I sat down with LionTree’s Antal Runnebom. I gave my perspective into the opportunities around vertical-specific marketplaces; the process by which FJ Labs selects the hundreds of start-ups we invested in; why humanoid robots and AI-powered investors are coming faster than any of us realize; and a vision for how future energy abundance will enable us to “waste energy” inventing the next cool thing.
If you prefer, you can listen to the episode in the embedded podcast player.
In addition to the above YouTube video and embedded podcast player, you can also listen to the podcast on iTunes and Spotify.
Transcript
Antal
Welcome to KindredCast. I’m Antal Runneboom. On the occasion of our 200th podcast, we’re diving into the world of early-stage investing, focusing on marketplaces, software, and other forward-looking technologies. We’re thrilled to speak today with Fabrice Grinda, founding partner of FJ Labs, a stage agnostic venture fund headquartered in New York City.
Fabrice is a serial entrepreneur and super angel. In 2006, he co-founded OLX, a global digital classifieds marketplace disruptor that was subsequently sold to Naspers where he stayed until 2014.
Antal
Since then, he’s been on a mission to revolutionize the marketplace and technology ecosystem through FJ Labs, where he’s led investments in over 1100 startups globally. Some of Fabrice’s investments have included Alibaba, Airbnb, FanDuel, Palantir, Beepi, Vinted, Quince, and Figure AI, amongst others.
Antal
Today we’re going to explore Fabrice’s entrepreneurial journey, his investing philosophy at FJ Labs, and his thoughts on the future of marketplaces, AI and B2B startups with insights into his key portfolio companies. Fabrice, it’s an absolute pleasure to have you here on KindredCast today.
Antal
Let’s start with your background. What first brought you to the US and into the technology sector?
Fabrice
So I grew up in Nice in the south of France. And in 1984, at the tender age of 10, I got my first computer. And it was love at first click. And from that moment on, I knew that computers and I were going to be together forever. And the thing is, it wasn’t obvious to me it was then needed to go to the US, but clearly in France in the 1980s loving computers, I was N of 1. I was an extraterrestrial. There was Minitel, there were a few things in tech, but it wasn’t really tech forward. And I skipped a few grades. I was top of my class. I was winning all the Olympiads. And I was building my computers, programming, building BBSs, and starting to think about what to do for college.
And I went and interviewed at the best French school, which was Lennar. And I went to speak to the main professor there. And he was like, « So what do you want to do when you graduate? » I’m like, « Well, I want to follow in the footsteps of my role models Bill Gates and Steve Jobs. » And he’s like, « What? You would betray the ideals of the French Socialist Revolution? » And I thought he was kidding and I burst out laughing. And of course he wasn’t. And I’m like, « Okay, perhaps I belong in the US, American Dream here I come. » And so applied for colleges. I mean, it was hard back in the day, because there was nothing online. I had to go visit them, pick the applications. I didn’t even know you could study for the SAT. But went to Princeton and the rest is history.
Antal
Amazing. And so, one thing that’s interesting about you is that before OLX, which is probably the most well-known business that you founded, you had already started two companies prior and sold them. Tell us a little bit about the formation and foundations of your entrepreneurial journey that ultimately led you to found OLX and then sell that again to Naspers.
Fabrice
So went to Princeton, finished off my class. Studied mostly economics because, to me, it explained the way the world worked. Upon graduating, I was like, « Okay, I want to be a tech founder and there’s a bubble going on, but I’m this shy, introverted, 21-one-year-old, let’s go to McKinsey. » It’s business school except they pay you. And from there, I’m going to learn oral and written communication skills, public speaking skills, and it’s going to make me a better businessman for when I go build my company. Two years in, I felt I’d learned what I needed to learn. I was 23, it was ’98. I’m like, « Okay, let’s go build tech companies. » Now the issue is the constraints at the time were very different. If you want to build a tech company in ’98, you need to build your own data center. You need millions of dollars to turn the lights on. And as a 23-year-old, it created complexities.
And so, as I thought through, what is it I want to build? Well, I was an economist by formation. I studied market design. I had fallen in love with the model of eBay. I felt it was acid light, a winner-takes-most and extremely scalable. And I’m like, « Look, it has its own complexity of matching supply and demand and fixing the chicken and egg problem. But this is a problem I’m perfectly suited for solving. » So I decided to take the idea, bring it to Europe, and also bring it to Latin America. And that was the beginning.
Antal
As a firm, we spend a lot of time in digital classifieds and marketplaces, and these categories have evolved tremendously. You’ve had the emergence of software-enabled marketplaces, we’re going into a world of AI agent-driven marketplace models. But back then in 2005, 2006, when you first started scanning the market for what to do next, you basically had a bunch of newspaper companies that were building digital portals and trying to sell classifieds online. And then you had the emergence of Craigslist here in the US, but it was a pretty wide-open field at the time. And so, take us back to that foundational moment and what got you, how did you see the opportunity in that moment?
Fabrice
Let me go back to what happened in the prior two companies because that’s what led to OLX. So the first company, which is in eBay of Europe, I grew from zero to 10 million a month in sales, leader in four or five countries. Ultimately grab sadly the feet from the jaws of victory. As the company who bought us, their stock fell 99.98% QXL Ricardo. And even though it had gone from zero to hero, but back to zero again. So, in 2001 I was like, « I want to be a tech founder. » And my second company I built was not one whose products I loved. It was like, « Okay, I need to build a company that is going to be profitable very quickly because there’s no capital available. And that’s why I built a ringtone company. So Zingy went from zero to 200 million revenues in four years. It was a huge success, it was super profitable, but I didn’t have a particular affinity for it.
And so, I sold the company in 2004, we did 50 million revenues in 2004, for 80 million in cash. I’d learned my lesson from last time, better cash than stock. Actually hired bankers, which ran a process, doubled the price, it was amazing. But I knew I wanted to go back to my true love of building marketplaces. And Craigslist was coming of age. So in 2005, as I knew I was going to leave Zingy, I went to Craig and Jim Buckmaster, who was the CEO, and I’m like, « Hey-
Antal
Both still there.
Fabrice
And they are still there today. In fact, the site hasn’t changed all that much. So I went to them and I’m like, « Look, what you’re doing is providing an amazing public service to humanity by having this free site that’s allowing to match buyers and sellers and products who are finding jobs. It’s amazing, but frankly, it could be a lot better. You could actually monitor all the content and make sure that there’s no scam, spam, phishing, et cetera. So why don’t we moderate everything, make sure we create the highest quality content possible? And I realized you may not be willing to do that, but I’ll do it for free. Let me run it. You don’t need to pay me anything. If you like what I do, maybe after a year we can discuss you giving me some equity, but I’ll do it frankly as a service to humanity. » They said no. Then I said, « Okay, I’ll buy it from you. » And I offered, and that’s pre- it being very well monetized. I said hundreds of millions, I think maybe even billions of dollars. And working with another private equity company, they’ve also said no.
So, I’m like, « You know what? Fine. The best competitors to have are people that don’t want to compete and don’t really care. » And they were like, « Look, if you do something great, fantastic. » So decided I was going to go after Craigslist and build a better classified site catering to women, which ultimately are the primary decision maker in all household purchases and moderating all the content, created the safest space possible. And so raised all the VCs that had wanted to fund me in the last company after we were profitable. And I turned them down because, of course, we were profitable, and they had not been willing to fund me at the very beginning. Then were falling over themselves to fund me. So I got a $10 million essentially pre-seed check, which in 2006 is extraordinary on a PowerPoint from General Catalyst, Bessemer, Founders Fund. And I was like, « Okay, let’s go and build this. »
So, Build OLX. Launched in a hundred countries. When you build a marketplace, you want to start by going to the sellers because the sellers are financially motivated to be there. So I told all those people, « Look, we don’t yet have liquidity, but we’re free. Why not us? And see what happens. » And so I went to car dealers, real estate brokers, especially in countries without MLS, to people that were listing inventory elsewhere. And then I doing SEO and buying long tail SEM. So we spent 50K in a hundred countries, which is $5 million. And it just so happened it really, really, really worked in four. So it really, really worked in India, in Brazil, which of course creates massive economic value, and really worked in Portugal and Pakistan where essentially overnight we became dominant leaders.
