Kevin Jiang of Mangusta Capital joins Nick to discuss Investing in xAI, Wiz, and Flexport; Masayoshi Son’s Superpower; How Elon Will Win the LLM War; and Whether AI Is an Extinction-Level Event for SaaS. In this episode we cover:
- Choosing Early-Stage Investing Over Growth Investing
- Masayoshi Son and SoftBank’s Investment Decisions
- X AI and Elon Musk’s Vision for AI
- Vertical AI and Industry-Specific Solutions
- Scalability and Expansion in Vertical AI
- Challenges and Opportunities in AI Adoption
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.
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.
Are you a founder looking for your next investor? Visit our free tool VC-Rank and we’ll send a list of potential investors right to your inbox!
0:18
Kevin Jiang joins us today from San Francisco. He’s the CIO and Co-founder at Mangusta Capital, a VC and growth equity firm whose mission is to shape the world for future generations by supporting technology that moves humanity forward. Prior to starting Mangusta, Kevin spent time at Goldman, Apollo, and most recently, was a founding member at Softbank’s Vision Fund. Kevin has invested in companies including xAI, Wiz, Flexport, Lifeforce, Caliwater, TaxGPT, and Eight Sleep. Kevin, welcome to the show!
0:53
Thank you
0:53
so much, Nick. It’s great to be here.
0:55
Such a pleasure to have you. So tell us a bit about your backstory and your path to
0:59
Vc. Yeah, absolutely happy to do that. Well, I started out in a pretty in the heart of Silicon Valley, actually. So, so I grew up in Cupertino, which many people may know as now the headquarters of Apple. But when I was I was growing up there, it was still kind of not really known for that yet. So my parents were first generation immigrants from China. They came here for college and then grad school, and I grew up just surrounded by everything that was happening in Silicon Valley at the time, a lot of semiconductor boom and bust, and it was a lot of fun. And I got to hear about IPOs and startups and stock options at the dinner table every day, and I just wanted to learn more about it. And so for me, that’s kind of where, where the interest originally came from, of learning about technology and getting into it. And I ended up going to the east coast for college. I studied at Harvard, majored in economics and just wanted to get into business and learn more about what all this is about. And so I started out at Goldman, was part of their banking program, worked in private equity at a firm called Apollo, doing leverage buyouts, and then for the vast majority of my career, was at Softbank for almost the last 10 years, helping build out the vision funds. So that’s the late stage growth equity vehicle and fund that was started by Amasa Yoshi San CEO and founder of SoftBank. As as you and your listeners may know, masa raised about $100 billion around 2016 or so to essentially invest in the next generation of unicorns, technology companies and we invested in everything from Uber to WeWork to DoorDash and coupon as well as bytedance And so, some amazing companies out there, some big stories out there, I’m sure, some headlines that you and your listeners have seen as well. And so it was quite a ride, but it was there from kind of very beginning. I was one of the early founding investment team members there, helping lead our investments across, I would say, logistics supply chain. So I led our investment in Flexport, which is one of the darlings in Silicon Valley, as you may know, one of the largest freight forwarders that takes a much more tech forward approach, and then one in your backyard as well. Shipbob is another deal that I sourced and led sat on the board of as well. It’s one of the unicorns of Chicago, and they do e commerce fulfillment for small, medium sized businesses that want to be able to use something other than Amazon, or if they’re running their own Shopify storefront, being able to have a fulfillment solution as well there. So that’s kind of, you know, the time that I spend at Softbank, and, you know, really grateful for the experiences I had there. Over the last year or so. I met a number of different LPS that and investors that I’ve collaborated with, done just some deals with, and really got along very well with a family office that was that’s an Italian family office behind Luxottica, which is one of the largest businesses out there that no one’s a lot of people have not heard of, but it is the largest eyewear business in the world, publicly traded, 100 billion dollar company that it’s funny that they’ve been able to stay under the radar for so long, but they own brands like Ray Ban, and they also license exclusive contracts with basically every major luxury brand out there. And so when I teamed up with a couple of the folks from that family, I was able to work on some great deals together. And, you know, saw this opportunity within early stage AI investing that, frankly, is out of the scope of what Softbank does. And so that’s when I came up with a thesis of investing in AI applications focused on specific. Industry solutions, and that is essentially the fund that we’ve started, which is mongoose, the capital that Italian family office is now our anchor LP, and we’ve been off to the races for about six months. And so it’s a pretty new journey for myself, but excited to share more about what we’re building and the excitement that we have behind this early stage thesis that we have within AI, very
5:24
