E434: Warren Gibbon on How AI is Changing the Private Markets

How I Invest with David Weisburd

Warren Gibbon discusses the necessity of re-underwriting investment assumptions in private markets due to higher interest rates and increase

Key takeaways

• Hyperscale investment currently represents approximately 7/10ths of total corporate U.S. capex.
• A massive bifurcation exists in venture capital, where 12 firms account for roughly 75% of this year's fundraising.

Main topics

• Private Equity Re-underwriting
• AI Infrastructure & Hyperscalers

Notable quotes

If you're not at the table, you're on the menu.

Conclusion

Investors must move beyond historical precedents to find value in spec

Transcript preview

Speaker 1 (0:00) When you look at these companies plowing in these massive amounts of earnings into infrastructure, specifically data centers, what do you think about that? Speaker 2 (0:10) If you add up hyperscale investment this year, it's around 70 % of total corporate U.S. capex. It's an enormous number, and it's probably going to increase next year, too. What I do wonder is that the beneficiaries of the AI infrastructure spend, chip companies, memory, I'd be careful about underrating that. There's a massive bifurcation in venture where 12 firms are accounted for about three quarters of the fundraising this year. You Speaker 1 (0:34) are not bearish on private equity today. Why? Most investors today still believe that the private markets will outperform the public markets. What do you think? Speaker 2 (0:44) I think investors today really need to take a hard look at that thesis. When I look at private markets, a lot of the data that people focus on is data that is really over the last 20 years or really over the last, say, 40 years. And one of the biggest things that happened in that period, as we all know, is we had a period of declining interest rates, which acted as a strong, I think, tailwind to returns. And more than that, we're looking at a period where... equity markets had some volatility. I mean, one example I can give you is I've seen quite a few studies looking at the period 2000 to 2020, that 20-year, two-decade period. And obviously, you're really comping off an equity market that was very overvalued in 2000 and then subsequently had pretty poor performance. Speaker 1 (1:33) Today, my number one Speaker 2 (1:35) recommendation for investors and how we're thinking about it here at BFA is re-underwrite what you're thinking as regards private investments. why you think it makes sense today, and really getting back to it, sort of a first principles approach, which is, what are my underlying securities? What are my underlying businesses I'm going to own? Why are they going to increase in value? Because if you're looking at the past and saying, that's what I hope is going to happen, I think that's going to be a disappointing strategy. Speaker 1 (2:03) What's changed over the last five years? Speaker 2 (2:05) One is... A real plethora of new players and the amount of capital that's gone into these markets is obviously very significant. One of the beliefs I have in investing is large sums of capital coming in is antithetical to go forward returns. It gets much harder to generate the returns you're kind of hoping for. When you think about the endowment model, which was popularized by Swenson and others, one of the reasons why it was so successful is that you had they were the first movers in that thesis you had an inefficient private market You had lack of sophistication to how you can create value. At the same time, you had some great tools in the toolkit to increase value at companies. You had the benefit of declining rates, which meant you could put more leverage on these businesses. And you had a lot of wind at your back. And so the amount of capital coming in was also driving returns, I think. And so now we've got a lot of capital in place, more firms chasing. the same amount of deals. And you can see that most clearly in just multiples, right? We're looking at, let's say, 12 times EBITDA for a typical middle market buyout. A few years ago, that was eight. Now, declining corporate tax rates have kind of justified some of that expansion, but it is very tough today. And to underwrite the same kind of returns, you have to make some pretty aggressive assumptions. And that's what we do challenge GPs on today. They're going to be forced to really focus on the operational side, which a lot of them say they do, but I think we need to see more of that. Speaker 1 (3:51) On the public markets, the multiples have also gone up. Absolutely, David. And I think if you think about Speaker 2 (3:57) What many large investors have, they've really adopted the endowment model. They've increased their allocation of private markets. And as we've got further into this decade, I think we've seen the fact that one of the corollaries of that is that investors have really underestimated the value generation of the public market. And yes, multiples are high. But the thing I think people are still really struggling to appreciate is the immensely powerful combination it has been for large tech public players to combine revenue growth with very, very strong profit margins and free cash flow generation. Today, that free cash flow is being used to fund the AI infrastructure build-out, and