AI: The great disruptor

Welcome to the latest episode of the Astute Investor, an Australian podcast that gives you access to some of the world's leading minds in wealth management. Hi and welcome to the Astute Investor. I'm Martin Randall, Head of Private Markers at LGBT Wealth Management in Australia. Today, I'm joined in the studio by Isaac Kim and Michael Romano from the Light Speed Venture Partners. Light Speed is a leading multi-stage venture capital firm founded on a core belief that technology would permeate every sector of the economy. The firm has been investing in enterprise and infrastructure companies for over 25 years and in AI specifically since 2012, deploying over $7 billion US dollars across more than 155 AI investments. Today, we'll be taking a look at the big picture on artificial intelligence, framing how it's changing the landscape today and into the future. We'll break down the various layers of the AI stack, discuss AI's impact on traditional software businesses and explore the societal and environmental impacts of increasing AI adoption. Please listen to the disclaimer at the end of the podcast and keep in mind that this is not nor is it designed to be advice. The podcast is of a general nature and any statements by the speakers are their own. Welcome. It's so great to have you here today. Before we dive into the ever-evolving world of AI, can you maybe give it a background in your history and the light speed journey on what your role is today? Sure. Happy start. My name is Isaac Kim and I'm a partner here at Light Speed. I joined the firm a couple years back. Going way back grew up in Seattle, son of a couple of immigrants from Korea. They came to the States and didn't have very much and ended up being small business owners. Which was helpful for me as I thought through my career and what I wanted to do over time and I always took the path that was a little bit less fraught in and started insulting right out of college. But jumped into tech private equity at a small firm called Golden Gate Capital in San Francisco. Then in 2015, a little known hedge fund called Elliott Management Cult. And they asked whether I would come to join and build out their West Coast office and build out their tech private equity strategy from scratch. Which we did over about an a half year period. And then I joined Light Speed as I said two years ago to build out our growth private equity practice. So that's me. And then Michael Romano, the chief business officer at Light Speed, been there for 13 years and like Isaac also, yes, son of some Italian American immigrants, both also founders and always wanted to work in venture. At Light Speed, I really help architect our global LP base and we want that to reflect also our values as a global firm. The great founders can come from a lot of different and diverse places, whether it's here in Australia or back in Silicon Valley, great founders are increasingly popping up in more diverse places. I also run our co investment practice. So we built a really strong partnership working with investors to access, identify and lead a great investment opportunities. And then increasingly capital markets is a new kind of area in V.C. And so how do we consistently drive liquidity out of the asset class and secondary is actually been a really big part of that. So both IPO, M&A, and now the secondary market's been quite robust for sure. Absolutely. Thanks guys. So let's start with a big picture. Light Speed has been around for a very long time. You've both been investing for a long time. You've lived through various cycles and waves from a technology perspective. So I can have a look at what AI is doing today. Maybe comment and give your perspectives on the velocity of change we're seeing across the landscape and what you think AI could evolve into into the future. This is more consequential than all the prior disruptions and waves that have occurred. AI is fundamentally changing the way that we do business across the board. It is not just focused on where software automation has been, which globally is about a trillion dollar business today. It's going after and thinking through how we do work across the labor spend as well, which is globally $58 trillion. And one space that we focus on a lot is the business services area, which is knowledge workers in that space. And that's a 13 trillion dollar business. So AI is much broader than just a software automation play like we've seen in the past. And then the velocity at which AI is changing things is faster than we've ever seen. It is building businesses that are scaling quite rapidly. You're seeing businesses that are growing to $100 million in revenue in just a couple of years when prior iterations of software, it took typically a decade or so to get to that scale. And what that does then is it can permeate different parts of the economy much faster than any other innovation cycle we've seen. So it is, as I said, a consequential change in something that we track very closely. Yeah, from I think the internet cloud mobile to now AI, we've seen a lot of different innovation and the hype cycles that come with it. Well, we know a lot of values going to get created. One of the things that light speed we spend a lot of time thinking about is in what sequence. And how do you underwrite and understand, you know, the differentiation and the defensibility of some of these companies versus a lot of the big headlines and a lot of the promises and platitudes that AI can kind of provide. And so it's not an easy job. And it's one that we believe you really need to be more specialist in whether your computer