good. And give us the profile of this stage that you enter at. And you know, are you leading investments? Are you co investing? You know, what’s your approach at this stage? Yeah,
5:34
absolutely. So we focus on pre seed and seed stage investments within what I call AI applications. Our check size will be anywhere from 500k to about 1.5 million in terms of first checks. And then we’ll continue to double down into our best portfolio companies. So we will end up reserving about half the fund to double down into follow on rounds, say, series A and onwards for the best portfolio companies that we first write first checks into. We will lead selectively. So call it out of a portfolio of about 30 companies, we’ll be looking to lead around a third of those of those rounds that we write those first checks into. I you know from my experience being on the late stage side of things, sitting on a bunch of different boards, you know, 1010, plus different board seats at Softbank, I often see where, you know, board members may not have a bunch of value to add. And so I want to be extremely strategic with both my time as well as the company’s time, in terms of which board seats we do take, which rounds we lead, where we do actually have some strategic value. And so I think one area that makes a lot of sense is in consumer AI, where obviously we have the Luxottica ecosystem that we can lean in. That’s one area that I think makes a lot of sense for us to be leading those rounds, taking a more active role, making more introductions, helping them drive revenue. Opportunities, very
7:00
good. And with your experience doing growth investing at the Vision Fund, you know, why? Why do a seed stage fund? Why not, you know, raise something at Series B or later, that’s
7:11
a good question. You know, I think one of the things I’ve realized from and maybe, maybe a lot of us have seen for the last few years, is, I think the growth stage and the growth market has just become very, very saturated and competitive. You know, I actually think Softbank and maso is probably one of the earliest folks that started on this trend of and created this market of growth investing with the raising of the Vision Fund. You know, after that, Andreessen started raising billion dollar funds. Sequoia started raising billion dollar funds. I mean, Tiger was, I would say, relatively early as well. But they, you know, they really put their full force into private market investing as well, around the time of the Vision Fund getting raised. And so I think over the last call it four or five years, you’ve just seen a ton of capital come into the growth side of the market, and as a result, you have a lot of capital chasing very few deals, and very few good deals at that, I would say you can probably name on, you know, two hands the number of companies that all these, all These growth funds are chasing, because everyone knows, like Stripe, SpaceX, rippling, you know, the large late stage companies that are likely to go public, that are amazing companies with great metrics, Databricks, figma, you know, everyone kind of knows what they are, and so you end up playing this game where it’s Like, okay, who has a slightly better brand name, who has slightly more capital, who can provide a slightly higher valuation to the company. And so it’s a very competitive space, because you’re competing on the size of your wallet and your check writing ability and and also your brand name, to some extent, where it’s very, very small differences, I think, between the different players that are in that space. And so for me, you know, as I thought about, hey, where do I want to play? You know, I just felt like the earlier stage is actually a place you can have a differentiated advantage, where, if you build a great relationship with a founder and they really want to work with you, they may not care that you’re the door. You know, the name of the door says Sequoia or Andreessen. They may just be like, Hey, I love working with Kevin, and I want him to be on my board, and I want him to support me on this journey. And we have a really great working relationship. You know, I think that’s a very differentiated advantage that you don’t really have the ability to lean into at the growth stage. The other thing is, I think there’s a lot more opportunity for these outlier returns on the earlier stage side, when you’re looking at pre seed and seed investing, where, because of the competitiveness of where growth investing is today, you. Know, I think what you really underwrite a lot of these late stage growth opportunities, too, in terms of return profile, is anywhere from maybe a two to 5x return, because these are companies that are effectively public companies. They just don’t need to go public, because private capital is so plentiful now that they don’t need to do that. And so really the upside is, I would say, quite limited, whereas, when you look at early stage investments, if you’re investing in a future unicorn at the seed or pre seed stage, you still have the opportunity for 100x outcome or 1,000x outcome in one of the maybe, hopefully, you have a shot on goal in a couple of the companies near portfolio, and you don’t really see that in the growth stage. And so I think for me, that’s a much more interesting, exciting profile to be investing in, and a space to be playing in where I think I actually have a differentiated advantage that I can bring to the table, where you just can’t do that in the growth stage.
