not all that complex. It doesn't require lots of governance privacy. It doesn't have connections to all these other systems. You are definitely at risk. But if you are, you know, what I described, it's complex.

6:17And you could add value to the organization of data. I think in the world of AI, these systems of records are critical. Because if you don't have the data well organized, you're never going to get any value out of data. I'll stop there and please follow up. How can these companies restructure the business model? So they're not as impacted by the destruction we're seeing today? Yeah, again, for the software vendors, the so-called legacy ones. Again, the first of all, they do have to move really, really fast. I do think there were a lot of players that will be disrupted. We've been incredibly careful with our portfolios. We are not making, quote, unquote, a software bet. I call it being carefully, selectively, puttally, contrarian. So we're being contrarian in places. Again, we're really focused on companies that can capture more generate unique data. Certainly cybersecurity is a space that we've been investing more in. What you might call the data management space. Other system vendors that are going to benefit from all the movement of data. With AI company like Cloudflare, we're very, very careful in what you might call the application space.

7:22So again, you know, one of the ones that we own is a company called Samsara, which is trying to bring software to what you might think of as the industrial base. They have lots of business clients who are either in transportation. Construction is one of the biggest vendors. And they're basically trying to bring all of these different physical assets into the technology world and add AI. But you cannot capture data on any of those assets unless you put your system. So they have systems on trucks, buses, cars, and other, you know, assets such as bulldozers or even high value assets that are moving around the world. And when they talk about, you know, what the data they're going to capture on their conference call, they say this data is not available on the internet. So, you know, anthropic or open AI cannot train their models on such data. So these are just giving you an example, but these are the type of things that we're looking for in the companies that we think will thrive in the age of AI. You've already mentioned this specific aspect application of AI. So what are the true capabilities or use cases of AI agents with so much evolving in the space?

8:24Yeah, listen, it's hard to capture any given moment because we're seeing, you know, such incredible advances. You know, Johnson Wong literally talking about the recent GTC. Right. You know, what AI could do today versus what it could do a couple of years ago. You know, the first moment of AI was, of course, the chat GPT moment. But last October one, you know, anthropic released, you know, their latest cloud model. We're kind of in the agentic moment now. And the first use case that is really, you know, had a significant change in adoption of AI is, of course, code generation. And so I can't remember the last time I talked to a company that does not have literally every single one of their internal developers. You know, using one of these products, whether it's cloud code, codex from open AI, you know, cursor, for example, all these other products. Oftentimes, you're using multiple products. So I think that's the first use case that we've really, really seen. And the productivity improvements, you know, are anywhere from 30% productivity improvements to, you know, much, much higher numbers.

9:26It's hard to know exactly, you know, what it is on average because we hear lots of different data points, but it's very, very clear. I think this will be 2026 will be the first year that we see a lot of other, you know, business use cases where you have the adoption of AI. I mean, everybody on my team, so we're in investing. Of course, we generate models every single person on my own team. It is utilizing AI and just like any other firm, we're not only we're using it on data and information outside of the corporate firewall. Now we're trying to bring it inside the corporate firewall. And so I'm sure people at UBS are the same thing. Everyone's clamoring to use it. And so I think there'll be lots of use cases such as, you know, anyone in the financial industry using, you know, cloud code for Excel to help, you know, generate. Models. Again, in our industry, there's lots of transcripts out there so you can summarize transcripts with really information that's important. And I think every single business is going to start using that. We're seeing lots of very successful use cases and customer service where AI could effectively do the first level customer service.

10:28We're seeing use cases of AI and sales where they could look at, you know, your different prospects and gather information about them. And for a salesperson, basically, rank the prospects that they should contact and even say, oh, you know, you should contact this firm and, you know, Mike lippard, this is his role. And we know what he searched online and, you know, this is the what you should product that you should, you know, first. You know, try to try to sell to him. So these are just a few examples of use cases. So I still think AI will get better and better and better and better. I'm glad we go through the year. Now, as impressive Mike of a technology, this is have to face the reality that it's also quite capital intensive. You think about the investment required when it comes to chip production data centers. The energy to power this all might that translate to mag seven companies adding to their market share. Or will the playing field level out a bit might we see new competitors enter into this space? Yeah, I mean, before I address that, you know, I'm talking positively about AI, but I don't want to be polyionic.

11:30I mean, AI is not perfect. We'll still make mistakes today. It needs to have guardrails. AI is what's called probabilistic by having more data, more guardrails. You want to make what's called deterministic. Make sure it gets the right answer. So I don't want anyone to sit here and listen to think that we think AI is there. I mean, it's amazing advances where we're going to be in three years, five years, ten years. Who knows? Listen, when it comes to technology, most of my career technology was not capital intensive. Even the first phase of cloud computing where you still had to build data centers and you had to fill them with network equipment and certainly chips. It was just less capital intensive. There is no doubt that the world of AI is capital intensive. We are now in a phase of technology and will be honestly as far as I can see that will be very, very capital intensive. We're going to be able to build as Jensen Wong of NVIDIA says, you know, the factories to generate the so-called tokens that are put together to give you an AI answer. I don't think that will change. So yes, I do believe that AI is a scale game today, meaning those who have capital, those who have access to capital will have advantages.

