432. Lessons from Building the Iconic Firm QED, the History of Fintech, the Efficiency of Carbon Vs. Silicon Processing, and Decentralizations Role in the Future of Finance (Frank Rotman)

This episode dives into the insights and experiences of Frank Rotman, co-founder and chief investment officer at QED Investors, exploring the evolution of fintech and the future of finance.

1. The Origins and Evolution of Fintech: Fintech innovations are closely tied to broader technological developments, evolving through distinct stages from basic automation to sophisticated, AI-driven solutions.
2. The Role of AI in Finance: AI is poised to be a transformative force in fintech, potentially reshaping everything from customer service to back-office operations.
3. Decentralization in Financial Services: Decentralization could address inefficiencies inherent in centralized financial systems, although it comes with its own set of challenges and learning curves.
4. Venture Capital Strategies: The conversation sheds light on how venture capitalists must adapt their strategies to stay relevant and effective in a rapidly changing tech landscape.
5. The Impact of Silicon versus Carbon Processing: Discusses the efficiency and potential of silicon-based learning machines in replacing traditional carbon-based human processes in various sectors.

1. Introduction to Frank Rotman and QED

Overview of Frank Rotman’s career and the founding of QED Investors. Frank Rotman: "The story of QED is really about scaling young fintech companies into major players."

2. Evolution of Fintech

Discusses the stages of fintech evolution, from early automation to the integration of AI. Frank Rotman: "Fintech innovation followed major technological S-curves."

3. Future Trends in Fintech

Explores AI's role and the potential shift towards decentralized finance systems. Frank Rotman: "AI and decentralization are the next big S-curves in fintech."

4. Venture Capital and Fintech

Rotman shares insights on adapting venture capital approaches to align with new technological trends. Frank Rotman: "Venture capital needs to evolve with technological advancements to remain impactful."

1. Embrace Technological S-curves: Stay updated on emerging technologies to leverage early opportunities.
2. Evaluate AI Opportunities: Consider AI's potential to streamline operations and create value in financial services.
3. Consider Decentralized Solutions: Explore how decentralized finance can solve inefficiencies in current financial systems.
4. Adapt Investment Strategies: Adjust investment approaches to include support for transformative fintech innovations.
5. Continuous Learning and Adaptation: Stay agile and open to learning from new developments and failures within the industry.

Frank Rotman of QED Investors joins Nick to discuss Lessons from Building the Iconic Firm QED, the History of Fintech, the Efficiency of Carbon Vs. Silicon Processing, and Decentralizations Role in the Future of Finance. In this episode we cover:

Decentralization, and Stable Coins in Finance
Financial Nihilism and Is Impact on Savings, Investing, and the Economy
Megatrends, Including Aging Population, Climate Change, and GLP-1 Drugs
Carbon-Based vs Silicon-Based Learning Machines, Strengths and Limitations
Balancing Investor and Founder Needs
Venture Capital's Role in Startup Growth and the Need for Innovation in the Industry
Investing in Startups, Importance of Reserves, and Portfolio Construction

People

Frank Rotman, Nick Moran

Companies

QED Investors

Books

None

Guest Name(s):

Frank Rotman

Content Warnings:

None

Frank Rotman
Welcome to the podcast about venture Capital, where investors and founders alike can learn how VC's make decisions and reach conviction. Your host is Nick Moran, and this is the full ratchet.

Nick Moran
Frank Rodman joins us today from Richmond, Virginia. He's the co founder and chief investment officer at QED Investors, the premier venture fund investing in fintech. There, he has invested in companies including credit Karma, Flywire, Sofi, Braintree, Avid Exchange, and Nubank, just to name a few. Prior to QED, Frank was one of the early architects who helped create what became Capital one. He has been named to the Midas list six times, invested in over 200 companies, and the firm currently has 4.3 billion AUM.

Frank, welcome to the show. Happy to be here. So I covered a bit of your background. Can you give us kind of the short summary of your path to QED? Yeah, I'll give you the two minute rundown, because that's about how interesting it is.

