Mark Peter Davis of Interplay joins Nick to discuss AI’s Effect on the Demand Curve, Investing in Durable Barriers, Large Cap vs. Small Cap Seed Deals, and Building an Internal AI Machine. In this episode we cover:
- AI’s Role in Interplay’s Operations
- Investment Strategy and Early Signals
- Impact of AI on Software and Venture Capital
- Job Displacement and New Opportunities
- Barriers and Durability in Software Companies
- Challenges for Traditional Growth Companies
- AI Native Services and Market Segmentation
Guest Links:
The host of The Full Ratchet is Nick Moran of New Stack Ventures, a venture capital firm committed to investing in founders outside of the Bay Area.
We’re proud to partner with Ramp, the modern finance automation platform. Book a demo and get $150—no strings attached.
Want to keep up to date with The Full Ratchet? Follow us on social. You can learn more about New Stack Ventures by visiting our LinkedIn and Twitter.
0:19
Mark Peter Davis joins us today from New York City. He’s the founder and managing partner at Interplay, a seed and Series A stage venture firm, studio, and multifamily office. Before founding Interplay, Mark was an investor at Primary. His investments include Coinbase, Warby Parker, Jackpocket, and Ali, amongst many others. Mark, welcome back.
0:44
Thanks for having me.
0:45
It’s been a decade, my man. It’s been a long time. It’s crazy.
0:49
It’s crazy. You get old. Did that happen?
0:51
Well, you don’t look a day older, but yeah, you look pretty good
0:56
too. We’re survived. All
0:58
right. So it’s been a decade. Back then, I mean, it was a fun episode. You gave me a lot of good advice in the early days. Pre fund one, amazing. But back then, you argued argued that venture firms should build companies, not just fund them. You had a bunch of interesting takes. What I’m curious to hear to start off today is sort of what’s changed most about your approach to venture in the past 10 years?
1:26
It’s a broad question in a good way. So the firm has evolved. We now have effectively five divisions that are kind of at the forefront. Core piece still venture investing. We’re typically doing early stage checks into super high growth companies around the U.S. We’re based in New York, but we invest across the country. Part two is we have an accelerator. We work with a very small number, typically three to five a year, of pre-seed or seed stage companies that have really unique narratives and potential, and we get operationally involved. So we call this an accelerator because there’s not really a better name for the category. But kind of internally, we refer to it as becoming an interim co-founder. One of our partners actually embeds in the company for six months, and the goal is to take as much friction out. It’s not a class program. It’s manoe mano company and interplay person, and we’re we’re in it. And the goal is to have a high success rate versus a high volume of companies through, and that’s gone really well. Part three is we have a studio, and in the studio we are starting companies de novo completely from scratch. We’re in a way on kind of our third thesis with that. The first thesis was service companies for startups. I started doing that at the end of 2011, and we launched a bunch of companies through that cycle. And now we’re also doing American dynamism companies and AI application layer companies. Four is we have a secondaries practice now, so we’re helping to transact in that market. And five, if you do all this for long enough, it necessitates a family office. So we have a multifamily office where we’re stewarding capital for currently 14 ultra ultra high net worth families, and we’re helping them manage kind of their full book, top to bottom, full service operation. So those are the five pieces, and as per the name, they all have interactions with each other in very healthy and productive ways, and that’s what we’ve been doing.
3:36
Unbelievable. So I suspect you’ve had to put together some platform and structure around this. Do you have what, like a GP, an MD, a GM running each of these practices, or how have you structured it?
3:50
Yeah, we’re eight partner. We’re eight partners currently, and so essentially each of the business lines has, and there’s no hierarchy in the partners really, except for I guess me, but the each of the business lines is a partner that’s focused exclusively on that model, and then there’s three of us that kind of support across. And in addition to that, we’ve built out a tech function and an operations function that support all five of them. So there’s there’s some scale advantages built into this that have allowed us to really streamline and build forward. I think, as far as I can tell, in all the conversations I’ve had, kind of coffee and lunch with other GPs, other folks, I think we are on the absolute frontier of AI integrations. We are pushing the envelope and building proprietary technology internally. So there’s a lot of things you can do when you get these scale advantages to make each piece better than it would be on its own.
4:48
What do What are you most excited about that you’ve already built, and then what’s kind of the highest priority on the the roadmap for internal AI?
