Episode #24: LP POV on Secondaries, CoreWeave/NVIDIA Growth Loop, Questioning Old Reserves Strategy

The investing part of the fund manager job is only 1/3 of what you have to do. You have to fundraise, then you have to actually run the fund. A lot of people who wanted to try their head at fund management quickly realized they like angel investing and the capital deployment side of the business. They did not like the fundraising or funded management side of the business. And it's a bundle. You have to do all three it's.

The week of March 17th. Welcome to the Learning Corner, a Precursor Ventures podcast.

Welcome to the Learning Corner by Precursor. I'm Mia Farnam and I'm joined by our GP and Managing Partner Charles Hudson, where we will be talking through the articles and Goodreads that we found across the past week. Topic number one and another great episode from Strictly VC Download where Connie Loises chatted with Michael Kim, founder of Sadana Capital around fund investing. Focused specifically on pre seed and Seed Stage fund, Sedana manages roughly 2. 6 billion in fund investments across Pre Seed and seed funds. And contrary to most other LPs, Michael said that last year was among one of the most liquid from a returns perspective. Diving deeper into liquidity success, Sandana's approach to secondaries played a huge role. This of course is a continuing thread that we have on the subject, but I felt that Michael shared a really great line of sight to the outlook perspective around secondaries. Sandana recently launched a Secondaries focused fund alongside Klein Hill, which is a private equity secondaries focused fund to buy positions in high quality seed stage companies from other investors in their fund portfolio. So through this strategy the fund is able to see the winners early on, get consistent updates about them and work in tandem with sectors secondary experts to ensure that exit is done in a timely manner. I'd love to get your perspective Charles, and how the Sandana team is approaching secondaries and is this something that every LP should be thinking about? So partnering up with secondary specific funds or what are the attributes of an LP that make this actually this stronger strategy than others?

In a lot of ways Sandana is one of the first and probably best known fund of funds focused on sub 00 million funds. If you're a limited partner for these small funds, when you get a winner, you might own 8 or 10% of a company where your pro rata is going to just really outstrip what you can possibly fund out of the funds. So Your options are SPVs or Opportunity Fund. But at some point you also might want to sell those positions not because you don't believe in the companies, but because you want liquidity And I think for the Sandana team, it makes a lot of sense to have a vehicle where if a manager reorient LP in the fund has a position that they want to sell, you could use that vehicle to buy it. And I think it's really complimentary. And so I think it's really a really smart strategy. And I continue to believe that for sub 250 million funds, second secondaries are going to continue to be probably the primary way that we deliver liquidity back to our LPs in the near term without IPOs and Mega M and A. Like I don't know how else you get the money back. I think it's a very complimentary product to what they already do. Makes a lot of sense.

Probably calls into action though LPs to build even more of a higher touch with their gp. So I guess I'm just trying to think of like, how do you strike that balance of making sure that you're not being overbearing, but you are getting the best information about the winners early on.

And if you think about it, if you're Sandana, you sort of have two ways to play. You could do a direct secondary. Michael talks about this on the podcast. A direct secondary into a company that you believe in. Or you could find a family office or a small limited partner who's got exposure to a fund that has a company that you care about. I think you used the example of Mucker. Well, we know what's in that fund. If we could acquire more interest in that fund, we would buy it. And so because as these fund lives get longer, you're going to have LPs who at year 1012 might say, I know this is a great fund, I really need the cash back and I'm willing to sell my stake and give up the upside in order to get that liquidity now. And I think that is going to be an increased trend as companies stay private longer.

Michael also expanded on the concept of tourist fund managers, folks who just want to try on the fund manager title as opposed to really just live and breathe it. So he talked to Connie about backing founder magnet funds, funds where founders are simply drawn to work with the fund managers through their strong reputation and experiences and knowledge. And my question is, I'm sure actually all tourist fund managers claim that they are founder magnet funds. So what actually stands out when you're looking at fund managers to that?