Because of the dynamics in these marketplaces where you want to be hyper local, you want to create liquidity, we decided, you know what? It doesn’t really work in the US because it’s too expensive to acquire customers. Craigslist already has liquidity. It didn’t really work in most of Western Europe because Schibsted and Adevinta were already there. Even if you’re a better product and mousetrap, liquidity trumps everything else. So we’re like, « You know what? We’ll double-dime on where it works. » After haven thrown the spaghetti on the wall, said, « Let’s do India, Brazil, Portugal, Pakistan. » Hyper focused there. Became the dominant players there, became very big, profitable. And then we decided to then use this capital to expand to 30 other countries.
So even before the Naspers deal, we’d grown over a hundred million unique visitors, hyper-dominant in most of the emerging market. And frankly, I was very happy and everything was set. And I thought, « Okay, next 10 years are pretty clear in terms of we do CDC transactions, and we dominate cars and real estate, then perhaps jobs, and we go transactional. And there is a hundred billion company to be built there. »
Antal
Yeah, but then fairly quickly it turned out differently and you sold the company to Naspers.
Fabrice
So OLX is the company I wanted to build and run my entire life. We were part of the fabric of society in the countries we were in. Now, we have 300 million unique visitors a month, like 50 million people make a living off the site. It is profoundly impactful in the lives of people, especially in countries where payment systems and trusts and shipping systems are not working particularly well. What happens to my great chagrin is Schibsted, which is the publicly traded European competitor, did a joint venture with Telenor, and they attacked me with hundreds of millions in TV spending in my core countries, especially Brazil and Portugal. In Brazil specifically, they spent, in year one 60 million, year two 80 million, year 160 million. So hundreds of millions of dollars.
And we had raised at this point, 28.5 million. And while we were dominant, we had liquidity, little by little they were catching up. And at some point, it was like 60-40 for us still, even though they’d spent hundreds of millions. And I went my American VCs in 2010, and I’m like, « Look, I have bad news. Last year we were very profitable and did fantastically well, but now we need to stop monetizing, spend a hundred million on TV and do it across multiple markets. » And had this conversation happen in 2015, perhaps with DSC or Tiger Global or others, we might’ve been able to do it.
Antal
You would’ve gotten the funding.
Fabrice
Yeah. But in 2010, going to American VCs and telling them, I need hundreds of millions to spend on TV in Pakistan, it didn’t go over super well. So the people that I did manage to convince that this made sense were Naspers. And Naspers were interested in going into the classified market. And in fact, they were the perfect partner because as aggressive as I was, Kuss specifically was even more aggressive. I would go to him and like, « I need a hundred million. » And he would go back to me, he’s like, « No, no, you mean you need 300 million. We are bigger, we are stronger. We are faster. We are going to win this war. » And so, they gave me almost $1 billion to go fight the war with Schibsted. Plus, and that’s not even including the acquisitions of Avito of Dubizzle, et cetera. We ended up winning the war. We merged 51-49. And yeah, things were settled.
And these businesses, once you’re the dominant player are extraordinarily profitable, I mean we were talking like 75% EBITDA margin, and that’s even before going transactional, et cetera. So the deal with Naspers of 2010, they gave me all the money. I ran it for three years after the deal. They hadn’t bought me. It was just fundraising to win the war. And then 2013, I’m like, « Okay, we’ve won. Things are pretty good, but it’s boring to be part of a big publicly traded company. I’m no longer playing with product. My life is creating an annual report, then a quarterly report, then update the quarterly budget and stuff. Let’s go back to the basics. Let’s go back to zero. » And so, I left in 2013, in a way at the peak of the performance of the company, and it’s continued to do extremely well since.
Antal
Amazing. And these businesses are still dominant in many of their markets. Very strong in Eastern Europe, very strong in Brazil, very strong in other economies.
Fabrice
Yeah. In Southeast Asia, yeah. Leader in the UAE, leader everywhere. The only downside, the most profitable crown jewel of the company, Avito, which is a billion in revenues, 750 EBITDA, essentially stolen by Putin and his cronies. And that was paying for the aggressive growth and in India and Pakistan, et cetera. So the companies were then sold to other players who could fund the continued growth. They’re still dominant and they’re still doing extremely well, but they’re no longer all in the OLX umbrella, because the cash cow was taken away, sadly.
Antal
Let me just turn briefly to the next step of your journey before we go deeper into a few of these topics. So after Naspers and OLX, you changed your modus operandi from there and became really an advisor and an investor behind many, many, many successful entrepreneurs. What led you to pivot from being the builder, so to speak, to being the person who’s the person helping the builder and helping back them?
Fabrice
So, I would argue nothing has changed all that much. So back in ’98, when I built my first company, I was already a visible internet CEO. And as a result, other founders started approaching me saying, « Can you invest in our companies? » And I thought long and hard, should I be an investor in parallel to being a builder? Because it is a distraction from my core mission of world domination. I ultimately decided, look, if I can articulate lessons learned through others, it makes me a better founder. I’m running these multicategory horizontal sites. If I can actually meet all these founders running the verticals and keep my fingers on the pulse of the market, it makes me a better founder. So as long as it’s not too distracting. So if it’s in a one-hour meeting, I only do marketplaces, I decide if I invest or not, then it’s fine.
So by 2013, I’d already invested in almost 200 startups. And in fact, in the public eye, I was known more, despite the fact that I was running one of the largest websites in the world because it’s not big in the US or in Western Europe, when I went anywhere, people knew me as an investor and a super angel, not as a founder. So I was already a founder-investor. And the path I chose of building FJ Labs, I’m actually continued to be that, because FJ Labs itself in a way is a startup. I didn’t join a fund, I created a fund. And creating a fund is the same thing. You need to raise capital, hire a team, build a structure, have in effect a budget, and a business model and a strategy. And within that fund, I continued investing and I continued building companies. So even there, I’ve been executive chairman, co-founder of multiple companies, keeping my hands busy. So I’ve always been an operator-investor, and I continue to this day to be an operator investor.
Antal
Fantastic. The piece I wanted to just chat about at the moment, which is very topical with our clients, is the advent of AI within marketplaces. How is this dynamic going to go between the ultimate customers and the platform in this AI agentic world that’s coming up?
Fabrice
Let me give you first a little bit of evolution of marketplaces and then talking to how it leads into how AI is impacting that.
Antal
Perfect.
Fabrice
Marketplaces started as basically a la OLX, Craigslist listing platforms where you list what you’re looking for or you list what you’re selling and someone contacts you. These are called double commit marketplaces. Upwork is one of these like, « Oh, I’m looking to hire a programmer. » 200 people apply. You need to do the work to select. Same thing with OLX, if you’re listing something for sale, people contact you. You’re only going to sell to one person. There’s a long conversation. And it’s a low barrier to listing, but actually it’s a high barrier to transacting. It’s a lot of work. The next thing that happened is we had transactional marketplaces. So the buyer could say, « Okay, I want to buy this item. And that was done. The next evolution, people had called it managed marketplaces, but the better terminology, and it’s more effective, especially when it comes to services marketplaces are marketplace pick marketplace. Meaning the marketplace picks the supplier and makes the transaction happen.