10:57
What would you say to those that that suggest some of the dynamics happening at the growth stage are similar at the seed stage. You know, as these multi platform investors, the andreessens of the world, are overpricing things and being more active participants for serial, you know, entrepreneurs at the early stages. You know, what would you say to those that say, you know, the competition is pretty fierce at early stage as well. Yeah, no, that’s a very
11:24
good point. I absolutely agree with you that Sequoia and Driessen are taking this kind of playbook from, I would say, the late stage as well, to the early stage where, for example, you know, Sequoia just recently backed Vlad, you know, the Robin Hood founders, new company at a pretty crazy, you know, valuation and round size. And I think it’s pre product, pre revenue, but hey, it’s, it’s a amazing serial co founder, you know, founder who’s had a huge exit, right? And I get it for them, the game is a little bit different, and that’s, that’s the game that they want to play. And I can kind of talk about why they’re playing that game. It’s not the game that I would I’m playing, frankly. And that’s really the answer to your question, which is, what is Sequoia and Dresden doing with that strategy? Well, when you raise a billion dollar fund, or multi billion dollar funds, how do you return, you know, a five to 10x on that multi billion dollar fund, if you get, you know, 10 to 20% ownership in some of these businesses, well, then you’ve got to have some decor or 100 billion dollar outcomes right to be able to return that size of a fund, because you’re raising so many dollars, then you have To return, you know, a lot of dollars back. So the strategy for them is, okay, well, how do we find those founders and those future businesses that we can feasibly see, hey, we actually need to take a swing at this. And hopefully this becomes a deck of corn, or, you know, $100 billion exit, public company exit. And I get it, I think for them it makes sense, they’ve got to raise, they’ve raised a lot of capital. They got to return a lot of capital. For me and for a lot of other smaller, emerging VCs, we don’t have to do that, because if we’re raising 50, $75 million funds, getting a 10x return means we need to return $500 million of capital. I only need probably a couple of unicorns at, you know that ownership, you know, whatever, 10 20% ownership, to be able to return that. And so for me, it’s a little bit of a different game. I don’t need to back Vlad and hope that he builds another 100 billion dollar company. I can back a really amazing entrepreneur that I like, say, cash at tax, GBT, and if he exits at a billion dollar valuation, you know, we can still basically return the fund off of that or more, right? And so I think the calculus is a little bit different when you’re looking at a lot of these later stage, or, sorry, not later stage, when you’re looking at these platform shops where they’re actually able to raise billions and billions of dollars, you’ve got to be able to then return billions and billions of dollars. The calculation for us is a little bit different. We don’t need to return billions of dollars. I just want to return a 10x or at least a 10x on the fund, which would mean 500 million plus, which I think we can do with a couple of unicorns. So
14:20
you mentioned masa San at SoftBank. I’m curious, is he as bold as he’s portrayed in the media? You know, there’s so many stories, some I’m sure, are very truthful. Some have reached kind of mythical steps. You know, talk to us a bit about masa and his decision making, and you know what it was like being part of that team?
14:42
Yeah, absolutely. I mean, you know, I would say the media portrays him in a very bold way. And to some extent, it’s true, he is a very bold individual. He makes huge bets, some of which work out really well. You can look at. Alibaba by dance arm. You know, there’s a lot of way, you know, the but there’s also some that haven’t worked out well, right? And so I think in some ways, the media dramaticizes, obviously, who he is as a character in his personality. I will say he’s an extremely charming individual. He has a great sense of humor. He makes Star Wars jokes. He’s He’s a really cool guy, I would say, you know, and I think the media probably takes a little bit of a bashing to him, in terms of portraying him to be, you know, almost maybe inhuman in some ways, but he is really like a human, just a really charming, charismatic human. But he is a very bold individual too, that makes big bets, big swings, and that’s just the way that he thinks about the world. He is a super optimistic person, probably one of the most optimistic people out there in the business world. And that’s why you can believe that, you know, when you take these big swings and opportunities, that you can actually make something work, right? And so, you know, I have a huge amount of respect for him, because he’s obviously built, he’s built a bunch of businesses. So he was an operator and entrepreneur before he became, you know, an investor, and he still is very much an operator in terms of building businesses today as well. So I would say that I think the media doesn’t give enough credit for the ability that he’s had, to be able to build a lot of businesses almost organically, and to actually pick take these big swings and focus on the swings that have actually worked out really well, like arm you know, there were a couple of months where people talk positively about the IPO and people forget that, like he bought this thing for 30 billion, and now I think the market cap is probably over 150 billion. So it’s, you know, he’s, he’s made, you know, 100 billion plus of profit on this one investment. But the media tends to focus on, you know, the not so not so glamorous things, or the not so positive things. And so I wish sometimes the media would focus a bit more on the amazing things that he has been successful at. What was the
17:04
hardest investment decision that you were a part of while you were on that team with masa? Wow, that’s,
17:10
that’s a tough question. I can only speak to the ones that I was directly involved with, and so, you know, unfortunately, I was not involved with we work, and some of the other more kind of flashy stories out there. But, you know, maybe I can talk a little bit about about Flexport, because it’s certainly one that is is really interesting, and it’s gone through a lot of of ups and downs itself as well, right? You know, I think in many ways, they are still doing an amazing job at disrupting the logistics and supply chain space with a technology forward approach. I think, you know, one of the the toughest decisions was when we were doing the deal back in 2017 2018 you know, valuations for a lot of these companies were quite high, and so I remember that the valuation ask was somewhere around, I think, a 3 billion pre or $4 billion post valuation, and we were looking to invest a billion dollars. And so this was when Softbank was at the peak of everything that we were doing. And, you know, 100 billion dollars. It’s a lot of money to deploy, and this was one of the, I would say, core positions within the logistics and supply chain space that we’re really excited about. And so I think at that point it, you know, in a lot of ways, it goes back to how masa thinks about investing is, do you, are you valuation sensitive, or do you believe that this is actually going to be a huge, you know, winner. And so whether you enter today at two or three or 4 billion, doesn’t really matter if this is going to be a 10, $20 billion outcome. And so I think that was probably one of the the decision making processes where I got to see, hey, you know, if you’re investing in a generational business. It’s not going to matter too much whether you entered at two or four. And I think you can see that how that’s played out with say, bytedance was another example that just taken off like a rocket ship, and they’ve done super well too. But one of those situations where, when you have such a large market opportunity in front of you, and you have someone who’s able to execute so well on it, you don’t really you can be a little bit more valuation and sensitive and still make a great return interesting.