4:58
We’ve built a stack. Of AI systems that you know, we’ll actually we actually have a bunch of people knocking on the door to get access to the system we built internally. But we built a proprietary platform that works across all the models and all the integrations. So it’s kind of like what the future of AI we think we’re all imagining it will be. I think we have a version of that genie in the bottle right now. The KPI that we measure internally, we’re actually looking at man hours, is we have a 50% lift on our team in terms of throughput and productivity. So there’s a bunch of technology layers that go into that on top of the things that are available, but we’ve been really pushing it forward. And next week are going to start sharing that with the first couple beta customers. But we think it’s a an application layer and an application of AI that this or equivalents will end up being integrated in I think every operation in the country,
6:01
so so give us some teeth. You know, like what’s what’s the priority? What do you what are you trying to build next? It sounds like you’ve built a lot, but is there a workflow you’re trying to solve? Is there you know a key pain point within the firm that you think AI can help sort of collapse that pain.
6:22
See, that’s the rethink. Is instead of thinking about it from a one-off tactical tool, I’m going to open Claude CoWork and I’m going to have it do this routine. We’ve built a system that acts like 15 people who have jobs that repeat and recur, and so everything from triaging emails, diligencing companies, writing reports, keeping up with portfolio companies, internal communications, enforcing policies and standards within the firm-anything that, if you were to think like, hey, if I had an unlimited number of people on the team, kind of floating around the rim. I could, I would eventually do this whole long tail of activities. We are methodically rolling all of those out through this one system.
7:11
And and how many FTEs at the firm?
7:14
We’re about 30 total across all five of those divisions, and that that seems you know it’s like anything in these venture and family office land. Usually it’s kind of headcount light and get a lot done, and then you layer on top of that AI. And we don’t need a whole lot of bodies to get it
7:33
done. So so did you have to bring in somebody to lead that initiative to build all this out? You know was this something that you started building and then had some tech help. You know how did how does this come together? Because I I find even at my own firm it’s small, right? There’s only five of us. We’re not running a platform at the scale of Interplay, but you know it’s like each of us are building stuff. We’re coming together. We’re kind of Frankenstein putting it together, and we’re we’re all doing some really cool stuff, but I’m curious how people unify and also you know bring some some leadership to it.
8:09
There’s a general motion we have for building out the firm, and it it kind of comes from a very bootstrap, highly efficient startup ethos. The kind of organization you’re talking about, we started there. I bootstrapped the firm. We couldn’t afford an office 15 years ago, and layer by layer, we’ve built this. So the general motion we go into now is we identify something that we think might be a frontier opportunity for the firm, either an operational capability, a technical capability, a new type of service to offer that we think is relevant and synergistic, and I will put my hands on trying to market test it and set it up personally. So I’m the tip of the spear. If I get it going and it starts to get some motion, we will then staff against it, and that could be reorganizing this existing team, it could be bringing in new bodies, but the the game for us is we do this very methodically and slowly. So to get to this point where we now have five divisions, it’s taken 15 years of layer by layer because we want to lay a brick into the wall, make sure it’s extremely sturdy, pressure test it. Make sure we have no operational constraints on time, mine included. And then when we have space, we’ll look at like, okay, if we added this, it would make us better at that. It would make if we added this other brick, it’ll make that first brick stronger, and that kind of builds in this synergy concept. So that’s our method. So in terms of like the the AI application, just as an example, we started toying with AI like everybody else. Then there’s some questions come came up internally about hey, can we push the envelope forward and get towards the frontier? Someone raised their hand, went heads down on it, came back with an answer. We spent a couple months alpha testing. Eventually, we staffed up to a handful of engineers working on it, and keep leading into it.
10:10
Awesome. Well, maybe we can come back to this, but I do want to jump in to the investment side, right? Yeah. So you back companies like Coinbase, Warby Parker, Rowe, Jackpocket, Ali. What’s the early earliest signal that a company can become a category leader? These are very different companies, but you’ve seen a lot of journeys now. Like, walk us through some of those early signals that you have a breakout.