This is something that comes up a lot in the work I do at Screen Door on the investment committee. I almost always ask people, why are the founders that are perfectly inspector strategy going to choose you instead of someone else. And I think in 19, 20 and 21 we met a lot of founders who are executives at high profile companies who said, you know what I'm going to do? I'm going to use this position I have as a VP at such and such company because my company's hot, people are coming to me for advice and interest. I'm going to use that as a lever to start a fund. And I think what a lot of people realize is the investing part of the fund manager job is only one third of what you have to do. You have to fundraise, then you have to actually run the fund. And what I've seen is a lot of people who started operator funds on the side or wanted to try their hand at fund manager management quickly realized they like angel investing and the capital deployment side of the business. They did not like the fundraising or fund management side of the business. And it's a bundle. You have to do all three. Ironically, we're also seeing some of those folks realize, hey, I like the fund management, I like the full bundle, let me scale this up. But there were a lot of people who tried their hand at it. I think that's fine to try their hand at it and realize this isn't for me.

What would you say for founders as they're diligencing, maybe different fund managers to lead their rounds? Like what are the flags that you say are there to kind of signal maybe they're not best long term partners?

It's probably a bit self serving, but I would say someone who's doing this as a hobby or as a side project, who's leading around, there's a chance that the demands of their day job will pick up or they will decide they want to do this. And then the person who wrote you the biggest check and ledger round is sort of out of picture. So I think the better role for those folks is a supporting capital. But founders got to make their own decisions.

Our next topic is an article by Corey Weinberg and Anissa Cardisi from the Information discussing Nvidia's role and the growth and success of Core Weave, a cloud infrastructure company specializing in high performance computing and AI workloads. More specifically, the article sheds light on how Nvidia played a pivotal role in Core Weave's growth using a mix of strategic investments, GPU allocations and cloud rental agreements to fuel its rise across the AI infrastructure powerhouse. So the article shares there were circular funding arrangements between Nvidia and Core Weave Nvidia didn't just invest in Core Weave. From an equity standpoint, nearly $100 million. It bought back its own chips from the company and accounted for nearly 15% of to your sales, Ensuring Core Weave success while keeping its own supply chain in check. It's important to note that Core Weave is predicted to IPO in the near future. But findings from this article call into question how sustainable its own revenue base can be beyond major customers like Nvidia and Microsoft. So Core Weave have positioned itself as an alternative to hyperscalers like aws, Azure and Google Cloud by focusing on high performing computing and AI workloads. But when you look at its backers, Nvidia supplying GPUs, Microsoft is a key investor, and even Nvidia renting cloud services from coreev just raises that question. Is Core Weave truly competing with the major cloud providers or is it just becoming another arm of the same dominant AI cloud ecosystem?

I'm very curious to see how people respond to this S1 because it feels very circular and I mentioned this to a friend of mine, he goes, well, it's all circular because Jensen and Nvidia are making all the money. And so all of the money kind of rotates around GPU access. And if you think about it, building data centers is really expensive. And if you find yourself in need of capacity, if your choice is like, wait till you build it yourself or pay a premium to get capacity from someone who has it on demand, even if they charge you a pretty penny. Just where we are in the AI cycle, I think a lot of people are like, I need the access more than I need to save the money. And I'm very curious to see how Wall street responds to this. But if you're Nvidia, it totally makes sense. You want people on your hardware and if they can't get it from you and they can get it from a partner, why not support that partner? So I'm very curious to see how this plays out, because if you look at the data, while there are the hyperscalers, Nvidia has been the far and away principal beneficiaries. Boom. It hasn't really trickled out. And I'd be remiss if to not mention another article that the Information wrote, which was the one about Klarna. We've talked about this internally. Klarna has made a lot of claims about the GNA savings. And I read that article and I'm like, they certainly don't appear on the bottom line. They could be embedded elsewhere in the org. It doesn't look like those lofty claims about the AI savings they were getting.