19:32
So you’re an investor in x ai, we’ve seen a variety of leaders emerge, creating, you know, New llms, including open AI, anthropic, even meta with llama. Why is x ai positioned to build the world’s best large language model? Great
19:49
question. So, I mean, I think the answer inevitably probably goes back to Elon. You know, as someone who’s obviously the most, probably the most the general. Operational entrepreneur of our lifetime, you know, he’s, he’s just someone who has extraordinary vision for pushing the boundaries of of what we are able to do with AI, whether it’s software as well as hardware. You know, I think that the supercluster that he was able to build recently is a testament to, you know, his focus on really pushing the boundaries on hardware and being able to develop, like, the largest, you know, training model there is, I think it really comes, I think, back to, you know, a bet on any Elon company is a bet on Elon. And for us, I think that’s, that’s really the main thing. I think outside of that, though, you would, you can look at a lot of other data sources that they have which are proprietary and interesting, right? So if you look at the x data that they have, I think that’s super interesting. When it comes to being able to have real time data sources, that’s you can train your model off of you can take, I mean, with Elon’s approach to things, you know, he doesn’t take a biased or woke approach to anything like maybe a Google or llama would. And so I think you have a very unique approach there, along with the Twitter data or x data that you can leverage to be able to train the model. And so I think that’s one really unique, proprietary data aspects and approach aspect. I think the other thing that’s really interesting about Xai is there’s the ecosystem of companies that he has around it, right? Because he’s able to leverage Tesla, you know, one both, because there’s hardware that you need to be able to build. AI, I know that he’s able to essentially leverage some of the procurement vendor relationships that he has from Tesla to be able to leverage that for for x ai as well. But also, in the future, you could think of Tesla cars being a distribution channel for for the AI products, right? You’ve got millions of cars on the road that literally could use the AI products that x AI is coming out with. And so I think across a bunch of different companies within his ecosystem. I mean, you could totally see where neuralink could also come into play in the future as well, as they continue to develop the product and add more AI elements into it as well. And so I think that’s another super unique aspect that, obviously you’ve got to give credit to Microsoft and open AI and the ecosystem they have as well. But I think what Elon has through his network of companies is something super, super unique that you can’t replicate with, you know, autos with, with, you know, game changing, you know, humanoid robots with, with, you know, neuralink as well. It’s just a really vast ecosystem that I think you can’t under appreciate as you think about all the different applications that AI will have in the future. Yeah, it seems
22:51
like there could be some conflicts there. But does he set up like preferred licensing agreements between all these companies? Because they are standalone companies? You
23:00
know, I think the way that, you know, from what I can tell, the way he operates, is, is hard, is just bat, you know, Shoot first, ask questions later, which, you know, it’s interesting to see, you know, especially with kind of the position that he now has with, you know, the government. But, you know, I look, it’s, it’s, it’s kind of the startup mentality, right? You if you want to push the boundaries, you know what? What’s the saying that we have in Silicon Valley? It’s like, move fast and break things, you know, is, is, in some ways, I think what you have to do to really push the boundaries in the way that he’s doing it. And obviously, you know, not to say that we shouldn’t have some guardrails, but I kind of admire, you know, the fact that he’s willing to push, you know, push the boundaries, bend the rules in the pursuit of something greater. Which of
23:51
the other major players do you think is most strongly positioned, and why? Well,
23:56
I think the the natural second answer, if you had to make me make a second answer, would be open, open. Ai, just because, I mean, they have such a great brand, you know, they, they’ve, they’ve built an amazing, I mean, when chat GPT came out, it was, you know, the hottest thing since pancake, right? Everyone was using, everyone and their grandmother was using chat GPT. And I think that’s, it’s valuable, right to have a consumer brand like that where everyone literally knows your product is is supremely valuable, and as a result, it’s also allowed them to raise as much capital as they want. And so I think the other thing that they have going for them is the fact that they have such a great brand is has allowed them to be able to raise globs and globs of capital that, you know, I’m sure x ai, you know, Elon can as well, but it’s such a great brand that I think everyone wants to be a part of, part of that, that story, right? So, if you think of every single large VC firm out there, they all want a piece of that pie, and the Microsoft part. Ownership is obviously super valuable too, right, to be able to provide them with the compute to the hardware, all the ecosystem benefits and fine, and also final distribution, right? Obviously, you know, Elon has a bunch of ways they can distribute, through Tesla and others as well. But, I mean, arguably, Microsoft is probably one of the largest distribution engines out there as well, with Microsoft Office and and all the different products that they have as well. So I think that would probably be the second choice if you, if you had me choose a second a second place,