Frank Rotman
So the story of QED is really the story of Nigel Morris and myself working together for the past 30 years. So in the first incarnation of my career, it was helping build what became Capital one. So I was hired directly out of graduate school from the University of Virginia into this sleepy little regional bank called Signet bank. And a few years later, we ended up spinning off Capital one. And I was at Nigel's side during building Capital one into the giant that it is today.

So we did that for more than a decade. Nigel left in 2004, and I left in 2005. I actually left to build a student lending company, so tried my hand at entrepreneurship, and it was a heck of a lot of fun actually building something from the ground up. Did that for a few years and then joined backup with Nigel to form QED investors. And the whole premise was that we had good, solid operating backgrounds, and we thought that we could be helpful to young companies and help them figure out the path to becoming real companies.

And scaling. We used our own capital at first to basically become investors. And after a while, we proved to ourselves that we were pretty good at it and started scaling the team and started scaling the Aum with other people's money. And here we are today with over 200 investments made over the past 16 years. We're a global firm, US, UK, Latin America, Southeast Asia, a little bit in Africa.

So we've done a lot globally. At this point, we have over 20 investment professionals, and we're out there just trying to help young fintech companies become scaled and established fintech companies and hopefully do something real. And what is the best stage entry point for you, Frank. So we have an early stage fund and we have a growth fund, but our preferred entry point is still very early. Being ex operators, we like to actually help crack the code.

We actually use the language and we find it funny, but it's really true that we're really operators masquerading as investors. So we like getting involved before the story has been told or the business model has been cracked. And part of our job is to help the founder. We're very hands on. We get involved deep in the teams with the 20 plus investment professionals that we have on the team.

We have expertise that crosses lots of disciplines within fintech. So when you get QED, you get all of QED, and we unleash ourselves against the company wherever we can be helpful. So, you know, this is a deep question. I'm going to ask you to kind of give us an overview at kind of a macro level, in brief, but can you walk us through the evolution of fintech? How would you describe the stages that we've been through and then kind of the stage we're currently in?

So it depends how far back you go. Fintech wasn't even a word or an established thing until even after QED was started. So we didn't even have a name for what we were investing in. We started in 2008 to put that in perspective, a few years later that fintech was coined. So if we think about the origin of fintech being somewhere in the past couple of decades, the first part of fintech was really about snapping out a product that a bank was offering to their customers and doing that one thing better than the bank did as part of the corpus.

So if you go back to some of the fintechs of the 1990s, or the two thousands, like going way back when Capital one, as an example, would have been classified as a fintech, we snapped out the credit card business from the bank. We actually did not have a banking charter. Did one thing better than the banks did, and then eventually we came full circle into buying banks and becoming a full service bank off the backs of this one product that we did better than the other players in the space. So a lot of the early innovation in fintech was saying that if a traditional bank runs 350 products, which is about the number of products a bank has, they're precisely zero that they are world class at when you actually look at the typical bank. So snapping out a product, doing it better, figuring out how to have better experience for the customer, hopefully lower pricing, less friction, all of those things enables you to actually build a company with purpose.

So a lot of the early fintechs looked like that, but the innovation in fintech really followed s curves of technology as well. So you can actually stare at big technological innovations. And a lot of the fintech innovation followed. So there is a major s curve around mobile. There is a major s curve about UX and UI being very different than it was historically.

The ability to actually procure product online or procure product through your phone with less friction than it would if you actually showed up in person to fill out paper applications and talk to bankers. So you can think about the early wave when fintech was coined as fintech in the 2010 timeframe. Sometime around there, it was about UX UI and APIs and reducing friction for the application for fintech products or for financial service products. You can look at cloud as another s curve innovation for a lot of the b two b type of platforms in fintech, making things cloud native, making things available at all times, reducing friction for back office. There's a lot of things that happened in fintech around that, but a lot of the innovations are built on top of these s curves.

So if you look at products like cross border money remit, cross border money remit used to be a face to face physical process. You would actually go to your bank and you would convert digits to physical cash to then go to another location to actually wait in line so that you could take that physical cash, fill out a paper application, and turn it into digits so that you could then pick up your phone and you can call someone on the other end and tell them that you just wired the money so that they could wait in line, fill out a form and basically claim the money and turn it from digits back into physical cash. And with the advent of mobile and the advent of reducing friction, like, you could reduce a lot of those steps because of that innovation. Why do you have to actually convert it from digits to physical to digits to physical? Like, you can skip a lot of those steps if everything is mobile native.