we can talk more at length about that. But investors have really underestimated. If you go back, David, to the middle of the last decade, we were coming out of the great financial crisis. We kind of had a jobless recovery. Things didn't feel great. The economy was doing okay. But people were not particularly positive on the equity market. The valuation was somewhat high. It was unclear what the next real driver of Google and Meta were still relatively young companies. Microsoft was having a big change. And this has really been a game changer. And if I think bring it right up to today, S &P earnings in the second quarter are up 29 % year over year. Your tech earnings are up 50%. And even outside of that, the rest of the S &P earnings are up 19%. So we have an earnings boom right now. And so if you just think about that, are we seeing the boom in earnings power of middle market buyout companies? Probably not. One area we are seeing a boom is late stage venture tech, right? Some of those companies are doing fantastically well. That's been a great story. But that's one of the crawlers of the down model. I think investors have kind of missed what's happened in the equity market. And so have some public equity investors as well. Speaker 1 (6:00) I want to get into private equity and venture capital in a bit. But first, you come from the public investing side. That's where you... built your career. When you look at these companies plowing in these massive amounts of earnings into infrastructure, specifically data centers, what do you think about that? Speaker 2 (6:21) It's Speaker 1 (6:22) a Speaker 2 (6:22) great question, David. And we've debated that a lot internally. For us, we're more in the believer camp. I think there's three camps of investors right now. There's believers, there's sort of AI skeptics who think this is sort of a house of cards. And then there's just folks that don't really know what to think. I think Right now, it does give me a lot of pause, the amount of money that's being invested. When you think about just the hyperscaler investment today, if you add up hyperscale investment this year, it's around 70 % of total corporate US capex. It's an enormous number and it's probably going to increase next year too. The source of this is the free cash flow from those big tech companies and that's being eroded, which in general, if we look back in time, as you know, anytime you've seen these big CapEx booms, it has really been followed by some kind of bust, whether it's railroads, whether it's internet, more recently, the late 90s. And so I think we're right to be cautious here, be concerned about this. What I do wonder is that the... Beneficiaries of the AI infrastructure spend, so chip companies, memory, that I'd be careful about underwriting that. Right. Because tech, well, there's so much incentive today to find better, more efficient ways to do things. And so that's something I'd be questionable. But I think the hyperscalers, my view is that they're in a very good position. And we saw that in the second quarter with hyperscaler revenues were up 50 percent year over year. And they're really cementing themselves. They're right in the middle of everything we do and around AI. So however AI kind of plays out, they will be in a good position. I think the broader question for public equities and private companies too, both private equity and venture, is how do we use AI? Revolutionary technology and what are the implications? What are the applications that we really haven't seen yet, probably thought of, that in 10 years from now, 15 years, 20 years from now, it's going to be? change kind of the way we do business in whatever sector. I think that's the thing that investors should be thinking about. This episode is presented by Juniper Square, the operations partner for private markets. Speaker 1 (8:32) I think the unknown variable is how far on the S-curve we are, how much AI is going to really improve. There's a lot of views on that. Some people believe that we're going to have ASI, superintelligence, meaning that AI will be smarter than the entire collective intelligence of human beings. That's probably the most extreme view. And there's everything in the middle. But one of the things that's missed is that even if tomorrow China, U.S. and other countries decided to halt AI development. there would still be significant economic growth just based on the models that are out there today. I Speaker 2 (9:05) completely agree. It's right to be skeptical on how AI is changing everyone's lives. I'm sure everyone listening and in their private conversations with friends and family is debating that. It comes up all the time. I've kind of resolved with some of my friends, I'm not going to talk about AI because it's... We kind of default to that on every conversation. But I think that I do subscribe to the view that in the same way the internet and the digital transformation played out, we didn't really know all that we could do from our phones in 2004, 2005. And the same thing will happen here. I mean, I just think of the market for insurance, the way we purchase insurance policies and the way that gets transacted. I mean, that is one area that could be completely revolutionized through AI agents doing that for you, essentially. We have a lot to navigate between there and