10:38
So when I first started in venture, time frame is like roughly 2006. I started interning, so 20 years ago at a venture fund, and that became a full time job. And then I went to another firm, and then 15 years ago, started my own shop. It feels like every 234, years, you you kind of think you know exactly how to invest, and you can you can see the matrix, and then the windshield wipers go across the windshield, and you look and you’re like, “Oh, it was foggy a moment ago, and now it’s clear. And you have this relative, like this advancement, this leap in seeing more. And after you have three or four of those moments of clarity where you level up a little bit. For me, at least, I started a pattern match that there’s probably every time I think I have 2020 vision, I can see exactly what’s going on. There’s probably another level. There’s probably a way to be better. So this is an ongoing journey. One of the things that clicked eventually midway through my cycle of this to date, is that it’s just really about stage appropriate risk profiles, and to identify winners at pre-seed and seed changes a little bit in titles and product stages, but let’s say pre-seed more. You’re betting on teams and ideas, and so you have to have a really good sense of strategic frameworks, how to win. I lean a lot on barriers as a concept, even if companies don’t start with them at day one, but they layer them in. And then I, over time, developed a bunch of muscles around how to identify capable leaders and entrepreneurs, and that that’s taken a lot of let’s say leveling with the windshield wipes to start to see patterns of what are those characteristics, and it’s not the usual signal. It’s not where they went to school, right? There’s other dimensions that are far more helpful. The Series A for me is now the point where the companies usually got material financial KPIs. We’re talking. We’re usually looking at companies doing a million plus in top line revenue. You can start to read down the PNL. May not have a positive bottom line in our industry. That’s normal. You’re reinvesting for growth, but you can see a lot of the healthy operational margins throughout. You can see CAC LTB, and so the math starts to tell the story, so at Series A, which is where I primarily write checks out of our venture fund, we’re very much looking for operational and financial KPIs that we think are strong signal. Yes, we want to love the strategy and the idea. Yes, we want to believe in the team. Those are given, but the A phase we get the we get the benefit of proof points of the financial KPIs, if you can bet on that those two dimensions together, I guess like three: idea, team, and KPIs. There’s a lot of risk you can take out of a deal. Now it’s venture; it’s early stage. Even with all that de-risking, you still get it wrong. But the game is to get it wrong less, and to get it right a little more, and when you’re putting portfolios together, playing those odds can be very lucrative for the LPs. So that’s the thinking. I’m looking for the triangulation of those three checkboxes.
13:52
Got it. So we started off talking about AI within interplay. Would love to hear how you’re thinking about AI in the portfolio, and a decent place to start is some of the narratives I’m seeing around. So we have some VCs that are fearful of the effect of AI on software, right? They they contend that if AI lets anyone build software cheaply, then software gets commoditized, prices collapse, margins evaporate, and defensibility disappears. So you know, in this instance, you know categories can get worse according to some VCs, but you know this is a mindset shared by some, but not all. I’d be curious to get your take on AI’s effect on software at large, in sort of the demand curve, as as you and I have traded some notes on.
14:54
Look, when anytime there’s a major technological shift, the media. The outsider on the industry wants to distill it down to a black and white answer: AI is good or AI is bad. The truth usually lies in a myriad of different effects that play out. So I do think AI will be disruptive to a lot of software players. I also think it will create a lot of software opportunities, and so the game is to get a little bit more nuanced than good or bad, and to find the trades within. We are on a very significant wave of change coming through this industry. It’s already here. Three years ago, AI was floating in the ether, but you couldn’t really use it. I would say a year ago, for most of us, it was kind of a chat. It was like a search tool that gave you better reporting, and now it’s getting into workflows in a very meaningful way. The story on AI for me and how to evaluate the market is that every company is using it operationally to get productivity gains. That’s already happening. There’s a lot of complexity. Actually, it’s a side conversation for how organizations are now having to adapt to the fact that every member of the team is kind of managing people, but those people, air quotes, might actually be deployed agents. So there’s organizational change. Roles of everyone in the team is a little different. We’re in a new framework for how people work. I felt like this year it was kind of pre iPhone to post iPhone workflow for me, having my first smartphone. The same parallel in being able to leverage AI fully for the first time and completely change my workflows. So that’s coming. That’s internal operational implications of AI strategically and how it determines which companies will be successful. We’ve been on a longer term trend than just AI. It’s the latest chapter of a longer trend of the commoditization and the reduction of capital requirements to build software. At the end of the ’90s, you had to get the first 5 million bucks, and I’m making up a number, but it’s in the right time zone of capital raised just to put servers in the corner of your office so you can run the website. Yep. Okay. The cloud shows up AWS first, and then everyone else, and the cost, the upfront cost of you know getting compute plummets. We’re now seeing increasing productivity in coding, which means less labor required to build. Eventually, all of these will continue to asymptotically approach all these layers of cost will asymptotically approach zero. I don’t know if they’ll ever hit full zero, but they’re coming down. This is really significant strategically for the venture community, in my opinion. In 2015, 2010, 2005, even closer to 2020, you could get away backing companies that didn’t have a really baked-in barrier, because it used to be hard to start a company. It was hard because you had to have a good idea, sure, but you had to convince people to give you money. You had to convince people to join your team. You had to know the method, listening to customers. There’s all these motions that had to go into the process of making a successful business. So, if you had a company that had inherently no barriers, but if you could spend $1 and make three, you probably were only you know even if yeah as you fast forward, you’re only going to be one of 345, competitors in the space, and in these large markets, you could still make a bunch of money. I like to think back to Constant Contact and Mailchimp, newsletter platforms that now take under a day to make with current AI systems, right? But those were huge technical lifts. It was just hard. Well, now it’s easy. What’s left is not companies that have no barriers exclusively. There’s a group of those. There are a lot of companies now where there’s no inherent barrier, and if if there’s money to be made, 100 entrepreneurs will show up. The money will be made. It’ll just be cut up into 100 little pieces, and there won’t be venture scale yield for investors. Founders will do great, but there are still real barriers in the market. They’re just not the setup barriers. The barriers now are all those traditional barriers.