scientist, your engineer really not just understanding what problem that founder is building, but exactly the architecture and engineering decisions that she made to build that. And so we have kind of combined deal teams across a lot of different stages sectors and geographies to find those great founders wherever they exist and earn the right to be their partner as they built their business. Indeed, he talks about specialisms there. Let's dive into some of your comments historically on AI. You've noted that it's critical to own the foundational layer, so foundation models. Can you unpack that and where you see those going because they all seem to be diverging into their own separate swim lanes rather than this all takes all mentality. How do you think about that? And then from a light speed perspective, how does that help inform where you invest beyond that foundational layer? Yeah, well, each of the models are at least the the large global models, the frontier models. They're trying to do three different things that wants one their research shops, right? And they're trying to get as close to human level intelligence or beyond as quickly as they can. And I would say that that space and that research and where it's getting to, we don't know yet whether that's winner take all or not. I think the first company that gets to, you know, AGI or super intelligence, I think they will have an advantage versus the rest. That's the research side of the house. Then there's the intelligence model layer that is fueling the AI economy today that is behind the applications that are out in the market. And that I do think there can be multiple winners and it is, you know, there will be open source, there will be close source, there will be enterprise, there will be consumer based models, there will be smaller models for different use cases. And there's, you know, difference between inference and other areas, right? So that piece of it, I do agree that it's an area that we will see different winners and there was different use cases. So different models can play out well. Then there's a third piece, which is more recent, which is the model layer going up the stack and into the application layer as well. There have been announcements around certain models going into cyber security and into financial services into the legal world, what have you. And that's a newer trend that we're seeing and I think that goes to the power of the models. And this is something we saw at light speed when there was the chat GPT moment. It wasn't the first time that we thought about AI. We have been investing in AI since 2012. We've made over 155 investments in AI. And so when we saw that come about, I think there was a lot of debate in venture circles and investment circles that gosh, the model layers kind of feel like they will be commoditized. And they're just part of the infrastructure layer really that's going to support other companies coming up kind of like what the cloud providers, you know, became to to to rest of sass world. And we had a firm view that that wasn't going to be the case. It's why we invested in a number of the models very early on. And that's come to be proven out as we've seen how quickly some of these model companies are growing. They're growing at 10X, year on year on year, often. And that's something that at the scale that they're at today, we haven't seen before. So just coming back on that. So how do you think about let's say the traditional mode of of a foundational model, you know, so why anthropic relative. example. Because it's a different mindset relative to a traditional industry or sector, for example. Well, I think the first thing that comes down to people and what you've seen with the types of aqua hires that have been occurring or just, you know, Zuckerberg out there and hiring this AI research talent for really, really, and godly sums of money that, you know, normal people can't really fathom. It's because that research element is so important and that people element is so important. And so I think we at light speed always look at the founders and think about the founders first and that founding team and it's not just that founding team, but who can they go and recruit, who can they go retain, who can they go and really motivate to join a mission. Right. And so that's I think the first thing. And then the second is you always have to think about, you know, obviously the technology and what you're building and the level of intelligence that we're building into these models and how they are occurring. But that's been somewhat a leapfrogging game, right? You see certain models come out and then, you know, a few months later, you see another model, it's leapfrogged, et cetera, et cetera. I think that's just the competition there breeds better outcomes for enterprises and better outcomes for consumers over time. But that does take a lot of research and energy from the companies. And I think for us, we just are focused on which team, which founding team and therefore which company can win out in that game. And then the last piece is, you know, these models do take a lot of capital. And so you have to understand the capital base that they're going after and the business model too. One of the things that got us excited about Anthropic was not only just the founding team, not only the vision that they had, not only the mission that they had to be focused on, you know, being safe and secure and reliable. But also the focus on enterprise because we think the enterprise business model is one that is a better business model overall than just paying subscriptions on consumer. If you're a