So innovations like that create new form factors for products and companies, and we've invested in some companies that have taken advantage of that particular trend, like remitly. So I think you have to follow the technological innovations if you want to really stare at what's coming next. And I think it's just really interesting because we have a couple new s curves that are kind of playing themselves out. And that means there's going to be another wave of innovation that could be a decade or more long. Yeah.

Nick Moran
What does this next decade look like, what are the biggest s curves that you're tracking? And how might that make fintech look differently in the next ten years? I mean, the obvious one everyone is talking about is AI. Sure, you can't avoid a conversation talking about venture without talking about AI. For what it's worth, it's something that I've been thinking about for over 30 years.

Frank Rotman
It's what I went to school to actually learn. A lot of the things that we actually talked about 30 plus years ago. Now we have the compute power, now we have the aggregated data sets that we can start really making AI a valuable technology, right? A lot of it's been theory over the past couple of decades, and there have been some amazing AI products that have been built using technologies over the past 30 years. So it's not like nothing has happened.

But I think we've gotten to a place where a combination of compute and the aggregation of datasets is allowing us to actually turn AI into what it's always envisioned to be able to accomplish. And it's still in the early wave, but it's hard to not imagine the world being different with AI behind the scenes powering a lot of basically what people are actually doing within the banking ecosystem. Very good. How does decentralization play a role going forward? Do we see more decentralization of finance or not?

So decentralization is just another form factor of finance. Right. So if you think about the banks as centralized finance or traditional finance as centralized finance, decentralized finance has both features and flaws. Centralized finance has features and flaws. So I think a lot of what we've seen in the crypto world and decentralized finance deFi, it's been speedrunning all of the learnings of traditional banking, making a lot of the same mistakes as traditional banking.

I'm trying to figure out where it can get its sea legs, where it actually serves a purpose, and it's not just a technology looking for a problem. And I think there are some places where decentralized finance is actually superior because it solves a real problem that centralized finance can't solve very easily without the decentralized infrastructure. But I think the advantage that traditional finance tradfi or centralized finance has is that it has 200 years of experimentation to eventually make lots and lots of mistakes and then build the safety nets around those mistakes and figure out how to fill in the holes. A lot of what we're seeing is this playing out very quickly in decentralized finance, with lots of mistakes being made along the way and a lot of the people in the space learning some of the exact same things that centralized finance learned over a much longer period of time. So again, I think there are going to be some areas where decentralized finance really does take hold.

An example would be stable coins. If you think about what blockchain technology is, it's really about ledgers. It's about managing a ledger where it's not a central party that's actually controlling the read and the write to that ledger. And if you think about where ledgers have gotten to be a bit of a mess, which I guess you can call that a technical term, bit of a mess. If you think about all of the banks, like every bank runs their own ledger.

And if you think about how many banks there are, there are 10,000 or so depository institutions in the United States alone. If you include credit unions and community banks. If you think about internationally, you can add another couple, tens of thousands of banks. And every time that you want to move money from one bank to another, it has to move ledgers. So there's a bank settlement process associated with moving it from one ledger to another ledger.

And because of that, the movement of money between banks, and especially cross border, it's just slow. It has lots of mistakes that are made along the way. It's prone to fraud. There are lots and lots of issues. And if you think about some of the biggest markets in the world, think about international trade settlement.

International trade is a pretty big market, and it involves moving money from ledger to ledger. So the concept of stablecoin, it really is about having a single ledger that basically is working 24/7 it's uncensorable, and it has the ability to have good money on it at all times. So that if you conduct trade on that particular ledger, there's no friction involved. So I think stable coins are going to be a magical thing for cross border trade. It's going to be where things really take hold.

Stablecoins of other value as well, in terms of being able to invest or denominate your savings in a currency that you think is more stable. It enables people internationally to denominate things in us dollars, as an example. And I bet people in Egypt are very happy if they were holding stablecoins recently or in Nigeria. So I think there are multiple purposes of decentralized finance, but stablecoins are one that I think will take hold. So if the killer app for bitcoin is store value is the killer app for the stablecoin.