19:25
I like to call them academic barriers. It’s real switching costs. It’s you know, which is like data and also having data advantages and network effects. It’s all those things we studied in school. Those academic barriers were being invested in for the last 2030, years, but not exclusively. Now we’re in a world where investing outside of those academic barriers puts you at risk of tremendous competitive forces, and it puts VCs at risk of not being able to get to a venture scale yield because the market gets chopped up too many ways. So for us, we’re deploying in. I think the opportunity. In front of us are massive. The number of new companies going to be created. I think we are at a growth inflection point of opportunity set for investment, but there will be a lot of long tail that are no competitive barrier dynamics, where there’ll be a lot of entrepreneurs all making a few million bucks, and there won’t be investor yield to be found. There’s going to be a lot of other companies, maybe new models, maybe disruptors that will come out with true barriers and spaces, and that’s where the yield is going to be for the VCs in the game.
20:31
And what about the demand side, right? The demand side of the curve, like how does this cost collapse on the AI side interplay? No pun intended, but with
20:43
I’ll take it.
20:43
Sort of the disruption concept, right? You collapse price, you increase volume and access. How are you thinking about that as as new markets are created by agents, AI, machine learning, computer vision, other tools?
21:01
I think the story you’re kind of dipping into on the cost reduction is really a story about jobs, and there’s a big debate and a very healthy debate happening. We’ve been having a lot of debates in our firm. We’re all constantly reading, not just the headlines, more academic books and papers. We’re trying to look for historical models and comparisons, and I don’t have the answer, but I do have a perspective now. When AI hit and I saw the the cost curve collapsing of producing software, my first intuition was massive job destruction. We can already see places that we all know are going to get hit, right? We know there’s millions of drivers in the country that are going to be displaced by AI-driven cars. It’s not an if; it’s a win. They’re already forming up in lobbying groups, trying to pass legislation, which is always the last-ditch barrier to preserving a job that’s going to go away. And there’s been a lot of these jobs throughout history. There used to be a group of people in New York City before electricity was prevalent that walked up and down the avenues, lighting the lamps at night. They would light the candles, and that job’s gone. No one misses it. People found other roles. Now there was a difference when things were moving a little slower. Workforces have trouble kind of reorganizing. You got to get educated. You got to find the right people. Job discovery is not easy. New companies have to pop up. They have to get to a scale to hire people. So there’s all this friction that is very real. And I think for people in that those those I would say high risk workforce categories, I think it’s pretty scary because there will be friction and there will be a need for transition. I would be getting ahead of it if I was in one of those roles. The reality, though, is historically all of those jobs, in theory at least, have been more than displaced, more than replaced by other new roles that came out. The question people are asking now is, “Hey, we know a bunch of jobs are going to go away with this. What we don’t know, since it’s moving so fast, is what are the new jobs? I can’t see them. Someone might be thinking, you know, I can see the new startups coming out, but they’re early. The industries aren’t robust. They’re not hiring enough people yet, so the speed of innovation may create more friction in this transition period. But I do have a bit of hope, and for me, the hope comes from a belief that there is so much more opportunity in plain sight that no one’s looking at. Here’s a great signal. This was the aha moment. If you can take, let’s say, the cost of building software down to zero, okay, maybe it’s not profitable. It’s profitable to be a coder. I get it. How many products would we have built through the last 1020, years that we just couldn’t justify economically because the opportunities were too small? That whole part of the demand curve opens. Easier to see this less abstractly in the physical world. Same parallel from AI displacing software roles to humanoid robots, which are kind of in the dugout, about to get on, about to get up to bat, right? Them coming into the blue collar workforce, they’re going to displace jobs in construction, etc. What gives What gives me hope is as I drive around town, I see a lot of dilapidated buildings. I see potholes in the street, and if the cost of repairing the world we live in keeps coming down, well, we should be solving more of those other obvious in plain sight problems, so it’s really easy to see that there’s a lot more work to do. There’s more demand for these roles now. Maybe it’s different. Maybe a human’s not holding the shovel anymore. Fast forward to human and robots being deployed, but maybe the person who is doing that is coaching and supervising the 10 robots. Doing it and making sure all the potholes are filled, but we have a lot of work to advance the society to a more utopian level of society, and that’s kind of the way to think about what this unlocked demand is.