consumer and you have two different consumers, one is using a million tokens and others using 10, but they're both paying 20 bucks a month. That's a hard business model to get behind over time. Whereas enterprise is much more focused on a business model that we think can scale. It's easier to switch as a consumer. That's right. We wanted just a little bit because I've been guilty in the past of focusing on specific applications or specific foundation models. Can you break down what you referred to with regards to foundational layers, infrastructure, operational layers, application layers, so the stack, so to speak? Yeah. So where it starts from is obviously the hardware layer of the chips. And that's where you've seen a lot of the public markets be able to invest in Nvidia and other areas there and to really see the early innings of growth and investment appreciation out of it. The picks and shovels around that have to do with the data centers and the energy that needs to to be created over time and all the businesses around that piece of it. You can go back even to the commodities that drive those, that infrastructure layer at the hard asset side of things. Then you have the models and that's the intelligence layers we talked about before. All the applications need to sit on top of a model and so that's the base layer. And then beyond that, you have infrastructure AI and you have application AI, native third-party businesses. That's across a lot of different segments of the market and can permeate, as I said before, not only the very software layers but also the human capital layers that are out there. Yeah, those are the basic stack as you go up. I think what's unusual is that so many of these previous technology waves were really driven around creating different tooling to help someone do their job better or faster or more productive. And what we're seeing now is just the work itself, the intelligence itself. And so why is that important and how does that manifest? We're seeing a significantly larger average contract values for these businesses that have reached size and scale. That's why in the old days, it would take anywhere from seven to ten years to reach a hundred million of ARR. We're seeing that in some cases 12 to 24 months. And so looking at any period of new technology transformation, we see a lot of speculation. There's always R&D spend. There's a very famous MIT study that circulated that said that hey, 95% of these AI pilots, they don't actually end up in people being customers. But that's actually not too dissimilar to what we saw in the mobile or the cloud. Whenever there's a really disruptive new technology, people have speculative R&D spend. They want to try and test new products. But what hasn't been talked about is the 5% of those companies that have found that product market fit. They're fundamentally transforming the way people work. And the amount of money you can charge for those sorts of tools, they're just not like anything we've ever seen in 25 years of investing in enterprise. Yeah, on that point, Roman was made. There's one of our growth stage partners, AI partners, that mentioned this to me the other day. He has a friend that works at one of the global hedge funds. So financial analyst. And you know, financial analyst, you pay $50,000 a year for a Bloomberg terminal, right? And that's something that they live in every single day. Well, that same financial analyst right now is run rating and a million and a half dollars of spend on one of the major models. That's orders of magnitude different. And why are they spending that? Because they see benefit of it. They see the benefit of the level of research they can do, the amount of research they can do, how quickly they can figure out different trends. And so to Roman's point, I think it's opening up new avenues and new areas of growth. It relates to the actual spend in the market too. Let's dive into that a little bit because we can talk about pure play AI companies across the stack, so to speak. But you've just touched on a financial services example there. We were talking about audio earlier, you've talked about accounting, the AI, sort of renaissance, as you put it, let's touch on that because as technology historically has become pervasive with every industry, this is likely to follow suit and what it is. You're seeing in the AI ecosystem as well. How do you frame that opportunity set now? We're kind of starting to get a feel for what the technology looks like and we move it out into the broader economy. Yeah, over the last 20, 30 years enterprise software has gotten to permeate roughly a trillion dollar annual market in revenue. As I said before, I think the business services space, so your audit firms, your legal firms, your accounting firms, your insurance firms, what have you. That market is roughly 13 trillion in size annually and that piece of the market has not been automated with software yet. But what we're seeing in these companies is even if the model layers didn't improve from today, a lot of those workflows can be automated with AI. And in a manner that is not just third party, AI applications being sold into those businesses, there's a reason why those businesses have not adopted software across their entire workflow. It's because it's hard, it's bespoke, it's specialized. Often you have folks that are high-end knowledge workers that you need to sign up on things and audit needs to be signed up by CPA, insurance broker needs to have a license to do that. But if you can go and own those firms, acquire those firms, and you can attract the right AI engineering talent, folks that understand