Medium of exchange, medium of exchange and denominating unstable currencies in stable currencies, or converting them from unstable currencies to stable currencies. So it enables international people in certain jurisdictions to get exposure to a more stable currency. Is financial nihilism a threat or an opportunity? That is just. It is a topic that I'm thinking a lot about because it's actually depressing when you really start digging in about this next generation of individuals that are coming up where they just don't see a way to a better life, they don't see a way to financial freedom, they don't understand the old way of thinking about the world, which is how I grew up and my parents grew up, where if you just studied hard and they worked hard, and you ended up in a good job, like, you'll end up with the ability to support a family, you'll be able to buy a house, you'll end up being able to retire, just work hard and good things will happen.

Financial nihilism is about not seeing a path to that outcome. And I think a lot of the younger people today, they see the lack of affordability in the housing market. They see the inability to take home enough money to not just pay bills, but have money left over to save. They see the printing of money and inflation taking over to a point where they actually can feel the difference in how much they're paying for things. Year after year.

Between 2020 and today, the US dollar has lost about 20% to 25% of its buying power. That's a very short period of time where you can remember when things cost less, because it really wasn't that long ago that things cost less. So financial nihilism is showing up all over the place. It's showing up with more gambling behavior, because a lot of people, if they believe they can't save like their solution is, maybe I should just take a lot of risk and see if I can hit a home run by getting lucky, rather than just working hard. So it's affecting everything from savings to investing to the stock market and how it's working, to crypto and people investing in things like meme coins, trying to overnight work their way out of these issues that they see as unsolvable.

So I think this is very profound. I don't have answers, but I think that it's going to change the nature of banking. I think it's going to change the nature of how people think about money and their jobs. So it has to have a profound impact on how businesses are built and what products really start solving these problems. Without going deep on any trends, specifically, what are the megatrends, what are the macro factors that you and the folks at QED kind of discuss and look at most closely.

So when you look macro, you have to look at certain trends that you actually can't deny. Right? So the aging of the US population is an example of a trend that's undeniable. You can't fight demographics. And when you think about the lack of savings in the United States and the fact that there's one thing you cannot borrow for and it's your retirement, if you think about the fact that people are by and large living longer lives and living them less healthy than they were in the past, it's creating huge healthcare expenses and aging in places becoming a very large issue.

So you can't outrun demographics. There's no denying it. And the aging of America is one trend. And if you actually look globally, you have some of the opposite trends. Some of the emerging countries are extraordinarily young, right?

So if you look at Nigeria or if you were to look at Turkey, like very young populations, and that has completely different implications than it does in the United States. But staring at demographics is important. Climate is another megatrend. Pretty hard to deny what's happening, even though there are climate deniers, but pretty hard to deny what's happening with climate. It's making certain products less affordable or maybe even unaffordable.

So if you think about home insurance, if you live in a location where there's weather, insurance prices are going way up. If you think about California and wildfires like these have big implications. And if we can't get carbon under control in the next ten years, if we can't control the warming of the world, it's game over. So nothing else actually matters. So this is a problem that in the next ten years needs to be solved, therefore it will get solved.

I mean, I actually do believe that problems that need to get solved will get solved. So I think it's going to, you're going to see a lot more time and attention and money being thrown at this because it's a problem that we don't have a choice. We have to solve. So those are a couple examples. In the fintech space, we're seeing a lot more embedding products being embedded within other products and other distribution channels.

Instead of going direct to consumer, you're now going to where the consumers already are shopping for products or in products. So a lot of embedded fintech is becoming real because we now have the technology to link lots of things together. So those would be an example of just a few. I don't want to get too off topic here, but how do you think about Ozempic? People have talked about how the economic impact of ozempic could be greater than AI because of its potential to impact healthcare and subsequently financial implications on consumers and those that may have bad health, bad eating habits.

Nick Moran
Do you think that's a viable position? So every, who knows, decade or so, maybe two decades, like a miracle innovation takes place within the medical world. And I think it's going to accelerate from here, by the way, for a whole host of reasons. But I actually think GLP ones are a class of miracle drug the same way that statins were a class of miracle drug. I mean, statins absolutely improve the health of the world, right?