Another company positioned to go. That's right, public this year. We talked about actually the cons of having a monopoly or oligopoly a few episodes ago when we talked about advanced AI chatbots and foundational LLMs. But this also clearly applies to AI specific cloud companies as well. So when Nvidia, Microsoft, AWS and Google are all funding, supplying and effectively shaping who controls the infrastructure, they get to set the price and dictate access to compute. We're obviously already seeing this play out. With smaller AI labs struggling to secure the hardware that they need to compete. Do you see a larger trickle effect into early stage startup land happening?

This is like a fundamental question I wrestle with, and I'm not a public markets person, so I don't have to wrestle with it the same way we are right now in a moment where the inputs for air are scarce, GPUs are scarce, data center space is scarce. The power, I mean, the big thing about Stargate's power, all the things you need to scale AI are scarce. And in a world where things are scarce, if you can provide a buffer or if you can be a provider, maybe not last resort, but a provider who can bridge a capacity gap, there's always a need for that. But the question is like, what is that business business look like? If and when the supply chains or access to these core inputs become less constrained, are these businesses still as good? I don't know. I'll be very curious to see. Is core weave still a good business five years from now, if the supply of GPUs expands beyond Nvidia and other people have ways to get access to chips, metal and all the capacity, will these still be interesting businesses?

Yeah, and it'll be also interesting because we might find out after it's hit the public markets and for when it's jumped in. So that's something that I don't think we've seen similar models 10, 15 years ago, when companies go public, I mean.

There was a time where like bandwidth and fiber were a constraint. And then we built all this fiber and everyone said, oh no, we overbuilt it.

That wasn't at all our final topic. We're discussing a piece from Clint over at ULU Ventures called Questioning the Conventional Wisdom of the Reserve Fund, which literally does exactly what the title describes.

The standard thinking is that a VC firm should allocate a significant portion of its fund, roughly around 50% or more, to follow on investments in existing portfolio companies. This logic is generally straightforward. You double down on winners, protect across dilution, and you maximize return returns from the most promising investments. But Klant argues that this approach is flawed. He suggests that reserves may not always deliver the best returns, especially when considering opportunity cost and the difficulty of consistently picking the right follow on investment. Instead, they propose that firms may be better off focusing more on initial investments rather than holding capital back for follow ons. We've seen and talked about this on the pod how firms take different approaches to reserves. Some set aside quite a large pool, others keep it pretty lean and rely on funds for follow ons. Given the data that shared in the article, do you think we'll see a shift in how early stage investors manage reserves? Or is there a more underlying reason why reserves are marked off and allocated that will continue to survive no matter what the numbers say?

I think it's very hilarious we're having this conversation.

That's true.

This is an older article. I will pull something out of the archives because I think for a long time a lot of LPs were like you're a seed stage fund, you should have reserves to like protect your ownership in your winners. And I've always felt that if we're getting in at very low entry prices, I really think about what's the dollar cost average entry for us and the more we go into companies at the A and B stage on follow ups, the more we're kind of pulling up our dollar cost average entry point. And I talked to an LP a week ago who said, you know, we've realized one thing as seed odds have gotten more expensive. If you're a seed fund that's 50% in the seed and 50% in the A, you start to look like on a blended basis much more like an A fund on entry point but with seed stage risk. And does that coupling make sense? I do think though psychologically the argument it makes sense, which is like hey, if you're going to make bets early, you should make sure you reserve a lot for the winners. And I said well that makes sense provided a few things are true, you can identify the winners consistently. Which I have found is very hard.

To do early on.

Early on you get false signal 2 that you are willing and have access to writing your full pro rata check or whatever check makes sense in the context. Also not always the case. And third, that you're actually right in the end. And I think as important as reserves are, it's one of the easier ways to blow a gigantic holiday fund is to Pile into companies that either have momentum or you think are working only to later learn that they don't. So my philosophy has always been if you keep your funds small enough and your first check underwriting is good, you can have a fund where the first checks can really dry performance. And I always worry more about underwriting decisions for reserves. They're just influenced by so much how well you know the founder, how big the position is, your perception. And look, we underwrite at seed. I don't know as much about underwriting at Series C. That's not what you.