25:32
I’m going, llama, Kevin, I’m curious,
25:34
what do you like? Well, llama, well, I just think,
25:37
you know, Facebook’s got the similar advantages to Twitter, right? They’ve got some data, and then the open source nature of it, I think that that will not, you know, it’s not just inference. I think they’ll use that to reinforce the training data quite a bit. And so I think that they can develop a lot of interesting models with proprietary data sets because of the way they’re going to market. But look,
26:00
I don’t think there’s just going to be, like, one or two winners. I think there will probably be a stable of, call it five, you know, five or six, kind of major, you know, large language models, some closed source, some open open source. And each will have, you know, kind of a place where it plays, and a specific use case that it, you know, covers, right? And so, you know, I think that’s one of the areas where I think it is going to get more competitive over time, which is why I like investing in the application side of the business, because investing in models is very capital intensive and very, very competitive. So
26:37
I want to get to that piece of the discussion before we transition though. You know, I keep hearing these narratives that AI is an extinction level event for SaaS. You know, what’s your position on this? Yeah,
26:47
look, I actually think in a lot of ways, AI businesses that I see are like SaaS businesses. You know, a lot of the application layer businesses that I see are utilizing existing large language models, whether it’s chatgpt or llama or grok or others, to be able to build their system on top of it, and build their product on top of it, and then they fine tune it with data that they have, and they deliver it in a very SAS Like way, whether it’s D to C or enterprise users, they deliver it in that way as well. And so I actually don’t think that it’s necessarily the death of SAS per se, but really the evolution of SaaS to be aI enabled software as opposed to just simple software. And look, I think there will always probably be some software that doesn’t necessarily need AI either, right? There will be workflow tools that you know work fine the way that they are, and you know, you know, simple applications or user interfaces that you know people will still use the same way that they have. But I think now we’re entering a second revolution of software, or SaaS, where AI has enabled it to do more than what it was able to do before. And I’m very excited about that. Yeah.
28:10
Well said. So Kevin, what sets vertical AI apart from other AI solution providers in terms of delivering value to niche markets. Yeah.
28:21
So for me, I really like this verticalized approach that we’re pursuing here at mongoose, because when you think about the businesses that are able to scale the most quickly, it’s really businesses that understand an industry, understand customers needs extremely well and is able to deliver value day one, right? And so when you use a vertical approach where you’re building a solution for a specific industry practitioner or specific type of customer, you know what they’re looking for. You know what pain points they’re trying to solve, and they’re willing to pay you day one, because you’re able to drive value day one. And so I think that’s something that’s really, really valuable. With our thesis that we’re pursuing is that a lot of the businesses that we are investing in, they may have only been around for three or four months, but they’re already starting to generate hundreds of 1000s, if not millions, dollars of revenues, and that’s because they’ve already found a pain point that an industry incumbent really, really wants and is willing to pay big bucks for. And so that’s one of the really unique advantages that I think we have as a thesis, which is finding those businesses that are driving value today, and ultimately, I think that also leads to faster exits for us down the road, right? If you think of, say, tax GPT as an example, a company like Intuit, which is a public company, you know, multi billion dollar public company, could see value in the product. Tax, or the team that tax GPT has put together. And so I view that as a really great opportunity for, hopefully, you know, potentially a quick exit, if it’s at the right price, at the right time, an M and A acquisition, and that’s something that doesn’t require, you know, going public, you know, 1015, years later, as a whatever it is, ten billion plus deck of corn, to be able to make a great return on a VC fund like ours, you know, an acquisition offer in the 500 to billion dollar range would be a huge outcome for us on that investment, right? And so I think that’s a very unique aspect that we have with this verticalized thesis that you don’t really see in a lot of other VC thesis. Yeah,
30:46
it’s funny. We were going through the math on our last fund and a $550 million exit on any of the 35 logos returns the entire fund. And so to your point, earlier, on these platforms, and just gross age investing in general, it can be tricky to make the math work, but, you know, we’re all in the scale game, and we understand this is a power law. So within the context of vertical AI, how do you ensure that these solutions are scalable while also being being tailored to, you know, a specific industry or a niche?