25:13
If it was really cheap, what would be all the other things we could do? And those will inherently create new jobs. So in my mind, there’s going to be job displacement. There’s going to be job creation. The really hard truth of it, from my perspective, is that there will be a lot of friction in the transition, and we as a society need to figure out how to help people make those moves and support them financially.
25:37
Mark, you you talked a bit About barriers,
25:42
yeah,
25:42
and mostly in the context of barriers to entry. I’d also like your take on durability and sort of long-term defensibility. So these these concepts are are related, but as you think about demand expansion, you know, costs coming down, and obviously the the companies that you’re looking to back, you believe will have a long term, durable competitive advantage, and those are the companies that are going to attract venture capital. You mentioned things like network effects and and some other aspects that are interesting, but if you had to describe sort of the software losers versus the software winners of the future, you know what would be the the characteristic on the lose characteristics on the loser side versus you know the characteristics on the the long term, durable, indefensible winners side,
26:45
I wrote a really boring book that I published in 2011. It’s a handbook designed to help entrepreneurs raise capital, and I wrote it over the course of five years. My first five years as a venture investor, so I was a young VC, and I was I was learning the game, and I’ve always fancied myself first to be an entrepreneur, now sitting in a venture role, and I was learning every day all these lessons: how to pitch, what these terms mean, how people are thinking about clever financial structures. I would write them in a form of a blog, but mainly I was writing it for me. I was writing it to help coalesce the concepts, the frameworks, the strategies. After five years of writing, I put it all out publicly for people to consume. There were a lot of people over time that asked for a physical copy. I guess they consume information, paper, old school, want to read a book, and so I published this book. It’s called the Fundraising Rules. It’s still available on Amazon. I do no book promotion, never have, but it is there. Most of what’s in that book is designed to be a pretty on-the-nose handbook for in the meeting. If they ask this, say this. It may not be a huge revelation. There’s one segment of it that I fancy to be my strategic contribution to a little bit of thinking in our industry. It’s a framework for how to think about value creation for the entrepreneur. And if you would ask an entrepreneur, as we might argue in the in the kind of first section or two of the book, you ask a founder day one, what’s going to make you money? What are the drivers? They’re going to say a big market, a great team. We know the checklist. Very few include on that list of a handful of things the idea that they need to align the financing strategy of their company with the company’s potential and nature. And if you raise venture for a company that shouldn’t be venture backed, that will, in most cases, destroy the yield the entrepreneur will face. There’s great examples, and any of the extremes is bad, right? And so there’s companies that, if to be clear, if you don’t raise venture, you’re going to screw it up, and it’s companies where if you do raise venture, you’re going to screw it up. The key is to find the match, the alignment. So a great example of this: imagine you’ve got a company. There was this guy I knew. He had a business, small, small market. He was doing $5 million top line. He was doing $4 million bottom line. But he owned the whole thing. So this guy was driving around in a Maserati, making 4 million bucks a year, totally crushing it. He bootstrapped the business. Now, if he had raised venture capital for that company, which did not have the potential to scale, a VC would have most likely sat around the table on the board saying, “Hey, 5 million is cool, but it’s not going to really move the needle for our fund. It’s not what we signed up for. It’s not what you sold us. So let’s now go back to the drawing board and innovate, and maybe we’ll raise more money, and dilute the cap table, and maybe take on more risk with a new idea. And they might go bust on that new idea and destroy the first $5 million of cash-flowing revenue. That could totally have happened, but this fella had. Really good alignment. He had bootstrapped the company that was a bootstrap styled company in its nature, and he had funded it correctly, and he did very very well. Maybe better financially than a lot of venture back founders. Now you take the other side of the game. We all know a lot of these marketplaces online. There were a lot of eBay’s in the beginning. Only one one, right? You’ve got to in some of these sprint land grab situations where there’s a network effect and it’s a winner take all. You got to fill the coffers, you got to deploy, you got to grow, and you you have to be first to arrive. So that’s a completely different game. And if you don’t raise venture in that construct, you could miss the whole opportunity.