When you go in Uber, you only say, I want to go from point A to point B. You don’t pick your driver. Uber picks. It’s not supplier pick either. It’s not as though they send the job to 20 drivers and the first one clicks gets it, they send it a one, and that person has a few minutes to take it, if they don’t take it, someone else gets it. So this is a mega trend that is happening right now in the disrupting the Upworks vertical by vertical, where now if you want to build whatever, replace your heat pump, there’s a company called Tetra, instead of going to AngelList or another one that you just take the photos, they actually do the pricing, they pick the contractor, they set the price, you’re done. So it really improves the user experience. So that’s been a mega trend.
The other big mega trend is services have been added on top of categories in order to create the category de novo. So, we’re investors in a company called Alpaga, which is a B2B marketplace for restaurant equipment. Restaurant equipment, basically people would wait until it depreciated to zero and would not change it. But now Alpaga said, « Okay, we have created this network of installers, a network of shippers to ship the equipment, and we will give this as a pass through. » And as a result, the category’s gone from zero to millions in GMV. And this is happening in many categories where there were no marketplaces. So, these have been the mega trends.
Now, the way AI plays into that role is, I’d say twofold. One is, seller flow and selling flow. And then I’ll talk about the buying flow. So, in the selling flow, every major marketplace is using AI to simplify the posting process. So, imagine the old days of going to eBay. You take 20 pictures, you write it, you select the category, you write a title, you write a description, you select the price, and then you wait patiently for someone to come and buy it or bid on it. And it’s a fair amount of work. And in fact, the eBay process is way more complicated than that because then they ask you how much it weighs, where you want to ship it, et cetera. And a lot of this, you have to guess.
The new model, and it’s better in verticals than horizontals right now. And often disruptive incumbents who have the data that can do it but is actually you take a few photos and because the marketplace has all the data, they tell you the name, the model, the category, the price, et cetera. So I’m an investor in a handbag marketplace called Rebag. Rebag, of course, has all the data. They’re the Kelley Blue Book of handbags. They know the condition, the model, et cetera. So you take a few photos, they’ll tell you, « Okay, this is a, » I’m not a handbag expert, but whatever, « A Kelly, it’s worth $10,000. It’s this category, it’s this condition. » So, it creates the listing, gives you the price, and boom, you can sell in five minutes. And so, the selling flows are being improved. Like TCG player already has a, they can take a photo of a trading card, instantaneously it will create the listing, will know the price and will list it. There’s a company called Col-X doing the same thing in collectibles. So the selling flows are massively improved vertical by vertical.
It doesn’t work in every category. If I’m trying to sell an iPhone or a computer, you can’t tell from the photo what storage capacity it has and what model exactly it is, et cetera. So the way people are doing it is you describe the item by voice, then you send it to OpenAI Whisper, they’ll transcribe it, tag it, and then create the listing. So there’s a company called Herosoft doing that.
On the buy side, there are three buying flows in marketplaces, and AI is disrupting only one of them. And so, most of the buy side of marketplaces will not be disrupted because it doesn’t need to be. So the three categories, the three purchase mechanisms or marketplaces on the buy side are browsing, so shopping as an experience. So people who go to Vinted don’t actually know what they’re looking for. They actually like going through 20 pages of items because it’s like shopping as entertainment, and they may or may not buy something. And so you don’t want AI, because AI will not optimize anything. You’re not trying to optimize. The experience of going through hundreds of items is the actual point of the experience. So AI will not change anything there.
Number two, search. You know what you’re looking for. And search, yes, you can improve it, but for the most part if you know what you’re looking for, you go to Amazon, type the name or the model, poof, you get it. And so again, AI will not fundamentally change that. The third category though, considered purchases, and this is where the Zillow comes in, if you’re trying to buy a car, if you’re trying to buy a house, if you’re trying to buy high-end ski equipment like on Curated, if you’re trying to go somewhere and travel, you don’t know the place, something like Fora Travel, there I think the agents are going to play a very important role in guiding you and changing the category and facilitating the transactions. So that’s why I think a company like Curated, which was using human advisors sold for not that much because there’s a very real risk of disruption by AI. And so there I think, for to your point, Zillow, so real estate purchases or car purchases, I suspect that the AI agents will play a very big role in improving transactions.
Antal
And those platforms are likely to become more valuable over time as a result, I would have thought.
Fabrice
So yes and no, right? It’s unclear, sometimes they come and do take advantage of the data they have, sometimes they don’t. So if you’re a disruptive incumbent, ala Rebag, so still a small startup, probably. If you’re eBay or Zillow, less likely. Especially since … Would I rather be eBay Motors or Carvana? And Carvana probably can do a better job at guiding you, making the transaction happen, et cetera, than eBay Motors. So unclear from that perspective.
Antal
Often marketplaces are known as playbook businesses, where you’re really making certain improvements to drive the liquidity, the two sides of the marketplace, et cetera, then you ultimately end up with winners that are dominant over time. We’re obviously in the M&A business, and one of the biggest comments that often comes is the lack of synergies in marketplace consolidation. But there are operators who’ve done well in terms of using playbooks that they’ve learned about to create value when they buy assets that are maybe a little bit further behind.
What’s the lessons learned around marketplace M&A that most people may not see? Why does a Naspers lean in so heavily into what have effectively been M&A driven creation models?
Fabrice
Within the same category, M&A is extraordinarily effective, because marketplaces are really winner-takes-most. For many categories, if you’re a seller you’re only going to list in one place because you can only transact once, and so therefore more sellers bring more buyers and creates a winner-takes-most category. So if you’re in a country, for instance, or any vertical, any geography where you have two players, it doesn’t matter really where the market share … Well, you want it at 60/40 or 50/50 or 70/30, it’s actually pretty terrible because you can’t monetize.
And this is really a category, so if you think of game theoretical terms, the dominant strategy, so if you don’t spend and they spend, they win. If you spend, they don’t spend, you win. So both of you spend, which you end up in a negative outcome where neither is monetizing and you’re spending a lot of money in marketing with negative margin. If you merge though, it’s one plus one equals 10, everything improves. The liquidity improves, the customer acquisition costs go down. And so we’ve seen many examples of similar companies of large size merging and transforming the category.
So Mercado Libre, if we go back to the history of that, I helped co-found a company called Deremate, which for a long period of time, until the failed IPO, was the dominant player. And then Mercado Libre, it was still a capital, we just funneled all the money to go public and we didn’t, started taking over. By 2005 it was 70% MELI and 30% us and we were a profitable company, and as a result neither company was doing well. When the two companies merged in 2005, it set MELI on a tear, and this was really one plus one equals 10. And this has been proven true for us at OLX time and time again.
Now, it doesn’t necessarily apply though if you buy a player in another country because then you have no, until recently, didn’t have cross-border transactions. It didn’t apply if you bought people in other categories where there was no cross-liquidity. Now, at OLX, the playbook though is you win CDC transactions buying and selling goods, because people transact multiple times a month. And then you use that to go and win cars or real estate where people only transact once every five years or 10 years. You can go in the other direction-
Antal
That’s the flywheel.
Fabrice
And that was the flywheel.
Fabrice
But there are a few interesting trends probably worth mentioning. Now cross-border transactions is a thing. So Vinted, which was one of our big successes and we’ll talk about it more later, they translated the listings automatically through AI, actually speaking of more impact of AI, the conversations between buyers and sellers are translated automatically in their native language through AI, and they have integrated payment and shipping cross-border in a way that you can ship anywhere in Europe for like two euros. And as a result, they’ve created the first European player, which didn’t exist before. It was like Germany, UK.