31:18
Yeah, absolutely. Look, we obviously look for areas where there’s a multi billion dollar market available, right? You know, for example, tax you know, you can look at the data on how many billions and billions of dollars companies and individuals spend on tax preparation software and tax preparation services. For example, longevity and medical health, you know, personalized medical health services. I mean, that’s, that’s a huge one as well. And life force is one that’s tackling that space in our portfolio of providing, you know, more personalized, AI driven insights into anyone’s health to help them live longer and healthier. And so, you know, we are always looking for these spaces which have multi billion dollar, if not trillion dollar markets to tackle. But I do think that it gives us the leeway of saying, Hey, if you’re building a very specific solution or starting with a very specific application within one of these industries, you can scale it to, you know, we want to be able to see that you can feasibly scale it to couple 100 million dollars of revenues, and then from there, you have different options available, right? You can either expand so you can continue to say, Hey, I’m going to build out like toast. Did you know In the vertical software playbook, I think of this as, like the vertical AI playbook is you go to another different, you know, solution or product within that space. So you start with payments, and then you go to labor, and you go to, you know, online ordering or inventory management, whatever it is. And so you can continue to expand the pie and go into other product areas where you bring that 100 million dollars to maybe a billion dollar revenue business. Or you have the option of saying, Hey, I’ve built an amazing application in this particular area. I can find an M and A exit. Or we can stay, you know, the standalone company, and find an opportunity to exit at still an amazing outcome, right? And so for me, it’s, it’s having that optionality of, yes, you’re tackling a huge market. So you can continue to expand into other products if you want. But if you, you know, happen to build a great thing and you think you can get an early exit, go for it. So you you’ve got
33:33
the unique and rare experience of being at one of the preeminent growth stage investors and and running, you know, a very sort of AI focused early stage fund. Question I have for you is on this expansion topic, right? So we’ve had some success in in vertical SaaS and often AI enabled. And once a company has success with customers in a space, it kind of earns you the right to expand, your customers begin to trust you. They begin to rely on you. They begin to ask you, can you also, do, you know, inventory management, you know, in addition to payments or what have you so, how do you weigh the decision to expand, you know, across maybe product sets with just double down, you know, doubling down on focus and just remaining best in class and driving volume and sales within, you know, a core product segment.
34:30
Yeah, no, it’s a really tough question. You know, it’s a very interesting one, because I It reminds me of one of the conversations we had around the board with one of the companies that I was I was involved with at Softbank, that was in the restaurant space, and they had started from more of a labor and scheduling labor employee management solution, and they were thinking about going into payments, payroll and other other products as well. And I think the answer. Your question is, it kind of depends. It’s case by case. It really depends on what is the current product and what is the ROI on the spend to continue investing in that product, versus where are you seeing returns or green shoots from other products that you’re thinking about rolling out, because obviously it takes a lot of effort to roll out a new product, right? You’ve got to do spend R D engineering resources to build that out, but then after that, you’ve still got to spend money on acquiring customers and marketing. And you know, those customer acquisition costs also add up pretty quickly if, if you don’t know what the LTV or lifetime value of that customer will be for the new product, right? And so I think it’s very hard to give you a blanket answer for that kind of trade off, but it is really case by case, and you’ve got to evaluate where is that dollar of spend if I’m putting it into what is the return on that dollar spend if I’m putting it into an existing product versus a brand new product that I’ve got to engineer and build and then, you know, market and distribute again. And in some cases, it makes a lot of sense, because there’s a poll, right? Sometimes a customer comes to you and says, Hey, I really want this product. I’m willing to pay you XYZ amount, and you know that the amount of time it’s going to take you to develop the product is relatively minimal, or the investment is going to be reasonable. And so in some ways, you’ve got that first marquee customer, or a couple marquee customers that will support you through building out that product. And in that case, it may make sense to invest in that new direction. But in a lot of cases, if there’s not that pull sometimes, I think the ROI, the known ROI of investing in your core product, where you still have a great say, LTV to CAC ratio, you know, can be a fine path to continue going on if, if there’s not another you know, area that pulls you in.