not only how to build out a native AI platform, but how do you go and partner with those folks that are the actual services workers? And how do you understand their entire workflow and what areas they really don't like doing, right? They don't typically like gathering data from their clients and cleansing that data and putting into the right forms and answering all these questions that they have. Well, AI can do that very well. And we're seeing in some instances 25 to 75% automation. But what's really important to us when we think about investing in these businesses is not just taking that automation so that you can drive margin improvement because that's not super interesting. You're not transforming anything when you're doing that. You're just driving really cost savings overall. What gets us really excited is can you increase the automation layer? Yes, you'll get some margin improvement there. But can you open up the capacity of those services workers to do more with their clients? Can you open up that capacity for them to sell more to their clients and cross sell other services into those clients? And therefore you improve the growth rate of these industries from where they are today. And then you took away all their root tasks. So aren't they going to be happier? Aren't they going to be happier that you increase revenue and growth so that their equity in the business of the own is worth more? And then you look at the end client and then customer of those businesses. You all of a sudden have services that are faster and better than they were before. That's the real unlock. If you If you can do those four things, we think you can transform entire industries that today haven't been serviced well with software, and that's really the unlock of different areas that we see with AI. We had a great case last week, it was an AI native company serving the Carey Co-System, which is heavily administrative, and so you've got all of these care workers that just want to do the human side, which is what they're paid to do, but they're spending 80% of their time doing the admin, and just being able to someone in the ecosystem bringing that together to say, "I know you don't want to do this, I know what you need to do, but this is how you can do it, and we can help, and you can focus on the human side." There's so many examples I imagine. Yeah, we have one in that space in hospitals, right, with a company called a bridge, and Shiv, the founder there, he's a doctor by trade. His family, he comes from family doctors, his father was a doctor as well, and he remembers when his father had to retire, not because he didn't love servicing his patients, and not because he had to for any other reason, but for he needed to spend two or three hours a night typing in his notes into the EMR system, and that is something that doctors, it's just, it's just try some crazy, right, it's not what they want to do. They want to spend time with patients, and they want to help heal their patients, and serve their patients, and so Shiv said, "Look, there's got to be a better way." Instead of typing notes while you're in the actual patient setting, you can actually focus on the patient if you have ambient listening in the background with AI, taking the notes for you, right, and so it completely unlocks, I think, satisfaction for the doctors, it's better for the patients, you actually have eye contact with your doctor instead of them looking staring at the screen and typing away. You know RSA with the right to answer your question. That's right. And you can't even read that anyway, the doctor's being around, it's terrible. And so that's the type of fuse case that we see a lot, and occurring a lot, that's really interesting. What's incredible about that example, too, is, traditionally, those markets have been terrible to sell traditional software, too, right, a lot of heavily regulated industries that are almost like the last adopters of technology, and interestingly enough, from healthcare to legal to even financial services, we're really seeing strong adoption. In fundamental new use cases, just things that you weren't able to do kind of previously. And because of that, there is a compounding network effect that often happens, where if all of a sudden a bridge can do transcription and capture this kind of critical information more accurately, well, then it also has a right to spin up a different workflow and to auto send you to a specialist, maybe even give you a prompt for, hey, you didn't actually ask the patient about this question. And so you start to see a lot of kind of widespread adoption and love that is a real reason why the company continues to grow so fast. So let's touch on the traditional software side, because this is amazing, but traditional software seems to have been challenged quite heavily in the public market, and it's what everyone's talking about the moment. Are you talking about the sassacre? The sass apocalypse, I think it's referred to you down here, I don't know, there's probably different names for it, but AI tools, obviously, they're cutting into traditional software notes. Isaac, you've been investing in software from a PE perspective, a golden gate on to Elliot and beyond. How do you think about the risks to traditional software companies, and how do you separate what's going to be good and bad and the ugly potentially in those scenarios, particularly given AI can also be a tailwind for certain segments, depending on that selectivity as you put it so nicely. Yeah, I'll be a little careful here because I have a lot of friends that are CEOs of sass businesses. I also have a lot of