Antal
I wanted to spend just a couple minutes on FJ Labs and what you’re doing there. You’re invested in over 1100 companies, which is an enormous number of companies. So I’m sure everyone would love to know how you manage through that, how you go about picking winners, how you go about full-on investing, and also where you’re focused.
Fabrice
So it’s not going to come to a surprise to anyone listening to this that I’m focusing on marketplaces, and network effect businesses. I love them, they’re capital efficient, they’re winner-takes-most, they’re scalable, and they use very little capital and they follow the venture flywheel. Like a pre-seed of million, your seed of three-million, your A of seven, 10-million, your B of 15, 20-million, and with that level of capital you can build very successful, effective companies. And that’s not true of many other companies and categories that need, if you want to turn the lights on in a fusion research company, you probably need hundreds of millions day zero, and that doesn’t make it very venture viable.
So FJ Labs is a venture fund, the last fund is 300-million, the next fund is probably 300-million as well, we’re raising it right now. But we behave like angel investors. So I think I call what we do angel investing at venture scale. So what I was doing already, while I was CEO of OLX and other companies, as an angel, I still do today as a venture capitalist. The difference is the number of checks I write. Now, it’s not as though I’m doing top-down portfolio construction to optimize returns, it’s more a reflection of my personality.
In the world there are many problems and I need many founders trying to solve these problems, and if I meet a founder I like and they’re solving a problem I like, I invest. And it turns out that these days there’s about 150 new investments per year. Now, the process, how do I pick the companies I invest in? So every week we get 300 inbound deals. So the beauty of having been building these marketplaces for at this point 27 years is I have a brand when it comes to marketplaces. And so we receive 300 deals every week by other VCs, and we share a lot of deal flow with other VCs, we don’t compete with them. They’re writing $10-million dollar checks, we’re writing $400-K checks. From a lot of founders we’ve backed in the past, we’ve backed 2000 founders, they come back with their next company, they send us their employees, they send us their friends, and then pulled inbound just because of the brand. We get 300 deals a week, we have nine investors in the company, the deals are randomly assigned to one of the nine investors, we review them, we take calls from about 50 of them, the others too out of scope, too early or whatever. We have a one-hour debrief where we evaluate four things: do we like the team? Do we like the unit economics and the business? Do we like the deal terms, and do we like the thesis?
Now, let me go through each of these a little bit more in detail. So do we like the team? Every VC in the world will tell you, « I only invest in extraordinary people. » The thing it can’t be like, « I know it when I see it. » So for us the way we define it is as exceptionally eloquent visionary salespeople who also know how to execute, and the way I evaluate it in a one-hour call if they know how to execute is the way they describe number two. How good is the business? What’s the total addressable market size? What’s the margin structure? What’s your customer acquisition cost? What’s the density of your keywords if you’re buying it has? And what is your estimated fully-loaded customer acquisition cost? How does it compare to your contribution margin per transaction? What’s your occurrence? What’s your CAC to LTV after a six-month, 12-month, 18-month, two years, et cetera?
And even prelaunch, I want you to have done enough testing, even with $1000, that you can articulate that. If you can’t, that probably means you don’t understand the business as well as you think you do. Number three, deal terms, and I’m price sensitive. Because we are such large volume investors, we know the median in our category of everything. So we know where the median seed A, B, C, et cetera should be, and we try to invest near the median. And medians are much lower than mean, because it’s been driven up by the AI and crypto deals in the last year.
And then number four, does it meet our thesis of where we see the future mobility, real estate, food, et cetera? And I need all four to be collectively true. So after a one-hour call we have our debrief evaluating these four things, it’s presented in our investment committee every Tuesday. We review the 40, 50 deals, and one of the four partners, including myself, take a second call with let’s say five, 10 companies a week, and then we invest in three of every week. So we have a 1% conversion rate from top of funnel to bottom of funnel that leads to 150 investments per year, 1200 investments to date, 350 exits, and that’s worked pretty well. We’re at a 30% realized IRR right now over 27 years.
Antal
It’s impressive to do it at such scale. I wanted to briefly dig into one investment theme that’s been very front and center for FJ Labs over the last year, which is B-to-B marketplaces. You’ve made some comments that these marketplaces have been left behind effectively while consumer has had all the innovation over the last 15, 20 years. Can you give us a sense as to where the pockets of opportunities are in B-to-B that we should all be looking out for?
Fabrice
So think about the life you have as a consumer, it’s been completely transformed by marketplaces. You can buy anything at Amazon, which is mostly a marketplace, and get it to your home in two days. You can order food on DoorDash, another marketplace, you can go and order groceries on Instacart, another marketplace. You can book a hotel on Booking or a place on Airbnb, you can get a car on Uber. And everything is so extraordinarily convenient, and yet when you go to the business world, whether it be large enterprise or SMBs, everything’s still email, pen and paper, and relationship driven.
So, if you want to buy petrochemicals, there’s no catalog of what’s available. I’m just saying I need a list, I’m not even saying there’s a buy button, there’s no connectivity to the factory in understanding manufacturing delays and capacity, there’s no online ordering, there’s no online payment, there’s no tracking the order, and there’s no financing it. All this needs to happen in every category, in every industry, in every geography. Likewise, if you’re a small business owner, like imagine you’re Luigi and you created your little pizzeria, why did you create it? Well, you like cooking pizzas, you like chitchatting with the customers, and yet what is the job you end up doing today? You’re creating a website, you’re answering comments on Google and Yelp and Tripadvisor, you’re negotiating with Google, with Uber and DoorDash, you’re negotiating with OpenTable, you’re getting a POS, you’re managing your delivery fleet, you’re getting your supplies.
It’s not the job you signed up for. And so, all of this also needs to be digitized. So the six theses we’ve been investing in is, one, digitizing inputs, and inputs writ large. So, petrochemicals, steel, lithium, scrapple, et cetera, but also machine parts, heavy machinery, things that go into making other parts. Number two, digitizing SMBs, helping the Luigi’s of the world. So, Slice would be an example of that company, that’s 20,000 pizzerias digitized, over a billion GMV. Fresha for barbershops, Cents for dry cleaners, Momence for yoga studios. Again, it’s happening vertical by vertical. Chowbus for Chinese restaurants.
Number three, moving supply chains out of China, given the Cold War II. And so mostly India, but also in Mexico, even on-shoring. So we’re investors in companies like Zeot, which is an apparel marketplace helping the little mom-and-pop apparel manufacturer in India sell to the ZARAs of the world. Because of course they don’t know how to do prototyping, they don’t know how to actually win an RFQ, they don’t know how to deal with customs and invoicing, so the marketplace will do all that for them.
Number four, Libre marketplaces to support all of these, so a blue-collar marketplace like Java Talent in Europe, or a company called War Cry which is doing oil services for workers on oil platforms. It used to be called RigUp, they were doing I think a billion in GSV as well. Number five, the infrastructure. So automation companies, last mile shipping and packing companies like ShipBob, digital freight forwarders like Flexport, the payments companies like Stripe or rapid!, robotics companies to actually automate all of the last mile picking and packing like Figer.
So, all of that is another big area of investment, and then sixth, re-commerce. So, a company like Ghost, which is helping excess inventory be liquidated instead of having to go to Century 21. Now small stores in retail can buy 10K lots without ever having to spend $2-million dollars. And so, these six theses which cover tens-of-trillions of categories, are basically at 1% penetration. Everything needs to be digitized; this is something that’s going to … A wave that’s going to last for 10 years, 20 years, and we try to be contrarian. The reason we’re doing this, and obviously being impacted by AI, uses AI, versus investing in just AI companies, because I feel these are way overvalued, there’s left field disruption ala deep seed, et cetera, it makes less sense.
Antal
And I imagine there’s also quite a bit of opportunity left in B-to-B service marketplaces as well.
Fabrice
Absolutely.