30:37
How you finance your company is strategically critical for determining the payout for a founder, so that’s that’s a given. This framework is little two by two to help people figure out which strategy is right for them, and I think that little framework, which is I guess now 15 years old, and they teach it at a few business schools, is a good signal for how to think about which software companies now, and all software companies are AI at some level, at least in the operations, if not in the actual offering, and almost all are figuring out how to put in the offering, right? How are these players positioned to get a venture scale yield or to have a smaller business, look if it if it can’t ever generate profit, it’s just a bad idea. Okay, let’s say it’s this $5 million business with an 80% EBITDA. That’s not a bad business. It’s only a bad business if you need it to be $50 million top line. It’s about setting expectations, getting alignment. To me, the core story again is barriers, and I think anytime you have a market, it’s barriers and market size. Those are the two combinations. So the fella who actually was at this $5 million top line had barriers, owned his market, but had a $10 million market size. It’s tiny, not venture fundable, and he made the great decision of not raising venture capital. The yield in early stage is still there. It’s in the companies that have real barriers and huge markets, and that small subset of the market is small. It always has been. I think we’ve been investing around that category because we could, because you could get away with it. You could invest in companies that didn’t really have barriers five or 10 years ago. They didn’t have academic barriers because there was a built-in barrier. It was hard to start a company. It’s not anymore. So to me, the investable set of real venture scale yield companies are still there. They’re the same. All of the you know free passes are kind of out of the system, and any of these companies without barriers, they’re going to face stiff competition. They could be very successful, but not as a venture bet. They could be very successful as a bootstrap small founder team. So that’s the line in the sand for me. That’s how we think about it internally, and it requires a sharper pencil on really strategically evaluating companies on the way in now, it’s less forgiving than it was in 2015.
33:06
Mark, you’ve described sort of a barbell and seed, right? The old two to $4 million round, the new $20 million plus mega seed, kind of an awkward space in between. Which end of that are you playing on, and can a fund of your size, you know, a fund like yours,
33:27
yeah,
33:27
compete for the mega seeds? We,
33:31
the answer is there’s probably a way to do it with a small fund size, but I do think you need to pick a lane. I think the venture market is bifurcated. The language I use internally and what our team uses is large cap early stage and small cap early stage. These large cap early stage companies, their seed might be a nine figure seed round valuation. Their A might be a billion. And while that sounds completely ludicrous, we have unlocked parts of the economy where these companies can go zero to 25, 50 million in 12 months, and the ratios and proportions start to actually look reasonable at those valuations using consistent multiples. So companies that have really unlocked the jet fuel at a level we’ve never seen before, really do fall in an orbit of their own, and there is a rational way to value them and play those game, play the play that hand. There’s the more traditional looking, what I call now small cap early stage, which is where I play. Our secondaries practice will touch more the large cap early stage, early stage, small cap early stage. We look at those plays, and they have very traditional characteristics. Where I think you get in trouble is if you try to build a portfolio of both, because it’s really hard to get a balanced risk reward diversification if you’re putting 25 companies in five. You’re buying up a crazy. Valuations, it’s not the same product anymore, and I think from an LP perspective, if I was a sitting on the LP side, I would be thinking about my asset allocation and risk diversification in this category, and saying I want a fund manager at small cap early stage, smaller fund. They’re going to seek really good yields, healthy multiples. They’re going to buy in at rational valuations and look for reasonable exits, which aren’t always $10 billion plus outcomes, but they don’t need to be because you’re buying in at reasonable prices. I would also be looking at these large cap early stage category, where you’re chasing these companies that can go big. Now, within both categories, there are companies that don’t belong there. They have been misallocated. There are large cap, early stage companies that don’t have the revenue potential, but maybe they have a famous founder or something else, and they get recategorized at that valuation based on the hype. Maybe they’ll live up to it. Maybe they won’t. My favorite is some of the companies that I think should be in large cap early stage, don’t know it, and they raise in the small cap early stage lane, meaning they’re coming in at what we would think are traditional early stage venture valuations or raising capital, and they might just kind of be right before they hit liftoff, and if we can grab one or two of those, they might cross over into a different orbit in the next round. But they can be very compelling when they’re priced right in a small cap or early stage portfolio.