Frank Rotman
Anti malaria drugs, like they absolutely improved the longevity of different societies. And I think GLP one drugs are going to be similar like they work. And in a world where, you know, cheap and calorie dense foods are available, like having the ability to fight the cravings for those foods, having the ability to change the hormonal response to those foods like that actually can be a miracle for people who are struggling with weight and a lot of people who are struggling with their weight. It's not a mental thing, it is a physical thing, like a lot of reasons why GLP ones might be necessary and not just a nice to have. So they're going to create their own issues, just like statins create their own issues.

But by and large, I think that the GLP one innovation is going to help healthcare overall. Frank, you talked before about AI. I'm curious where you think the ephemeral versus the durable value will be created by AI startups. So I is kind of the thing that you can't unsee when you see it. And once you actually see it in practice, it's kind of the art of the possible, right?

So if you think about OpenAI and everything that's now kind of in the public discourse about what's happening, everyone's trying these AI products, and what they really need to realize is this is really just v 10 of something that's innovating at such a rapid pace, in fact more rapidly than maybe any other innovation or s curve that we've seen. And it's because the aggregation of data and the availability of compute has hit a point where we can now create and crunch these amazing models that now can be tuned to do very specific things. And by doing those very specific things, you can replace the alternative for doing those very specific things and you can do it in a very cost effective manner where the input is a known form fashion that we know, right. You just have a little input bar where you can upload something or you can type something. Like we know how to use the input bar that AI is now using and we actually know how to use the output, right.

The output is either text or it's a PDF or it's a music file or it's a video. Like we know what the outputs are going to be of the AI technology. So if we know the form of the input and we know the form of the output, and you can put a big black box in the middle that's being tuned by people that are throwing hundreds of millions of dollars at it in order to produce these intelligent answers. Like we actually know how that works. Anyone who's watched Star Trek or anything in science fiction, like you just talk to the computer, right, the input device, and you get an answer and the answer is by and large something that you know how to use.

So because the form factor of input and output is like really obvious, I think the adoption curve is going to be much faster than it is for most things. Because of that, I think the incumbents are going to be adopting technology very quickly and they have all the money. Like big incumbents that are going to be able to use this technology to improve their product, build enhanced functionality and or replace a lot of the expensive people that are behind the scenes doing tasks that could be replaced by the AI technology. Like they are at an advantage because they spend the money and they can spend the money integrating this because they already know what the input and the outputs look like. This is very different adoption curve than for mobile or for cloud or for the web, where you really had to train people to do something very different.

I mean, I remember when the web was a brand new thing and we had, you talked about a company being a web company or a non web company. That's right, yeah. Did you have a website? Yes. No.

If you had a website, it required completely different technology, completely different technologists, like a new way of thinking about how you actually host what you're selling. Like completely different world Internet companies. Non Internet companies, yeah. And think about mobile. Mobile is a completely different form fashion.

If you think about having a screen that you can actually buy things on, you have a lot of real estate where you can position the different products and it's easy for people to upload things if it's on their computer. And computer has a camera like the first smartphones you have the smallest device with the biggest input device and you're trying to figure out how to make this thing work. And like, it was a completely different form fashion that actually didn't work at first. And it took a lot of iterations to figure out how to get people actually using mobile and shopping on mobile, and you needed more bandwidth and it took like many iterations before it kind of took off. Frank it does beg the question, though, does AI open new modalities or new mediums of interaction with technology?

Nick Moran
Obviously we're seeing audio base with Alexa and Apple's vision Pro, and maybe there's a future reality where you have a bunch of information in your periphery being served up by AI. Does it open up a new platform for consumers to get their inputs and outputs? But consumers are used to voice, they're used to typing, they're the form factors I think are now understood, right? Yes, you could put goggles on and yes, you could interact with the inputs and outputs a little different. But I think we understand what that's going to look like.

Frank Rotman
And it's, do you want to upload something? Do you want to talk about something? Do you want it to listen to something? What do you want it to do? And it's going to be able to do that thing in order to take your input, crunch it through a black box model and give you output.