37:00
So, Kevin, you mentioned GPT, you know, financial software. You mentioned healthcare. What are the industries that you see as the next frontier for vertical AI, and how are you preparing to, you know, lead on those at Mengistu,
37:15
yeah. So I think the way that we think about industries which are going to be disrupted by AI and specifically application layer technologies that are going to make a huge difference are, what are the industries where there’s a lot of manual labor in terms of hours, but also high cost of labor, and I think from the automation perspective, at least, starting from that, from that lens, and a lot of them are industries that you’ve actually already seen a lot of development in, in terms of AI companies and technologies emerging. For example, legal is one that we’ve seen a ton Harvey case, text, you know, Paxton, these. These are luminance. I’ve met a bunch of these companies when I was at Softbank, and you could see a ton of development and funding going towards these companies, because obviously legal and is a very manual, very high cost industry for for professionals to be working in. And, you know, the spend on legal services is so high every year, software coding, you’ve also seen a ton of, ton of funding go into this cognition, you know, Dev and so. So there’s a bunch of different, you know, companies as well in that space that, you know, I was tracking and continue to track. You mentioned healthcare before another here. Yeah, there’s, there’s, I mean, I there’s, there’s a number of larger names, but also just smaller names I see popping up helping with, like, payroll, billing, collections, all those types of things. And so, you know, for me, it’s you’ve got to look at, obviously, the bigger picture of, like, okay, which are the industries that we’re excited about, but then you’ve also got to look at which are the areas that haven’t been super well funded yet, or, you know, haven’t become super competitive yet. And so while coding and legal, I think are make a lot of sense, text in, text out, and I think there’s going to be some great companies coming out of there. I feel like, unfortunately for us, it’s become a bit more of a growth stage game in a couple of those sectors. So maybe we’ll find something really interesting in the early stage but, but those are ones that, you know, I saw a couple years ago when I was at Softbank, and I remain excited about but maybe is a little bit too late stage for US healthcare. I think we’re still very much in early innings, whether it’s personal health care or enterprise application solutions. I see a lot of new startups coming up there. I think it’s really interesting. And I you know, we do want to make a play in that space. On the more enterprise side, we have it. We’ve made a couple of plays on the personal side, for example, like eight. Sleep as well as life force, where we think that AI is going to drive a lot of personal insights and really, much more understanding of the data behind our health. I think the other area that I’m really excited about is fintech. I think it’s an area that we haven’t really seen a ton of AI solutions yet, and application solutions specifically. And I think when it comes to KYC fraud, I think those are things that one attackers are getting more sophisticated with, using AI to do impersonation, to do other cybersecurity attacks. And as a result, I think you’re going to see a lot more, companies also have to tackle that from a Fraud Management and identity impersonation perspective as well. And so that’s another area that I think we’re still relatively early innings on, and haven’t had a ton of funding flow into that space. And I see a couple, you know, I see some companies emerging there that I’m very excited about. And so for me, it’s really interesting because, you know, obviously really excited about the applications for these huge areas that I think, like legal or, you know, tax, that will, you know, obviously be great. But also it’s kind of finding those new segments and new niche areas that people haven’t really tackled yet, and that’s really where I want to be playing, is to be ahead of that curve before it gets, you know, before it gets too hot.
41:28
You know, in reading through your materials, you had highlighted a bunch of these categorical things that you just reviewed, like cost of labor and labor constraints and workflows. You also mentioned dark data. What does dark data mean? Sorry, I don’t remember where was that. It was in one of the articles you sent to me. So let’s see here.
41:47
Oh, it must have been from our blog, huh? Yes,
41:51
we can pull that out if, yeah, maybe we pull that resonating, yeah,
41:56
I think it was. It must have been with one of the analysts that I had worked on the one of the blog articles with but to be honest, I’m not entirely sure what that means. I
42:07
mean, my assumption is it means data that exists but is not being collected. So like the aura ring, it could be like, there’s all this biometric data that’s like, hidden, but it’s there, yeah and yeah, so it’s susceptible
42:24
to, like, breaches or being used in like, malicious ways, essentially, wow, maybe, yeah, I think that’s, that’s probably what, what it’s referring to is, essentially, you’ve got data that exists but may not be secured and may not Be. You know, is not meant to be kind of publicly used or used for other purposes, and then it ends up being used for malicious purposes. Got it
42:49
interesting? Okay, I’m gonna give you an impossible question here. Kevin, great. If you could only invest in one of the following categories, AI, co pilots, agents or AI enabled services. What do you choose and why?
43:05
That’s a really interesting question. Yeah, it’s an impossible question, because, to be honest, I think different industries and different use cases require each of these different delivery methods, and so I think, you know, broadly, I would say we want to invest in all of these. But if you made me choose one, I think it would probably be AI agents, because I think that I think about AI agents similar to API polls, where, if you’re able to, you know, I just think of it in terms of like a business model that’s just like the most beautiful, kind of elegant business model, in my opinion, where you’re able to deploy an AI agent that essentially is able to self run itself to accomplish a task, right? And because of that, you’re able to monetize off of this, you know, whatever it is, beautiful algorithm that you’ve built and you’ve constructed, I think that, in my opinion, is kind of like the most elegant, perhaps, business model of the three versus, when I think about AI copilot, it’s kind of like, okay, you have this, you know, chat bot that you’re using, you’ll pay a subscription fee to it, you know, whatever it is, for example, like chat GPT, $20 a month. You know, maybe you can charge more if it’s a more enterprise or professional grade product for, you know, specific industry specific use case, and so that, you know, is not usage based. And so I think, in some ways, is like more limiting in terms of the revenue opportunity there and then on the services side. You know, at least the way that I’m interpreting it is, you know, you still have kind of a human in the loop that’s delivering the services. And so the margin can. Ever truly be, you know, really like a software like margin, because you still got to have some humans in the loop, delivering, you know, helping deliver that service. And so, if you just thinking about it from like a business model perspective, in my mind, the first answer is the AI agent, one perfect.