Antal
Yeah, absolutely. You mentioned Vinted before, but for the audience maybe tell us a little bit about what Vinted is and how they’ve come to disrupt the European classifieds marketplace category, particularly in fashion. What we’ve observed from the outside is both the innovation and economics they’ve managed to gain on the shipping side by reducing cross-border transactions, but also they’ve innovated on the business model by shifting away from seller fees to this interesting buyer protection model. It would be interesting just to hear a little bit about why those things changed the game, and what it’s ultimately lead to, because this is now one of Europe’s most valuable scale-ups at the moment with a $5-billion dollar valuation.
Fabrice
So Vinted started out as Poshmark for Europe, and it did okay. It was funded by Berta and Inside, et cetera, it grew to 100 million GMV, it was taking 15, 20% from the sellers. The problem is it flat-lined and it wasn’t doing very well, and they were treading water. And they approached a person who was working with me, who is now CEO of Vinted, Thomas, to go and help them out. And we invested at the time when Thomas came in, and the plan was actually pretty easy, it was to implement what I defined as the classifieds 3.0 playbook.
I wrote where that playbook should be and in fact it’s in my blog in the summer of 2012, so pretty early. I’m like, « The future of classifieds is transactional, you’re going to built-in shipping in escrow and it’s going to be buyer paid. » And the thinking behind it was by going buyer paid the person paying for the service is the person getting the value. So if you do at a home pickup from the seller, the seller pays for it because they’re the one getting the value. If you have an escrow service where the buyer can say, « Okay, it wasn’t the right item. I don’t want to pay for it. » That’s the buyer getting the benefit of it and it being shipped to them is them paying for it.
So you should basically charge the person having value. We realized, especially in emerging markets, and it turns out it also applied to Western Europe, the sellers were extremely price sensitive. And the minute that you removed the seller fee, the liquidity and the inventory increased dramatically, and the willingness to disrupt the platform went down dramatically. And so step one at Vinted was apply literally the OLX playbook. Go free, do massive TV campaigns, and the volumes exploded. So the thesis of first of all going classifieds, non-transactional yet, rather than taking commission worked, absolutely worked in fashion, it worked especially well with the products that were the category of the 30, 40 euro type fashion items.
Phase two was implementing the transactional category where our business charging the buyer, which is an innovative business model, no one had done them until that point. I guess we’d done it at [inaudible 00:41:09] at the same, more or less the same time, because again, it’s applying the strategy I defined while I was at OLX, and it really, really worked. And it really worked in France where the uptake created was up to 90%. And so the effective take created of course, it first went from 15% of zero, or three with the bump up and ads, et cetera, and then went back up to 10%.
And as it worked, then Thomas kept innovating. He did the automatic translation of the listings, automatic translation of the conversations, he negotiated these cross-border deals with the post offices in order to be able to ship across Europe and created true liquidity. So, what happened, if you were in Leboncoin, which is Chrysalis of France, it’s only French listings. And having liquidity there did not help them go to other countries. Vinted was so dominant in France, they’re like, « Hey, why don’t we use that liquidity to then go to Spain, to Italy, to the UK, to Germany? »
Antal
Because you’ve got builtin liquidity on the supply side.
Fabrice
Exactly. And it worked extremely well, continuing the playbook of massive TV advertising, build up the supply and the demand, create the liquidity, have the lowest cost structure, so everyone else, the Bastiers of the world were taking 15, 20%, working to higher price points by having the entire team in Lithuania and being super disciplined in costs, they could undercut everyone and grow faster. And the company is on a tear, last year let’s say they did six-billion in GMV, 600-million in ad revenues, 80-million free cashflow, they have 400 or 500-million in the bank, they have a 50% EBITDA margin in the UK, 45% in France.
And the plan is bring that level of success to all the other countries they’re in, number one. Number two, expand to luxury. In one year they’re already as big as this year with no money spent by adding authentication of the luxury items. Number three, add collectables, and number four, add electronics. And maybe number five, in the distant future, go to maybe the US, who knows? But definitely number one, just winning the rest of Europe at the same level without going to new categories is a $15-billion dollar company, winning a few other categories, like luxury is a $50-billion dollar company, and then maybe 100, and then going to the US, maybe more. So I think this company is on a tear. I’m underwriting at the very least a three-X, maybe 10-X is reasonably high probably from here. Thomas is a machine; he’s not going to stop until he’s won.
Antal
Let’s pivot to Quince for a moment. This is a play on affordable luxury. Tell us a little bit about how this business has flourished, and what did you do in terms of playbook and business model innovation to make this business a successful disruptor in that category?
Fabrice
Direct-to-consumer brands are not new. They’ve come of age over the last few decades, but none of them actually ended up doing particularly well. You look at Warby Parker and you look at the different mattress companies, etc, the issue is they have reasonably high customer acquisition costs. They don’t have a lot of recurrence. They have a lot of inventory that they need to put in balance sheet, and or that stays on boats for 90 days as it’s coming from Asia.
And Sid, the founder of Quince, who’s amazing, had multiple insights. He’s like, « Hey, maybe I can turn this into really a marketplace. It could be an inventory-less business where I can have the factories in Asia actually bear the inventory costs. Number two, to make it actually way more efficient. Instead of shipping them on boats, we’re going to go and ship the items by plane. Obviously, you need a somewhat higher price point, and as a result, we’re going to be able to turn the inventory in every five days as opposed to every 90 days, which improves the capital efficiency dramatically in the margin structure. »
And from a positioning perspective, when you think of who’s done really well, the problem with Amazon is every category you have infinite items. Consumers don’t want infinite items. If they have 20 that are good, that’s good enough. So you want the selection of a Macy’s or Bloomingdale’s, with the pricing of let’s say Costco. The quality in fact of the Macy’s and Bloomingdale’s, and then the supply chain of the Shein or Temu. And he’s managed to build that. And the company has been on a tear in like three, four years. Last year they did almost a billion in sales. And if you, in the right demographic segment, like this is where they buy. The items are amazing. They’ve expanded categories, started as AAA grade cashmere.
Antal
It’s a young, affluent audience, right?
Fabrice
Young, affluent audience, but cashmere that’s amazing for 60 bucks. And I wouldn’t be surprised if there’s a 10 billion revenue company in there in the next five years. And that company is extraordinary.
Antal
Now let’s turn to a really exciting sector. everybody, this last year in 2024 has probably to see the advent of humanoid robots. There’s multiple players going after this market. And one of them is Figure AI, in which you’ve invested significantly in the early stages. This is a company that’s being backed by many large players including Microsoft, NVIDIA, OpenAI and Jeff Bezos, and they’ve had really meaningful orders already.
This humanoid robotics category seems to be hitting this point of optimal convergence where generative AI and other models that allow these robots to comprehend the counterparty and also respond, it seems to make them a lot more functional. In addition, in places like factories, it seems like they’ve become flexible enough to really complete very complex tasks with their hands and so forth, and with the machine vision. But I’d love to hear your view on this market, how big it could become, how Figure AI is positioned, and I guess per Elon Musk, are we just going to have a couple of $10,000 robots in our house in the next 10 years?
Fabrice
If you ask me what is the one thing that people in the mass public is underestimating in terms of the impact it’s going to have and the speed at which this is going to happen going to happen, I’d say humanoid robots. And Figure is probably is the leading dominant player here. And the founder’s extraordinary. I mean, the Tesla demonstration was a bit… They were cheating. I mean, they were not AI driven robots. They were actually remote controlled by humans.
Figure is a year ahead of Tesla, which is probably a year ahead of anyone else. So Figure is the leading player here. The founder’s extraordinary. He’s a third time founder, and we backed him every single time. He started building a labor marketplace, which is how we got to know him, called Vetteri. Then he built Archer, the EV toll company. Now he’s building this.