And the question then is what you do with the output? How good is that output? Is the output recursive? Are you interacting with it like a human being or are you letting it do most of the work on its own? And I wrote a piece not too long ago on this where I talk about carbon based learning machines versus silicon based learning machines.

And by carbon based learning machines, I mean us, right? Like, we are carbon based learning machines and silicon based learning machines. That's computers, right? And carbon based learning machines are actually bad students. If you actually think about most carbon based learning machines, there are a lot of topics you can't learn, whether you want to learn them or not.

Like, you can't learn them. Guilty. Yeah, I mean, we're all guilty of that. We're all guilty of there is something that we can't learn. For me, I struggle with learning languages.

Like, it's just something that my brain does not tune itself to. Like, I've never been able to learn another language. Just doesn't work for me. Other people, they take to that naturally. But for me, I'm really good at math.

Like, I can pick things up. I know how to code. Like, I can code incredibly well. So other people, like, they can't think that way. They can't learn those skills.

But if you take a carbon based learning machine, like, it takes 18 to 21 years of a carbon based learning machine's life to learn how to learn, that's a really expensive training period. And then you have to pick something that you want to learn, specifically that you're going to spend the next 30 years of your life executing against, like actually doing on a day to day basis. And you actually learn slowly. So you practice something, like, in the workforce for a while, and you get better and better at it. And hopefully, if there's new information that comes in, you're staying up to date and you're continuing to train yourself.

But what's interesting is carbon based learning machines, you pay for the compute in a per minute basis. So if you think about people, you buy them by the hour. So you can think about a call center employee as an example, and you might pay them $20 an hour to be in the call center. When you load up, all of the expenses on top of that might mean that by the time you add the overhead, it costs fifty cents a minute for computer, for a carbon based learning machine. And there are other carbon based learning machines like high paid lawyers, where it might cost $2 or $3 or $5 a minute for compute.

So when you think about these expensive carbon based learning machines that, by the way, as they train themselves, they get more expensive that lawyer, when they're one year out of law school, they might be $2 a minute. And when they're 20 years out of law school and a partner in a law firm, they might be $10 a minute or $20 a minute to get a more refined answer. Or hopefully you're getting a better answer. Or you're a venture capital investor, right? Yeah.

So carbon based learning machines get more expensive over time. They don't necessarily get better. They get more experienced. They don't necessarily have the most up to date information, and they don't continue to learn. A lot of carbon based learning machines are static.

And some carbon based learning machines actually aren't good even after they learn. So a lot of what AI is doing with these LLMs is basically saying, look, let's just spend all of this money, 200 million, 300 million, $500 million, training the general intelligence so that it knows how to learn. So instead of taking 18 years to train a single carbon based learning machine, you spend a couple hundred million dollars to train an infinite number of clonable silicon based learning machines. You then give it the data that you want to train, and it has the ability to train itself very quickly. Yes, you need people to reinforce the learning, and there's a bunch of other things you can do to train the model, but once you've trained it, you can pop out an infinite number of instances of that model.

Nick Moran
Yeah. Low marginal cost scanning. I was talking to someone today where they were complaining about the compute costs of AI, and they were saying, it costs me $1.50 to run 1000 queries. I'm like, it's a dollar 50 to get a thousand answers back from this training. Contracting that out.

Frank Rotman
Contracting that out, because someone spent the time training it with hundreds of millions of dollars. And now it's leverageable to create an infinite number of inexpensive answer machines. Right. So there's still a lot of work to do before things are trained to a level where they actually can replace people. Right now, they're better as copilots, they're better at providing answers that then can be refined by people.

So instead of the person taking an hour to do a task, they can take an hour and do ten tasks because they have help at their side doing it. But eventually it's going to get to the place where it can actually replace human beings because they're actually better learning machines than we are. It's like AI enabled humans versus human enabled AI. That transition will happen. Frank, while I've got you, I do.

Nick Moran
I want your insights on venture capital. I've been following you for many years, and you're very open with sharing, which is helpful to all of us. But I feel like the asset class, there's been a chronic lack of innovation, despite trying to invest in the most innovative companies. How does the asset class need to evolve in order to survive? I wrote a thought piece on this just about two years ago called the three body problem, which is actually funny, because now three body problem, very popular, great books, by the way, but popular show on Netflix, but it basically outlines some of the problems with venture capital and about how the VC world has to evolve, or it's really going to suffer from a lot of problems.