friends that are technology private equity investors and partners, and I still want to be invited to dinner with them. But what I will say is, I think the public markets, it's a broad, you know, overreaction overall, right? Because it is very idiosyncratic to specific segments of the software market, as well as the specific companies and whether they can adapt or not. I think the smartest investors and software and the most adaptive and smartest sass CEOs will figure it out. And so I don't think it's a broad swath like software is in trouble overall. Having said that, there are segments of the market that have been disrupted already and will continue to be disrupted. What are those segments? So because of the proliferation of coding agents and what's occurring there, and you know, we looked at a test automation business a year and a half ago and brought Guru, one of our early stage partners, he actually leads our early stage practice along with us to take a look at that business. And it was really interesting because he had told me after the meeting, he said, "This is a really interesting business. It will grow faster in the next three to four years because there's more code being developed and it's levered to the code development, both agents as well as humans." But at some point in the next, call it three, four, five years, this business probably will have a terminal value zero. And I was shocked by that, right? I was shocked by that. And how did he know this? Well, Guru also co-leads our investment in Anthropic. And even before Cloud Code was launched, this was a year and a half ago, Cloud Code got launched last year. He knew what was happening at the coding agent layer. And he knew where it was headed. It was headed to better and better coding agents developing better and better code. And also an infrastructure layer around those coding agents that would do a lot of development tooling. And so you take development tooling as a market to human engineers. And unless they're really able to pivot to AI agents and how to help them do their job better, I think that's a market that has been disrupted. There are other markets such as sales and marketing tooling that sit on top of the systems of record, right? It's a lot easier and faster now to build agents and apps that sit on top of other systems of record. You know, the marginal cost of developing software is going down very rapidly. And so I'll give another example there, one of my friends that works at a startup. It's about a hundred million dollar revenue business. He saw one of them in the tool because I don't want people to go out there and sell on this specific tool. But he saw sales tool that sits on top of their CRM and he said, gosh, why do we spend a couple hundred thousand dollars on that tool? And it's not perfect for what we need. It gets us about 60% there. He's a business person, so he's a chief business officer of this business. He's not a technical person, but he went over the weekend and used cloud code to vibe code that tool for his business. And they ripped out the third party tool and they use his now. So I think if you're a thin workflow layer tooling business, I think you're in trouble already as well. But then there are the horizontal system of record businesses, right? The CRMs, the human capital management software platforms, the ERPs, and you have the vertical market systems of record as well. I think those will be imminently harder to disrupt over time. And I think those businesses have time to self disrupt, but the path to self disruption is really hard because history somewhat repeats itself. If you look back in the day at the ITSM market, this is IT service management ticketing for IT, essentially. There was a large industry around that. HP had to play their IBM had to play their CA had to play their BMC had to play their BMC was an investment that I was a part of when I was at Golden Gate. There was a small little business called ServiceNow when we made our investment in BMC and I think it was roughly 30 million of revenue at the time, venture back, angel back. And we said, eh, it's fine, it'll never come up to the enterprise. No enterprise would would would implement, you know, that sort of. Well scroll forward a couple of years from when we made the investment and ServiceNow started to show up in our enterprise accounts, not as the platform, not as the ITSM or CMDB platform, but as a point solution that we had not created or we had not spent a lot of time on. And what that did was, as we call this adventure a wedge product, they wedged into the enterprise with that product. And what that did was, it allowed them access to that enterprise. And then they went to that enterprise and said, what's the next thing you want to spill? So they built the second wedge product and they built the third and so on and so forth. All the while in the background, they were building the ITSM and CMDB platform. And you scroll forward to today, ServiceNow disrupted the entire industry that way with that wedge product and building the backend platform, you know, BMC fortunately enough was able to innovate and is still standing, but the rest of the market isn't. I think that's what's happening with the system of records today as well. And so what do the system of records need to do? They need to skunk work a project with their best teams and products and build an AI native platform, not just the agents on top. Because if they just build agents on top, the third party AI companies will be able to do that as well faster than most likely. Number two, they need to go disrupt their own business models, which is hard. You