Get paid to do. Yeah. Even just the point you raised about founders, you could have an incredibly high touch relationship with a founder and know so much about the business. And maybe the information that you learn, it's maybe not the best thing to follow on. I know there's a lot of fund managers, they probably just feel like they've already crossed the line and they can't go back and they've supported them so much so they have to follow on. I think that happens when you have already allocated a specific pool and percentage of your fund because you know you're going to use it.

I can't stress this enough. I think once you're like this money is only available for this purpose. It's restricted and he will find a way to put it. I've also felt like know some of my friends who raise opportunity funds found themselves in the same predicament. Like hey, I raised this pool of money. It's only for the follow ons and my winners. I'm either not getting access, not getting the same number of winners or the amount of money I want to write isn't in line with this fund. What do I do? Because you can't redeploy that back to the main fund.

Yeah. And also there's this risk. What if you raise the opportunity fund and you actually figure out halfway through I don't want to deploy the rest of this into the things that I have. How do you tell that story to LP's effort and freeze it and set it aside? Clint makes the case that follow on participation isn't always necessary. Particularly true for breakout companies. So once a company reaches Palantir status, investors are tripping for themselves to obviously get in. And early VCs don't necessarily need to double down. But this was also part of their shorter list of exceptions to their broader argument that VCs shouldn't default to reserving capital for follow ons. One of those exceptions was supporting founders who haven't quite gained enough traction to attract new investors, but still have strong potential situations where maybe early VCs continued backing could be a meaningful lifeline. At Precursor, we've seen firsthand how not participating in a subsequent round can be incredibly painful for founders. And navigating that decision is often more delicate than it looks on paper. So how do you kind of think about that from our shoes?

Well, I mean, there's like two things I'll say. One is I look at Koia's path with WhatsApp and just how much of that company they ended up investing in. The reserves in a big fund make sense because if you identify a winner and you have exclusive access and they like you, you can end up owning a significant chunk. Not only owning a significant chunk of the company, actually putting a significant chun of dollars to work in that company. I look at Andrews and Horowitz and Databricks is probably another example, but probably one of the hardest videos, you know, and we have to make videos for our portfolio. Probably one of the hardest ones I made was I made a video for our portfolio companies and said, listen, we don't keep a lot in reserves and most of the reserves we keep are for the top 20% of our portfolio. And that is a moving target. That's based on my subjective assessment of where we can make the most money. I've told people, you might come to me and ask for money and I might tell you no. And if I tell you no, it means at this moment in time, you are not in that circle. Because to your point about having ring fence reserves, it's too easy to like, oh, we have the reserves. This company needs a bridge or an extension. We have a pool of money that we can use. It depends on your performance. It depends. But really we're trying to keep that money for the top 20% of the portfolio. And that's a moving target. It's at any point in time when someone asks me ask myself, is this one of the best investments we could make with these dollars? And because we don't hard ring fence reserve dollars, it's not like, oh, there's a pool of reserves, I want to dip into it. It's more like I have this check, I could put it into a new company, I could put into this company, I could save it for reserves for another company. And you know, Ashton did an analysis a long time ago where we looked at all of the follow on checks we've written and whether we were enthusiastic or maybe hesitant to write it. And it turns out we are generally right if we were enthusiastic, it was on average it was generally a good idea. And if we were hesitant, that hesitation was oftentimes well founded. So my conclusion is the hardest part about the follow on Reserves conversation is actually the relationship dynamic you have with the founder, not your ability to independently assess is this company a good use of follow on capital? The number of times where we thought it was a good idea and it turned out to be a bad idea wasn't that large. And I think that's the hard part. When you have ring fence Reserves is like the temptation is to use them. But based on the analysis we've done, the underwriting decision is easy. The relationship and the delivery of the method is hard. And I've taken the approach if we can tell people in advance this is how we think about following, it doesn't mean it's easy to tell somebody no, but at least they have some context.

Well, we are all out of time today. Thank you, Charles, for joining us in Palm Springs at the Equity Summit. Thank you for our listeners for stopping by. Please like and subscribe and we will talk to you with you all next.

Week, two weeks in a row in person.

Bye.