Frank Rotman
And one of the major things that you learn after you spend a little bit of time in VC is that you realize that you have two major ecosystems that you're trying to balance. You have the LP ecosystem, which is really the source of capital, the money behind what you do. Because ultimately, as a VC, you're selling two things. You're selling money and you're selling advice and connectivity and all of the things that come along with the people that are involved in the process. Without the money, VC doesn't exist.

You become consultants, but you need the money, which means you need the LP's. And on the other side you have this very large ecosystem of founders, an ecosystem of startups and ideas, and both of those ecosystems are extraordinarily large. And for VC to really be, for a particular VC firm to have a stable point where they have a reason for existence in the market and can pound out returns, you need to have a product that actually works for the LP's that at the same time is a product that works for the venture ecosystem. A lot of the VC's have a product that's tuned towards one or the other, but not both. There are a lot of flaws when VC's face off against the startup ecosystem.

And I'll give one analogy here, one way of thinking about it, to point out why some of the VC ecosystem isn't tuned to actually serve the startup community. So if you think about how the VC world works right now you have an incredible number of companies at the base of the pyramid. It's kind of a cambrian explosion of startups because it's become easier and easier to create a startup to get from zero to one with very little money and very few people like you can take ideas and actually put them into the world pretty easily now, which means the base of the pyramid is really big. So there are VC's that their entire purpose in life is to basically be talent scouts. What they do is they look at the base of the pyramid and they say look, let me give them a little bit of money and see what they can do with it.

In some ways you can think about it as an author and say, here's an author that's never written a book before and they're writing a book with a story that's never been written before. So let's give them a little bit of money to create the first chapter. So let's get a writing sample out of them. So you have the seed stage, VC world, the angel stage, VC world. Like that stage actually makes a lot of sense.

It's about talking to authors and listening to the story they want to write and say, you know what, I think they deserve money in order to put a writing sample out into the world. There's an audience for this and this is the right talent to write. So you get a chapter back and you put it out into the world and you say, you know what, this is actually pretty interesting. I want to know what's going to come next. I like this story.

I want to see chapter two. I want to know how this story actually plays out. So then you have VC's that are like, look, we're chapter one specialists. We wait for you to get the product into the market. We can stare at a writing sample, we can see what you've done with the money.

We can talk to the investors that have worked with you, like, are you a good fiduciary of capital? Are you going to take this money and actually write a story? And if the answer is yes, like these stage one, these chapter one specialists, like, they will fund the money and say, let's go write more of the book. The problem is a lot of the chapter one specialists, that's exactly what they are. And when you get that writing sample back, not just the idea and not just chapter one, you now shout from the rooftops about chapter one to go find a chapter two specialist.

And then a chapter two specialist shows up and they say, you know what? We wait for the writing sample to get in place and then chapter one to get written. And then we help you write chapter two. And if chapter two is good, we're going to shout from the rooftop so that you can find a chapter three specialist. And the problem is, the advice that you end up getting from VC's because it's so staged, is you're guessing what the next person is going to want.

You're not saying, I'm writing a story for me, right? Because you might get a good answer and you say, I really like this story, but sorry, I don't have the capital, I can't give you more money to complete the story. The desired outcome is the next gatekeeper instead of the ultimate. So what you're doing is you're saying, I can only advise you on what I think the next person is going to want. Not what I want, but what the next person is going to want.

And if you do this 4567 times, like, the company can really be on a bumpy trajectory because they might actually achieve exactly what they wanted to achieve. And then they go to market and it's crickets, because what they ended up building was something everyone around the table agreed would be great. And then there's no chapter specialist that agrees that they want to actually put money behind that company. So does that suggest a lifecycle investor that gets in early and stays with a startup throughout? Or what's the solution here?

So there are a variety of solutions, but it also points out a variety of problems. One of the problems is that the cadence of building a startup has turned into what I call Alphabet soup. So you have the seed round, seed plus the series a, the a extension, the b, the c, the d, the like. How many rounds of capital do you need to go into a company to de risk it before you realize you have a real business? And again, it's turned into Alphabet soup, which is not the way it used to be.