So most people first start by asking, « Wait a minute, why do we need humanoid robots? Why don’t you create a robot dedicated whatever task you need? » But if you take a step back, there’s tens of trillions of infrastructure that have been created for human designed worlds.
Antal
Jobs.
Fabrice
Jobs. And so, you have two choices. Either you redesign the entire infrastructure for robots or you create humanoid robots. Now, both will obviously coexist. There are many, many robots ready in the fabrication lines and car factories, and that’s okay. But there were a lot of humanoid jobs or jobs that can be done by humanoid robots. In fact, most of these, and sometimes people wonder, is that a good or a bad thing? But actually, if you think of the history of automation, and the jobs that are being automated, the ones that frankly humans should not be doing, we’re not meant to be doing this a million times over, moving a package from point A to point, B, et cetera.
So where we’re at, Figure 01 showed that you can have, essentially, a robot where you could pass the Turing test from a conversational quality. Like you ask it for food, it gives you the apple, that’s the only thing available, it can do the dishes, and you can have a completely fluent intelligible conversation as though you were talking to human and understands the world that’s in. And that’s obviously thanks to the LLMs.
Figure 02 actually now is already operating in a few places. They’re in a factory in BMW where the robots work 20 hours a day. I mean, it’s not a complex job removing machine parts from one place, putting it on the assembly line to another place. But they’re replacing about $250,000 worth of machinists, and it’s a leasing program where they lease the robots for five years. The key, though, is Figure has figured out each of these components that they’re building and buying, how to scale them, and how to scale them by to the point where the cost divide by 10. So they’re trying to make these robots, to Elon’s prediction, ultimately… I mean, we’re in years from that 10 K or 20 K per robot in the home. Maybe a leasing cost of 20 K per year is closer than that.
But Figure 03, whenever it comes, next point in two years will be a lot cheaper faster. And Figure 04 again, two, three years later, will be even more affordable and competent in terms of the number of tasks we can do. And so, they were ready. And this is actually public, I think Brad, the CEO said, « Oh, we have interest for 100,000 orders from a few large companies. » I can very well imagine a world a few years from now where frankly there are a million of humanoid robots in the world. Mostly in the factories, maybe at the few high-end homes in the five years, give it 10 years though, and the Jetsons gets a lot closer. And so, this company can transform the world. It could be a trillion-dollar company, it could be a $10 trillion company.
Antal (00:52:13):
It does have profound implications. I mean, we’ve seen AI begin to have profound implications on service type jobs and skilled labor jobs, but now this is potentially another threat towards more blue collar and factory and other types of service jobs that are more day to day. So there’s some very profound changes to the labor structure coming in the years ahead, potentially, as the costs of these robots and AI come down further.
Fabrice
I would not worry too much about that. The Luddites have always been wrong. So let’s go back 25 years. We had in 2000, and in 2000 we had this conversation. And I tell you in 2025, the top four job categories of 2000 will have disappeared. There will no longer be bank tellers, there will no longer be travel agents. All of car manufacturing will have been automated, and 500 billion of retail will have gone online, like completely vaporizing jobs in local retail. Please now describe the economic conditions in the environment in 2025.
And if I told you that, you would’ve told me, « Oh my God, the great depression, 25% unemployment rate revolution in the streets, » and yet we have lower unemployment, higher productivity. So historically, these jobs are destroyed. It’s easy to imagine. It’s hard to imagine new jobs being created, but because they improve human productivity, they improve our quality of life, they make things deflationary and cheaper, and they improve our wages. And so, the more technology we have, actually A, new jobs are created. But B, we worked less hours for more money and our quality of life improves. So I’m profoundly optimistic of the world of tomorrow,
Antal
Can I just go back to venture capital for a minute and talk about a different topic, which is exits in today’s environment?
Last year, PwC’s global top 100 unicorn report calculated a market value of over 2 trillion of basically private startups, most of which are in the technology sector. And there were almost 1,500 unicorns, so billion dollar plus companies. In this whole process as a venture capitalist, the path of liquidity has gotten much more complicated and much longer. You can invest, but it may take a long time before you have public liquidity. At the same time, the M&A markets have gotten more rusty. Higher regulatory bars, Facebook buying Instagram, no longer as easy. And the IPO markets remain challenged and the bar continues to grow in terms of what’s an acceptable size to be a public company, what’s acceptable growth. In that environment, you’ve talked about putting money into three companies, I think it was a week, but how do you get it out, and what’s the path to exit from your perspective and today’s exit environment look like?
Fabrice
Yeah, so I create a distinction between VCs and the private markets writ large and offset FJ Labs, because we’re in a very privileged position. So for VCs, to your point, it’s been a winter, it’s been an exit winter with no IPOs, no M&A, and mostly because of regulatory reasons. And obviously, the rate environment was not helping. I’m hoping that we’re coming to maybe a thawing of that. I’m not saying like bonanza, we’re going back ’21, especially with the current rate environment, but hopefully from both a regulatory perspective and in terms of the capital markets, we’re going to see more IPOs and exits in the coming few years than we have in ’23 and ’24. That helps other people.
In our specific case, we write small checks. We’re writing 4, 500K checks. We own 1, 2, 3% of these companies.
Fabrice
And so, we exit in secondary transactions. The vast majority of our exits in the last decade, but frankly… And well, the majority of the exits in the last decade, and the vast majority in the last few years have been secondary transactions. As the companies get very big, even though they’re private, there are secondary buyers. There is a real market for Stripe secondaries, for SpaceX secondaries, for Klarna, secondaries. And even the companies lower down the food chain, like if I had wanted to sell, don’t want, to Quince, or Vinted, I definitely could have. And there are buyers for those.
And often, so in the very least age, you have secondary marketplaces like Forge, or EquityZen, or SharesPost. In the earlier stages, the VCs on the way up end up being the secondary buyers. And so, if the company is completely crushing it and they’re raising their series B, and Sequoia, Andreessen, and Greylock all want in, but they all want 15% and the founder wants all three because he doesn’t want them to fund competitors, but he doesn’t want 45% dilution. He’ll take 30% primary, 15% secondary. He might come to me and say, « Hey, would you mind selling part of our shares as a favor? »
And I’m like, « Look, I love you, but 10 X, we’ll sell 50%. » And our role actually is 50% because sometimes trees do grow to the moon, even though rarely. And if it goes to zero, we’re happy because we made 10 X on the 50%, and if it goes to infinity, we still have 50% we’re happy. And so, we do sell on the way up to other VCs, and then we ultimately have the opportunity to sell in secondary markets if we want to. But I’m hoping the environment still gets better and thaws. It would be better for everyone.
Antal
And one other topic to follow on from a recent podcast we did with Reed Raymond at Apollo, around which we talked about the democratization of private equity and access to private equity products at the retail level. I’m curious, we’re in a world today where most public investors through their 401K are really only investing in the market more broadly, which means the magnificent seven is the driver of their wealth. But most of the growth is actually occurring in the private markets. I mean, do you see greater democratization and access to the private markets for smaller retail investors in the years to come as well?
Fabrice
It’s a regulatory issue, not a market demand issue. You can’t do general solicitation. People need to be accredited investors in order to invest in funds and in order to invest in startups, which means a pretty high threshold, actually. We have seen ways around it. Obviously, Angelists trying to get different exemptions for people to invest in startups, especially their friends startups or their family startups, etc. And in Europe, we’ve seen a few places where retail consumer has been able to invest in essentially feeders who invest in funds.