here, which is around security, around how this all works. So there's a lot to be done, but I do think there's a lot of potential. Speaker 1 (10:10) I want to get back to private equity. You are not bearish on private equity today. Why? Speaker 2 (10:16) What we see is unfortunately for GPs, a tougher environment. And that comes from the fact that real interest rates today are higher than they've been, and they're likely to move higher. In the last few days, we've seen higher yields on long-term debt. Speaker 2 (10:35) We'll see how that plays out, but clearly that's a headwind, right? It's increasing the cost of leverage. It's increasing the discount rate on future earnings. So long duration growth assets are going to have some headwinds there. Going back to AI, it's both a threat and a potential beneficiary. I mean, I can see a world where companies could be revolutionized by the application of AI technologies, both on the revenue side and on the cost side. In a similar way that, say, outsourcing was in the... 90s or the digital transition was in the 2000s the problem is that it can be very disruptive as we've seen that now with software right enterprise sass businesses to be clear on that i actually think some of those companies are going to be just fine but oversold a little bit yep Speaker 2 (11:30) But clearly there's a risk for disruption. So how are GPs navigating that? It's difficult and it's going to be case by case underwriting. That's one thing that I've realized is that there is no substitute right now for really getting in the weeds, figuring out company by company what the thesis is. Speaker 2 (11:49) You could have two similar companies, similar kind of applications in different verticals. One could be great. One could be just replaced. So that is a tough, tougher environment. So overall, we remain cautious. That's not to say we're not allocating today, but we are. We just want to make sure we were re-underwriting everything, you know, in a kind of a rifle shot approach rather than we're just filling an allocation sleeve. with strategies. Speaker 1 (12:18) When I look at private equity, and I want to figure out which managers are going to do well, you have to go to the underlying companies. So in many ways, a fund is just a wrapper on a portfolio of companies. So the question becomes which companies are going to do better and which companies are going to do worse post-AI. And I think it comes down to both the industry and the sector, but also a lot of times to the management. In other words, are the management integrating AI and will be the disruptors, or are the managers not integrating AI and will be disrupted? And I think we're going to see. a lot of consolidation across the field. One early example of that is Thrive Capital. They're doing this roll-up of CPAs and they're making everything AI-generated. And obviously, roll-ups in general have some favorable economics, but when you integrate technology into it, it's a whole other game. And I think we're going to see a lot more roll-ups, a lot more consolidations in sectors that previously were highly fragmented. Speaker 2 (13:14) I completely agree with you, David. I think that's the right way to think about it. As an investor, And as a GP, you have to have thesis around AI that is, these companies are going to be the winners of that process. So I think that the strategy that the Thrive team had there makes a ton of sense. And AI, we know it is most powerful in scaling a repeatable business process that needs a lot of data. And that accounting clearly is one of those areas. I think that is what investors should be focused. I mean, we think healthcare is a sector that is also ripe for the application of AI. You know, we can... debate. That's a separate podcast, perhaps, but just, you know, the amount of money this country spends on healthcare versus healthcare outcomes, there's a lot of potential there. And so we're looking at some of those opportunities today. But it is about that. It's not about we're just going to buy a basket of businesses that are pretty well run, and we're going to make some improvements, and we're going to end up selling it to another private equity company in five or seven years. I mean, that has worked. There is still a lot of capital on the sidelines. But I don't think that's what you should be underwriting today. Speaker 1 (14:30) Going to the GP level, I had Thomas Scriven, head of private equity at University of Pennsylvania, and he had just invested in a manager, three former Palantir engineers, and they would be deployed into their portfolio companies. So they would be FTEs, literally deploying themselves. It's not the most scalable model, but I thought it was one of the most fascinating models. Have you seen any innovative GP models on the private equity side? Speaker 2 (14:54) We've seen one that actually we're invested in, which is similar to that, which is actually Going in and fixing businesses. in the software space and really taking over the business and operating the business. And I think they've got their hands full with what's happening in the software sector, but that is real operational. That is getting in there and changing the business, changing how a business develops a new product, how they implement that, how they go to market, everything. I think that is what it's