know, it's mainly seed based business models. That's not where the future is necessarily going. And so you have to think really long and hard how you're going to change your business models. You do that. Number three, I think you need to think about how you're going to change. your entire business, not just R&D, not just product. But how are you going to change the rest of your business to help pay for that innovation? Because ultimately, the models, the token costs, will take your gross margin down, most likely in the short term. And so how are you going to pay for that innovation investment as you go along? So I do think a number of the system record businesses will survive, will thrive, will self-disrupt. But it's a lot harder than I think the past, because the speed of disruption that's coming is as much faster as we talked about before. They need to be far more adaptive, far more responsive to the challenges and get on with it now. That's right. Because he's not going to wait for them. That's right. And I think where we're seeing the public market just trying to digest this, you know, is, and there's a lot of fear that I think to Isaac's point is kind of indiscriminately being applied to a lot of companies when the truth is really idiosyncratic. And I think the question for us as, you know, a venture firm is, well, where actually is that challenge created an opportunity? Where is value going to get created? In some of actually these traditional industries, where, as Isaac talked about technology hasn't touched or penetrated. And where can you build really long, durable businesses by actually leveraging the power of kind of AI? And so that's where we spend a lot of time just thinking about what does that opportunity set look like? What markets do we think will be impacted first? And where, you know, can, can value get created in some of those pockets? I want to just pivot quickly before to my closing question. We touched on this a little bit earlier, but thinking about sort of the broader risk framework, bringing in the impact of AI on society and the environment. I could be thinking about the, let's say the negative impacts of AI, you know, you're thinking unemployment, you're thinking widely, you know, quality, you're thinking about the energy impact to climate change, you're thinking about this information. How do you, how do you within the ecosystem think about those risks moving forward and how do you seek to manage them? Yeah, it's something I think about personally a lot, you know, having kids that are growing up in this world and, you know, eventually, you know, over the next, call it 10 years or so, are going to have to come out of college and figure out what they're going to do with their lives. And so I think about the societal impact a lot. I know that the leading CEOs and founders in AI are also thinking about it. And speaking very openly about, hey, society and government needs to be ready for this because it is going to change the makeup of the work and the type of work that will be available to folks. And it is one that is much more focused to the knowledge worker. And that's not something that we have seen in the past in terms of disruption. So it's always, you know, I think college can cost at least in the US upwards of $400,000 to get through it. Now, if the promises go pay that $400,000 as a parent and you see that your kid comes out without a job, that's a pretty rough place to be. And I think that's just a microcosm of what can occur with AI disruption overall. And so it's something societally we're going to have to go figure out. And I think people have been focused on that and are at least the folks that are in AI that are vocal about it are trying to point that out. On the environmental side, you know, over the long haul, the beauty of AI is that it's getting more and more intelligent. And I think some of the larger problems that we face globally hopefully can help, it can help with drug development, it can help with figuring out how to get better and cleaner energy. And so I think a long-term view around that while data centers are energy hungry and we have to kind of turn to fossil fuels and other things in the short term. Hopefully in the long term, the AI can help us figure out the right way to do this in the way that will impact the environment as much. So I'm more hopeful around that side. On the societal side of things, you know, I think if people focus in the same way that we focus on trying to improve, you know, things for all stakeholders around these businesses, you know, there is an optimistic view there too that this will be a force multiplier in the world, force multiplier on productivity and what we can do as a society. But I do think that there can be some disruption in the short term that we're going to have to go figure out globally. Let's finish by framing how you think about building a portfolio across the AI ecosystem from the foundational layers all the way through to how AI finds its way into the broader economy. I think you're thinking about this from your client perspective and how to invest across this. And I think the key point there is portfolio, right? If you're an individual investor, I think it's hard to play this AI trend because ultimately what you're doing is you're picking off secondaries in the businesses that already have gone to scale. And that's at a very late stage. These AI businesses, as I said before, are growing faster than anything we've seen in software. And so to be able to access that early on, you can only do that if you invest in, you know, funds that are at the frontier, that are investing at