It’s very possible that it’ll happen. In a way, it’s happening in crypto in a completely wild, wild west way where people are investing in ridiculous meme coins. And so, I would like it to happen, but I can also understand how… The blue sky laws were created for a reason back in the 1930s. And so, you need to do it in a smart way, and perhaps not letting people invest in individual companies because that’s a recipe for failure in the tech world.
I mean, the reason we do so well is venture follows a power law, and whatever the top 100 companies of the decade return all the returns, by virtue of being in a 1,000, 1,500 startups, we’re likely to be in the power law companies. And frankly, it’s the top 10 of the top 100. If you’re invested in five startups, chances are you’re going to lose money. If you invest in 30 startups, chances are you’re going to lose money. So you really want a diversified portfolio of startups, and I’m not sure letting individuals just pick and choose… But to your point, most of the value is accruing to people on the private side. The companies are going public. Once they’re already worth hundreds of billions, they’re not going to go 100x from there. I, Microsoft went public at a 260 million valuation. Apple went public at a 1.3 billion.
Antal
All the value creation was really in the public markets-
Fabrice
Correct.
Antal
Whereas now you’ve got-
Fabrice
SpaceX’s
Antal
SpaceX.
Fabrice
3, 400 billion
Antal
You’ve ByteDance at 300 billion, et cetera. So all the value creation-
Fabrice
It’s all accruing to us and not to the public.
Antal
Peering into the future for a second. You had talked on other podcasts and other places over the years about your vision for what the world looks like 25 years from now, 100 years from now. And some of the things you’ve said in those prior comments have been around energy abundance and what kind of unlocks that creates for us as human beings. I’d love to maybe just get your latest take in 2025 of what the future looks like further out.
Fabrice
First of all, the beauty of being a venture capitalist, is in a way I live in other people’s future because I see all the ideas that are being pitched to me that are going to come of age in five, 10 years, and also see the cost curves and understand when we’re going to hit different price points where it actually becomes viable. So, first of all, taking a step back, the technology writ large and the technology revolution has been extraordinarily deflationary, and has transformed our lives for the better. Like go back 200 years, we’re all farmers. So, in 1825, we live to life expectancy of 29 at a global level, and highest life expectancy is below 40. And we go hungry multiple times a year.
Go 100 years ago, so now we’re 1925, we’re working 53 hours a week, we’re mostly factory workers and farmers, and our quality of life still kind of sucks. If you’re poor, you don’t go on vacation, you don’t have running water, you don’t have a car, definitely planes don’t really exist. And even the last 40 years or 50 years, our quality of life has transformed. When I was a kid, I remember my father yelling at my mother that she was spending too much money on long distance phone calls, and he met Nice to Paris. And now with these smartphones, we have free global video communications. We have the sum total of humanity’s knowledge in our pocket. We have an extraordinary intelligent advisor who has been able to help us in every respect.
So, you could be a poor farmer in India, and you have access to more information, communications and possibly intelligence that the President of the United States had 25 years ago. And we take that for granted. It has transformed our lives for the better in ways that are unfathomable. When I go in my crazy trips to Antarctica with no electricity, rehydrated food, I come back and they’re like, I can turn on the… It’s like hot water or shower toilet. I mean, we are so privileged, and we take it for granted. We don’t realize how privileged we are.
So how do I think of what’s coming in the next 25 years? It’s not to say everything is for the best and the best of possible world. This’s not a pinkalose state. Quite the contrary. We have many problems. We have climate change, we have inequality of opportunity, and we have a mental and physical well-being crisis. And all of these in a way [inaudible 01:03:13] are structurally incapable of addressing it. And actually, the good news is founders see these problems and go and attack them. And so, I’m profoundly optimistic because we’re finding abundance in multiple buckets. So first of all, energy abundance. Right now you have computing power abundance. Like back in the 1970s, computing power is extremely limited. You’d only use it very high-value transactions. Today, it’s so infinite, you have supercomputers in your pockets that you do anything.
The cost of electricity has been limited and limiting of energy writ large. And so, until the 1990s, we had actually perfect correlation between energy consumption and GDP growth. We’ve managed to become much more efficient. And so, right now we’ve actually managed to grow the economy without actually growing our energy consumption. But what’s interesting is the cost of energy, because of solar, is declining extremely rapidly. So solar is already the cheapest form of energy production and is declining by 11% a year. It’s dividing by 10 every decade, and it’s continued to divide by 10. So divided by 100 in 20 years, 1,000 in 30 years and 10,000 in 40 years. And it’s continuing to decline. And I can foresee a world 10, 20 years from now where the marginal cost of energy production essentially zero. Of course, solar doesn’t run during the night or on cloudy days, but battery costs have also declined. They’ve divided by 42 in the last 27 years, they’ve divided by 10 in the last decade. And now solar plus batteries is reaching the point where it’s becoming cheaper than all other of energy production. And as a result, we’re seeing massive penetration increases in many emerging markets where like 12% we were at 0.1% energy production from solar. Now it’s 12% on a global level, it’s actually doubling in many places year over year in many countries, both battery capacity and solar production.
And so, even with no subsidies and no government intervention, within 20 years, I can easily imagine almost all of the power being generated by solar and stored with batteries on a global level. We have infinite reserves of the minerals needed, and it is actually way more efficient. We basically are going to move to a world where all the energy is produced by solar and will be virtually free. All the cars will be electric and will not have emissions. So, our entire transportation and energy production system will be emission free, and the marginal costs will be zero. So, we’ll be able to increase your energy consumption once again and waste it. Once you can waste energy, you can do extraordinary things.
Once you can waste energy, you can do extraordinary things. People were like, « Oh, we’re running out of fresh water. » It’s actually not true. 70% of the world is water. You can desalinate fresh water, or you can just desalinate salt water with infinite electricity. I have a water desalination system in my house in Turks and Caicos. We’re generating 5,000 gallons a day. My water bill is 10K a month. The machine is 72K.
Antal
Wow.
Fabrice
I mean, it’s paid for in seven months and of course it’s run off solar, so I’m completely off grid and the marginal cost is zero. And once you have infinite fresh water, you can grow crops in the desert. You can grow crops from vertical farms. So, all of a sudden, the cost of food goes down, the cost of water goes down. I mean, you get to an abundance agenda and world, your underlying needs will almost become free. Now, it goes beyond that as you start having things like humanoid robots making the cost of services and labor go down, which actually is probably essential given that our birth rates are low, populations getting older and starting to decline. We’re going to have actually fundamental labor shortages in many categories. So, I’m not worried about unemployment. I’m actually more worried about labor shortages in many categories, and we’re going to be able to free up the labor to do the right things by having people help the elderly, et cetera. I’m so optimistic.
Antal
Yeah. So the categories where we never seem to have real overall benefit of technology deflation, are houses, education and healthcare, are we going to see turn around in the next five to 10 years?
Fabrice
So the categories that are highly regulated and or dominated by public services, have been much slower to innovate, but they’re such large components of GDP, we’re going to see innovation. We’re seeing innovation in three levels. So first of all, education is seeing it, you’re just not seeing it within the context of traditional schooling system. But right now, if you are a super motivated, intellectually curious student in India, you can actually go on YouTube and learn anything. You can go in Coursera and take classes with the best professors in the world in any category you want. And yes, you’re not getting the diploma, but you’re getting the knowledge. Many of my startups now when they recruit programmers, they can be any age, they can be in any country, they don’t need to have gone to school. We’re just going to give them programming tests. If they can program well, we’re hiring them.