12:38The Mag 7 today, most of them have the advantage of generating their own capital so you do not need to go to the capital markets. Just last week, you know, open AI, I went to the capital markets and raised over $100 billion. You know, there's rumors that anthropical public fix this year. So I do think access to capital will be a either competitive advantage or a competitive challenge for companies. At the same time, I do think there will be lots of disruption. So there were no easy answers here, which is what, you know, my job is a portfolio manager and instead of taking research of Baron, it's one thing to say, oh, all the hyperscalers or the Mag 7 guys are going to win or no, it's going to be all the disruptors. You have to look at it incredibly carefully because there's not one simple answer. And I think there are some of the Mag 7 that will thrive. And I do believe there are some of the Mag 7 that will absolutely be challenged by these AI disruptions. So to the extent that you could be transparent on this point, Mike, in terms of where to put investment dollars,

13:38a lot of avenues to explore could be overwhelming. Where are you seeing the most opportunity for AI investment at the moment, whether it be data centers, chip production, etc. Yeah, I mean, it's a little bit of everything to be honest. Like we're trying to run portfolios at our diversified. We just went on a research trip out to the West Coast and we talked to everybody. We talked to the application vendors. We talked to system software vendors. We talked to chip vendors. We talked to networking vendors, both public companies and private companies. And things are moving so fast that to predict exactly what the world is going to look like one, two or three years is very, very hard for everybody. So I do think any investor should be diversified. And that's what we're trying to be in our portfolio. So literally every area that you named, we are finding opportunities. So we certainly have investments in the chip area. And our main investments there are Nvidia Broadcom and TSMC, but they're not the only ones. You know, we certainly have some investments with the mag sevens and, you know, for example, Google and video who I mentioned Tesla with physical AI.

14:39We could talk about that. And in software, there are definitely places that we think the market is overcorrecting. And I would say, you know, site cybersecurity is one of those areas. I could go through a much longer list, but I'll just touch on a few things there with you. So there's a lot out there to be mindful of. I do want to ask, though, and we talked about this a bit last year. I thought it interesting, but bringing it to today, what are some use cases for AI that you feel might might be underutilized? That companies, individuals are not utilizing enough on the flip side. What use cases have been widely adopted and braced over the past year? Yeah. So the widely, most widely adopted, of course, is, is, is, is Codechan. I'd say the one behind that now is, you know, in the enterprise, not in our personal lives. Is customer service. I think AI penetrating, diffusing into the workforce is just a little bit slower. There's lots and lots of reasons. One of them is, of course, governance and privacy and control of your data. Companies, honestly, getting their data houses, what they call data houses in order to be able to take advantage of AI.

15:41And then I just think there's human inertia, right? Human beings like to, whatever supervise, ten other human beings and what happens when you're not supervising ten people, you're supervising two and you're supervising 20 AI agents. So these things will take a while to, in a sense, to permeate the fancy words that they use as diffused through the economy. And so we're, we're watching it very, very carefully. I mean, at, at the barren conference last November, one of the things that I talked about there, we're watching is effectively, what's the utility of AI? Right? What are the use cases that are providing value? Value could either be revenue generation value could be, you know, cost savings. And so we're looking at that very, very carefully and we're doing across every single industry. So we're talking about tech here today, but we're looking at it in healthcare. We're looking about our industrials. We're looking about what you might call physical AI, which might be, you know, robotaxies or physical robots or, you know, production where you don't need a human in the factory. And so I've just named it a few. So what we're, honestly, at our firm, we're trying to look all across the economy to see how AI is going to be impacting it.

16:42Not just, you know, whatever the tech companies that are in the limelight today. Well, I had the pleasure of attending the conference. Your team always does a terrific job. But I know Mike, we're coming up to time for the purposes of today's conversation. Any final thoughts, takeaways we'd like to leave our audience with? Yes, and the takeaway for the audience is, you honestly have to do the research, or if, you know, you're investing. I think it's very hard to do it on your own. So you want to obviously invest with someone who's doing a lot of work. I'm sitting here talking to you, but I got 20 people back in the office doing this work. And what I say to my team is, you know, the finance media, which I guess we're part of today, is, you know, at a different level. And it's been most of my career. We live in a world of, you know, ex podcasts of ours, you know, YouTube. You cannot believe everything you hear, everything you read. Even myself, you want to listen to me like I say to my team, everybody talks their own book. So understand what their book is, because that's where they're going to be. So, you know, you could, you know, watch a bunch of podcasts that some people are, I don't know, pro Microsoft and negative Microsoft.

17:45It depends on where their position is. And so we try to get past the slogans, the headlines, even when a company makes an announcement. It's not the announcement. It's like what's behind the announcement. There were a lot of announcements last year. None of them were deals. They were, you know, letters of intent frameworks. And so we try to really be, and I'm a former lawyer. And so I'm joking about it. Fact-based or evidence-based investors. What is real versus what is, you know, talked about? And I would mind everyone, you know, really think about that in the world that we live in today, trying to find out what's going on. It's real. So do your homework. Do your homework. Yeah. Well, Michael, let me get back to your team. You've been very generous with your time today. Thank you for once again, dropping by and the conversation will continue. We'll do it again. Thank you. Thanks for having me. Thank you for tuning in. Be sure to visit UBS.com slash studios to view the entire UBS studios suite of podcast channels along with their video offerings such as UBS trending. You can also follow us on Instagram for content highlights at UBS trending.

18:48UBS studios is part of the UBS Chief Investment Office within UBS Global Wealth Management. Visit UBS.com slash CIO to view the latest research. As a firm providing Wealth Management Services to clients, UBS Financial Services Inc offers investment advisory services in its capacity as an SEC registered investment advisor and brokerage services in its capacity as an SEC registered broker dealer. Investment advisory services and brokerage services are separate and distinct differ in material ways and are governed by different laws and separate arrangements. It is important that you understand the ways in which we conduct business and that you carefully read the agreements and disclosures that we provide about the products or services we offer. For more information, please review client relationship summary provided at UBS.com forward slash relationship summary. UBS Financial Services Inc is a subsidiary of UBS Group AG member FINRA SIPC.