36:38
So, for a company like that, right, when you raise like a big momentum round after a reasonable seed, you’re pricing in a lot of future traction into that. You know what? What do you think needs to be true for a company like that? You know, to if you’re paying a Series A price, or you’re getting a you know a Series B valuation on a company that that’s early, how do you reconcile you know the the degree to which that company has to grow into the valuation in order to you know start to reap some of the rewards of of the early bet.
37:27
I believe that revenue multiples are the best source of rational truth in early stage valuations. There’s not a straight line multiple. The curve looks a little funky. A company doing a million might have a much bigger revenue multiple than a company doing three because the multiple compress a little bit those early days. But let me give you an example. If you come into a company doing a million bucks in revenue, and you have real reason to believe next year they’re going to do 25, not 525, They’ve got contracted revenue. They know their LTV CAC equation. The market’s there. All the buy signal. It’s there. It’s like it’s it’s betable. Well, valuing it on a multiple on that 25 million, in some cases can make sense if you can really underwrite it. And boy, that’s a big step up in valuation because you might have bought in at 20, and now it’s worth 250. So, those are the signals that we can look for and make sense of. You can’t do EBITDA for anyone who’s not a venture player, because in VC, we believe the value of reinvesting retained earnings far exceeds the value of distributing them. So, we do not want EBITDA. We want to reinvest it because every dollar of profit we throw back into the fire should give us 10 bucks of upside in the back end, and that’s a good trade if the venture model is true for that company. So when it’s working, that’s the right trade all day. So you can’t really do EBITDA multiples because if it’s high growth, as we’re hoping it is, you’re always better off reinvesting it, having zero EBITDA, and so we’ll keep the burn up. But we can look at that revenue number, we can look at the margins in the business, we can look at the LTV CAC ratios, and we can triangulate on what I would call a very logical and sane way to underwrite real valuations that should hold up at exit, and that’s that’s I think the key.
39:25
How do you help the the companies that are on a more traditional, let’s say t2 d3, right? Triple twice, double three times. You know, companies that are growing 200, 300% per year, doing pretty well, but
39:40
doing awesome.
39:41
You know, when I send a company like that around to, you know, the the the large large platform tier ones, you know, they just kind of reply thank you, but the growth is not interesting enough. You know, they’re looking for these companies like you mentioned that are going from 1 million to 25 million. Inside of a year, so so how do you help the companies that are good companies, but they just are not measuring up from a growth standpoint? I
40:10
think the challenge is there’s still a bit of market confusion in the investment community about what lane people are in, because outside of the language, which I don’t think I’ve spoken about publicly this small cap early stage and large cap early stage before now. I don’t think people are looking at the market with that bifurcation yet, and VCs and investors really need to know which lane they’re in. So if you’ve got a company like the one you’re describing, which is an A plus in 2015, but today might be overshadowed by these crazy hyper growth companies. There are investors out there that are looking for that, but they may not be the same investors that we’re chasing in 2015. If we segmented the venture ecosystem a little bit more efficiently into two markets, and said, “Hey, look, I’m doing small cap early stage. We’re growing 300% I’m excited. I’m going to buy an enterprise where I can underwrite a 10x on this on the now come. There’s a lot of VCs out there that will make that bet. I’m one of them. Now, when I see the other ones going to the moon, I want to do that too, but in a different portfolio construction, a different pool. You’re taking on in most cases not always more risk, but you have more upside, and so the key is to play the entire portfolio with a similar risk reward profile across all the bets, so that they tend to weight each other out the right way. If you do some of one, some of the other, you might, if you get kind of a bad draw, you’re not diversifying away all the risk you’re taking on the upside plays with the smaller growth potential, so it’s either you’re in this early stage small cap early stage or this large cap early stage, and I hope someone finds figures out shorter language for the two segments. But it’s very clear that it’s divided, and it’s very hard for VCs now to help porkos that we know we look at. We’re like, this is everything everyone wanted 10 years ago, and I’m sending it to folks, and it’s not good enough. But they’re really in a different market now, and there’s a lot of folks smaller capital pools, maybe funds of 100 million or less, not exclusively because you could do the large cap early stage with really small ownership, right? But generally, there’s a correlation between fund size and which part of the market you’re playing in. The smaller funds, there’s a lot of people who want that deal that you’re you’re going to be sharing around. The key is you can’t send it to the same names that you’re sending the other deals to because they don’t do it anymore.