And so, it’s completely changing credentialing. It’s like literally, are you fantastic at what do you do? And by the way, part of the reason I don’t worry about the job thing is human plus AI will probably be a lot better than just AI or just human, and we’ll work together with technology. So education is already happening. I mean, Khanmigo is teaching kids in K through 12, using the Socratic method teaches them math. In science, I mean, we’re seeing improvements in drug discovery. Alpha Fold is playing a pretty big role 100% in creating new drugs. Most of the costs in the US healthcare system are actually non-medical. They’re the overhead of the fact that we’ve created a completely cumbersome, idiotic paper system. You go to Estonia and they basically automated everything. Like online medical records, everything. Go to the doctor, recommendation to go to the specialist, payments, everything’s online and automated. It’s been like that for a decade. So there’s huge opportunities that I think will start happening. And the same thing in public services.
Housing. The main problem in housing is actually regulatory. So we are creating artificial constraints to supply by having air rights and landmarks, et cetera. Obviously don’t build in Central Park, but most other things don’t need to be protected. And things like NEPA have been weaponized to basically empower-
Fabrice
… NIMBYism. So we don’t have an underlying construction cost issue, and yes, there are a lot of things we can do here in terms of prefab, et cetera, the fundamental issue is actually the regulatory regimes around building. 80% of San Francisco is zoned that you can’t build apartment buildings. I can decrease the cost of housing in San Francisco overnight by letting them build up infinitely.
Antal
Very, very close to home topic and very inexplicable. So you run these annual tech gatherings in Turks and Caicos, at your property there, and what you’re doing is you’re bringing together communities of founders. And what I’d love to maybe just hear from you is what some of the more standout moments have been of bringing these groups together in that kind of an environment, and what sort of insights come out of those events that drive innovation and other magical moments, so to speak?
Fabrice
Well, magical moments happen in many ways. So for instance, recently we had the founder of Somos, and he didn’t even speak Spanish. He went to Medellin in Colombia, and he built a super high speed fiber internet system. So he’s delivering like 100 gigabit fiber at five bucks a month. And he found way to do that. And he’s like, look, people were thinking incrementally, but you could do what SpaceX has done. The cost of a kilogram to send things to space with the space shuttle was like $60,000. And with Starship, they’re hoping to be maybe as low as $15, but let’s say $200. Are there ways to finally use existing technology to completely transform the cost structures of the world as we know it?
And the incumbents typically never do it, and he showed us how he did it, applying existing technology to solve the problem. We were mind blown. And we’ve seen so many countless examples of this. Now, there’s also serendipity where you have the COB Bay is there and he meets the founder of a company that he thinks is fascinating, then they go and buy the company or partner with the company. But it’s a great way to take a step back and really see what is happening, what are the overarching trends. So for instance, this one is all about AI, of course, in terms of where is there too much investment, not enough investment? Where are the orthogonal approaches and what are the contrarian plays in the category?
Antal
You’re doing it in an environment with very little outside noise, which is-
Fabrice
Yeah, you can’t do it in New York. Otherwise
Antal
… also great for stimulation. We’re going to wrap up, but we’re going to have a couple of quick personal questions for you. So what was your favorite recent travel destination?
Fabrice
Well, I went to Antarctica and I walked to South Pole holding my 100 pound sled, I’m completely off grid, eating rehydrated food. You actually have to poop in a plastic bag that you have to carry because it’s really leave no trace behind. You’re carrying your own fuel, your own sleeping bag, your own tent, and it’s so different and so harsh. It’s negative 50. You’re at 10,000 feet altitude. And it’s really an exercise in gratitude because you come back from that exercise where you’re burning 8,000 calories a day, you would lose almost a pound a day, and you realize how lucky we are to be living in the developed civilized world that we live in.
Antal
How long does that trek take today?
Fabrice
If you do coast to pole and you’re super fit, like 40, 50 days, and if you just do the, you’re ready, get dropped off in the glacier, which is the part I did, it took seven days.
Antal
Okay. I’ve also often heard that you’re an avid reader. What’s a book that has really stood out for you that has changed your view of the world?
Fabrice
The book that probably changed my view of the world because it made me look at it differently was Sapiens, in terms of understanding how many of these things that we take for granted are actually human constructs. Like currency, like borders. But honestly, I read a lot. I read about 100 books a year, but I mostly read for fun. And I find that the way to stimulate my mind and think of the future is actually reading sci-fi, to a lesser extent fantasy. So my favorite sci-fi book of the last few years is Project Hail Mary. It’s by the author of The Martian, but it’s way more fun and way more compelling. The Martian was like Robinson Crusoe in Mars trying to save himself. So he’s alone and he’s talking to himself, and the stakes are just him. This is like Robinson Crusoe in space with Friday, so friends, trying to save humanity, the stakes are higher.
Antal
Awesome. What is your favorite AI app today? And I’m going to give you a couple choices. ChatGPT, Claude, Perplexity or Gemini, or something else that we haven’t-
Fabrice
I am a GPT power user. So first of all, I build my own AI, which is a digital representation of everything I’ve ever said, written, thought, et cetera. And it took me over a year, and at first it really sucked. And in the process of doing that, of building it, I played with everything. Anthropic, Pinecone. I mean, I coded it, so it required infinite iteration and nothing really worked until 4.0 came out. And of course, I by then structured the data properly and completely converted all of the podcasts and transcripts, coded who was speaking. I used Azure or CR to convert my PDFs into intelligible content. So hundreds of hours of work. But GPT4.0 is fantastic and the new models are even better. I use it for everything. From research, questions. Yeah, it’s replaced Google. But beyond that, I use it as a, need an analyst to do research for me, it’s GPT. It’ll create summaries and analysis. I mean, sometimes it’s wrong, but it’s incredible.
Antal
It is pretty remarkable. And where can we speak to the virtual, Fabrice?
Fabrice
So if you go to my blog Fabricegrinda.com, there’s a link for Fabrice AI, and it is shockingly accurate. You can ask what our thesis is, what the feedback on your startup idea, what the mean valuation for a series seed or A is. You can even ask personal questions on how Antarctica was, what I’ve learned, how to live acid light. I mean, anything I’ve ever written about you can actually interact with, and it’s very accurate. And I’m currently coding the next two versions.
The next version is one where I will have a digital avatar of me with my face and likeness and voice, and you’re going to be able to talk to it or have it talk to you. You can either talk to it directly or via text. And then I’m trying to code, no promises of whether it will work or not, Pitch Fabrice AI, where right now, as I said earlier, we get 300 pitches, we take calls with 50 and we get the debriefs for them, and then I speak to maybe five. It will allow the 250 and frankly the 295 to talk to a version of me and get feedback on their idea. If it works.
Antal
Do you see the virtual Fabrice sitting on your investment committee and making investment decisions down the road?
Fabrice
It’ll make recommendations. We’ll see how good it could be. Look, I’m only going to make it live if I like the debriefs. So obviously the input is the deck, the transcript of our existing conversation with the founders, and then our debriefs and recommendation. If I feed it thousands of these, can it get reasonably accurate or good feed? This is really, again, a public service humanity or into the founders. Can I tell you, oh, your unit economics are underwater, or you’re going too early to talk to VCs, you need a bit more traction. Go to the friends and family round first. Or as a founder, you really need to be more eloquent in terms of the pitch you’re going after, or the category you’re going after is too small.
Antal
Right.
Fabrice
Can it actually give actionable feedback to founders to be helpful to them? The answer is no. I want make it live. The answer is yes, I will. So TBD. But that’s what I’m coding and working on at night right now.
Antal
Excellent. And the final wrap up question is, do you have a favorite podcast besides the LionTree podcast?
Fabrice
I don’t listen to many podcasts. I read books more than anything else. And on Substack, my favorite thing to read is probably Noah Pinyon or Noah Smith who writes about economics, public policy, et cetera.
Antal
Okay, fantastic. Fabrice, it’s been a real pleasure learning about the OLX history, the way you invest, and a lot of your perspectives on the world. So thank you very much for joining KindredCast today.
Fabrice
Thank you for having me.