45:19
Love it. So, so you and I are speaking it’s end of 24 this will be published in 25 and we’re at about 50% cloud software, enterprise adoption as we speak, as you consider AI adoption in the coming years, in the enterprise, where do you see adoption in five years? So in 2030
45:38
I mean, it depends. It depends. It depends what you mean by AI adoption. But my assumption is that everyone, every company, will be using AI in some form or fashion in maybe even earlier than five years. To be honest, I think we see, I mean, every single public company out there is talking about AI, and it’s become just a topic that you have to invest in, right? You have to you have to explore. You have to try it out. You know, whether or not you know, even if you’re running a restaurant, you’re a restaurant business, say you operate, you know fast food chains, you know they’re experimenting with drive through AI agents to be able to take orders in the drive through, like every single business has areas that they can use AI. And so my assumption is, before five years, I think every business out there will be using AI in one form or fashion, whether or not it’s driving, you know, a majority of the value to the business and where it’s driving value for the business is probably a larger question. But from what I’ve seen, everyone is trying to use AI in some form or fashion, even today.
46:49
Kevin, if we can feature anyone here on the show, who do you think we should interview and what topic would you like to hear them speak about?
46:55
And so I’ve got to answer this with masa, because I think he would be an amazing person to for you to interview. And I think it would just be, I mean, it would be talking about some of the lessons I think he’s learned along the way. And I think, you know, if he were able to be fully transparent and open about everything, it would be fascinating, I’m sure, and probably take hours and hours to talk about, but also, I think, his view of where AI is going to eventually go towards, and what he’s most excited about, what companies he’s most excited about, what developments he’s he wants to see and and what he’s building in the future. You know, I think he’s, he’s someone who’s, unfortunately, he probably can’t be fully open with everything that he’s doing, but I think he’s someone that’s always ahead of the curve, and so I’d love to hear a little bit more about what he’s building in the shadows.
47:52
Kevin, what book, article or video would you recommend to listeners? So
47:56
there is a book that I actually recently read called chip war, written by Chris Miller, and it’s about, you know, the title is chip war, the fight for the world’s most critical technology. And so it’s a super fascinating read about even going back to, you know, the 80s and 90s, when the first semiconductor companies popped up, to where we are today, where there’s a huge resurgence, obviously, in the importance of chips and semiconductors within this AI ecosystem that we’re now in. And so it’s just the really amazing, I mean, simply for a history lesson, to be honest, but also just storytelling about everything that’s happening, and then also the the international geopolitical considerations of everything that’s happening too. So that’s one book I would highly, highly recommend that you read. Awesome
48:50
Kevin. Do you have any habits, tactics or behaviors that are a force multiplier?
48:55
Yeah. I mean, it’s, it’s tough I have. I try to take on as much of these as I can. I think one is, like, the most important thing when you have a team and is building a great team that really supports you. And so I’m super grateful for a lot of the team members that I brought on with Mengistu that are able to really force multiply. Me, quite frankly, running, running a fund, is not an easy task. And so I think that’s that’s number one. I think the other thing that you know I learned from one of my previous bosses is actually to, I don’t know if you’ve heard of it, but I think some of your listeners may have it’s like the the four quadrant system of focusing on what is most urgent and most important. So if you have a quadrant of urgent versus non urgent, important versus non important, you know, finding that fourth quadrant of things that are both urgent and highly important, that you need to focus on first is one of the kind of tactics I think about in terms of prioritization, perfect.
49:59
And then finally, here, Kevin, what is the best way for listeners to connect with you and follow along with Mangusta
50:04
Absolutely. You should check out our website. Mangustacap.com on LinkedIn as well. I post pretty regularly, and if you know you’re building something interesting, feel free to shoot me a DM or send a message on our website. Awesome,
50:21
Kevin, thank you so much for the transparency, for sharing these great stories about masa and Elon and Flexport and many others. Appreciate the time and best of luck with the fun at mangusa.
50:32
Thank you so much. Nick has a pleasure being on the podcast, and hope we can do another one soon. Awesome.
50:38
Thank you, sir. You
50:45
all right, that’ll wrap up today’s interview. If you enjoyed the episode or a previous one, let the guests 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 over prepare, choose carefully and invest confidently. Thanks so much for listening.