42:35
Mark, what what’s your take on ains or AI native services? Right, lots lots being written about this. We had Jake Saper from Emergence on the show talking about it. Sequoia has been writing a lot about it. You know, the company that does the service with the proprietary AI backend versus sells the tool, the AI tool to all the services businesses, right? So they become the law firm instead of selling the tools to the law firms. Do you think that this becomes sort of the the new standard and the new winning category type of business and venture, or do you think this is you know just another model that’s intriguing but needs a lot of proof points.
43:23
Back to our first conversation, I don’t think extremes work. Anytime you’re hearing a “this is good” or “this is bad, in all cases, you’re probably not finding truth. I think these are viable plays. There are parts of the market where we, I believe, we still need humans. Let me take an example. Our framework for investing in AI law, we think the market has two sides, right? You’ve got the big law, and what customers like our venture firm are really buying when we pay for that contract review and those legal documents is we’re buying an insurance policy. We want to know that someone we can trust reviewed it, and if something goes wrong, we know how to go to. And so that’s an insurance product. If they use some AI in the background, maybe we know about it, maybe we don’t, but we’re going to pay a premium because the buck stops with them. Now there’s a bunch of form filling law, operational processing law, and we’re investors in this wonderful company Alma, and Alma has tackled the immigration law market, and they are growing insanely fast. In that part of the market, where it’s literally filling out forms correctly, it’s easy. There’s still some humans, but they’re a smaller part of the process. It’s easier for the customer to buy that, and the data is more quantifiable because you’re not buying an insurance policy, you’re buying an approved immigration application. Humans in their market tend to get approvals 80% of the time. Alma, I believe I’m speaking out of line here, gets approvals over 99. Percent of the time, because the AI is actually better than the humans alone. That’s an easy decision. So, I think there will be a lot of different scenarios, and I think there’s a lot of things to shake out. But the idea of dismissing either of the models, I don’t think is right. I think again we have to look at how does the market segment out. What are people really buying? There will be places for humans in the go forward market, but it requires a lot of careful thinking. Now it’s not obvious. We have to kind of reevaluate each space on its own merits and in its own nuance.
45:38
Love it. Thanks for that, Mark. Do you have a book, article, or video that you would recommend to listeners?
45:46
Well, I got a podcast. I’m bad at recall. I consume a lot. I read a lot of history. I don’t think anyone else wants to read that, but I have a podcast, as you know, a book out there. But no, I think what I would just say to everyone is keep learning. Once you think you know everything, you need to learn more, and the more you learn, the more you’re going to realize you’re never going to learn at all, and that’s the only way to be competitive long term.
46:11
And Mark, do you have any habits or behaviors that are a secret weapon?
46:17
Nothing too exciting. I’m just highly disciplined and organized. I work out every day. It maybe sounds trite to talk about. I reframe exercise as part of mental health and physical health, which you need in order to be a good steward of your family, a good partner to people, a good boss. So I do look at a more a fully integrated view of what it is to be a person, and I do think all of those pieces feed into being good at the job.
46:48
Perfect. And then here, Mark, any final thoughts that you want to leave with the listeners?
46:54
No, I think I’m a lot of wisdom.
46:56
All right, my man. Well, he is Mark Peter Davis. The podcast is Innovation with Mark Peter Davis, and the firm is Interplay. Mark, it’s it’s amazing to have known you 10 years ago. I saw what you had done then, and to see what you’ve done now, and all the AI that’s built internally, and of course the successful investments. I’m just you know such a pleasure to know you and to have you back on the show. So thanks so much for joining us.
47:22
Thank you for having me, Nick. It’s a pleasure.
47:29
All right, that’ll wrap up today’s interview. If you enjoyed the episode or a previous one, let the guest know about it. Share your thoughts on social or shoot them an email. Let them know what particularly resonated with you. I can’t tell you how much I appreciate that some of the smartest folks in venture are willing to take the time and share their insights with us. If you feel the same, a compliment goes a long way. Okay, that’s a wrap for today. Until next time, remember to overprepare, choose carefully, and invest confidently. Thanks so much for listening.