the earliest stages, and that are catching these companies at that level. But even if you could access that as an individual investor, you probably shouldn't because that game is a power lock game, right? There are definitely going to be winners and losers and more losers and there are winners when it comes down to it. And so you have to play a portfolio approach. The way that I think about it from a light speed perspective and why I feel so fortunate to be where I am, is we play across the spectrum of AI and where it's where it's meeting, you know, the world today. We have the earliest stage investors that are thinking about the frontier that are investing in ideas and founders and big ideas around that and big markets. And that's great. And that is something where we place a lot of bets, right? And hope that a few of those bets are generational businesses. Then you go up the stack and you get to where we are on a growth stage perspective, and we find the winner circle businesses to get behind and put a lot of capital behind. And again, that is something that is informed by our early stage often because we're already investors in those companies as they've grown up. But sometimes we may have missed something on the early stage. And we can get back into those businesses and we can do a lot of not only fundamental research, but asking our early stage investors, "Hey, why did we miss that? What did you see now? And is that something that's durable for that business as it grows?" And then in my business and my strategy, you know, we are trying to meet this intersection of where private equity sits and where venture sits, that latest stage of growth and venture. And where AI can be applied to real-world business, right? And so I think as it goes back to your specific question, "How can you play AI?" Well, you can play AI investing in funds like light speeds, right? And you see that portfolio approach and you can choose which funds meet your objectives and your risk tolerance level, etc. The other way is obviously going to the infrastructure layer. And there are funds that are dedicated to that. There are digital infrastructure funds that are dedicated to data centers, etc. And thinking about the energy and all that. But I think there's a few ways to play it. I think it's tricky if you're trying to be an individual investor, trying to invest in individual companies. I think it's hard to do that. And you won't see the same type of alpha that you may see by going to a diversified fund. I think one of the things, you know, we've learned over 25 years is that in periods of large technological change and sometimes over exuberance and hype, discipline and selectivity matter. You know, we spend most of our time saying no and respectfully declining investments. We looked over 15,000 portfolio companies last year. We did less than 70 investments. And a lot of people are very impressive. And there's big headlines and lots of kind of fomo generally in the market. You really have to be able to go a layer deeper to not just understand the architecture, but the engineering decisions. Why something can be defensible long-term. And ultimately, you know, even for us, we're learning every day. And it's just important to have a partner you trust. And I'd say we're, you know, kind of leveraging that expertise to also identify the right companies, but also at the right price. Because I think what history has also told us is that might be a great founder, might be in a really great space. It's got to generate a great return threshold, not just when the markets are all up into the right, but building a long, sustainable business that we think will continue to accrete in value. And I think that's what is so exciting, generally, on private markets is that so much of that value capture and that creation is happening in the private markets where historically a lot of it had had in the public markets. And these companies that we have the real privilege of backing, they're in the tens of billions of dollars of revenue and still private. So lots of opportunities in private markets, but yeah, pick a partner you trust, stay disciplined and level-headed. She'd take my job. Good never. Thank you guys. That was a genuine privilege. I really appreciate your time. Thank you very much for listening. If you haven't Any feedback or suggestions for future topics or guests, please reach out to your investment advisor. I'm Martin Randall. - Thanks for listening to the Astute Investor. To stay up to date, please subscribe to this podcast via your preferred podcast app. Or visit us on our website at www.LGTWN.com.au. This podcast is published by LGT World Management, ABN5005-311-937, AFSL231-127, and is of a general nature only. It does not contain all information necessary for an investment decision and is not intended to be advice nor to influence a person in making a decision in relation to any financial product or other investments. Any advice is general only and has been prepared without taking into account your objectives, financial situation, and needs. Before acting on any such general advice, we recommend you obtain professional advice and consider its appropriateness having regard to your personal circumstances. If the general advice relates to the acquisition or possible acquisition of a financial product, you should obtain and consider the financial product's relevant product disclosure statement or other disclosure document before making any decision to acquire it. This podcast has been authorized for access and distribution in Australia only by wholesale clients within the Meaning of the Corporations Act 2001.