And people will call me the grumpy old man, even though I've only been in this industry for 16 years. But I do remember when there only were four rounds of capital that went into businesses. You had a seed round, which was getting product into market and doing some testing to see if there was any pulse of demand in the market for the product. You had a series a, which was about de risking some of the critical assumptions around, can you distribute the product, can you manufacture the product with good unit economics? You had a series b, which was about scaling with improving economics.

So can I find a channel that I can efficiently put money to work to scale the business and start capturing economic rent and closing the gap in profitability? And then you had a series c, which was really growth equity. It was actually private equity at that point. By the time we got to the series c, which is about saying we can connect the dots between this being the company that it is and a profitable public company, and we're going to give the growth capital to make that happen. And growth capital might enable them to invest in new s curves, maybe geographic expansion or product expansion.

But you had a business that was working at that point. So I think part of the innovation has to come from the business side, where the concept of seven or eight rounds of capital going into a business that really does need to collapse, it does need to go away. It's actually very unhealthy for a host of reasons. Combination of lack of discipline and the amount of capital that's put to work before, you have a business that has actually proven that it should exist and is scaling and has good economics, a money machine at the end of the tunnel that's worth investing in. And you're hearing a lot of that now.

You're even hearing people talk about the flex might not be raising six rounds of capital, it might be raising one round of capital and being done. That might be the new flex. So with AI and with the ability to launch things on tighter budgets, it might be possible to actually have a team of ten people take you all the way to Nirvana, incredibly capital efficient, where you use retained earnings to actually grow the business. Imagine that. Imagine that.

So I think there actually will be some innovation on the business side about the cadence of how businesses are built. But I also think that venture capital has to change as well and be able to make commitments to companies where you say, look, if you accomplish the things that we collectively think are the right things to accomplish, there will be capital for you there. And it doesn't mean you can't go to market to find other funding partners who might also bring other skills to the table or network to the table, that would be valuable. But knowing that you have investors around the table that have the ability to invest in a business, when it's earning the right to exist, it's de risking itself and proving that it's building a good business, I think that's going to be incredibly valuable capital in the venture world. What stage going forward do you think has the best risk adjusted return profile?

So it changes, and it definitely is different. Based on geography, we're seeing very different things. But for any given market and for any given vertical, within any given market, there's always going to be what I've internally called the sucker round. And the sucker round is when you're overpaying for the progress that a company has made across the entire stage. Like, there are stages and there are eras where you can look and say the Series A was just strategically bankrupt to target, because the relative value relative to the progress that was made just doesn't make sense.

So this has changed a lot over time. And during peak Madness, I wrote a lot about this, where the difference between a seed round and a Series A was six to nine months. Right. What could you accomplish in six to nine months between the seed and the series A? Especially when the series a's were being raised at three x the price of the seed stage, like six to nine months later.

So when that was happening at Peak Madness, you were funding companies with a million dollars of ARR at 100 million pre and putting $20 million into the company. So that was what was happening during Peak Madness. Like, you would find these companies and you would say, oh, great. Over the past nine months, they grew from two hundred fifty k of ARR to a million dollars of ARR. Let's go fund them.

And the seed round was done at 25, and now you're funding it at 100. Like, where the sucker round is there in the market today. There are a couple cliffs around the Series A and around the Series B, where a lot of the companies raised money at those lofty valuations during peak madness that now they're coming to market and saying, look, we've grown by two x, we've grown by three x, we've proved that we have origination channels that work, we have unit economics that are starting to really show themselves as valuable. And the series a now might be priced the same as the seed round was two years ago or less. Yeah.

We're seeing for the best companies, right, the ones that really are growing by two to three x year over year, or at least two to three x from when they last funded themselves. A lot of them have grown into their valuations and you actually feel pretty good about a flat ish round. Got it. And we've seen a lot of our best companies now that did fund during peak madness. They're growing at flat to slightly up rounds, having grown three x over the past 18 months or two years.

Nick Moran
Yep. So you feel actually pretty good being on the other side of that divide. Right. If you are the series a funder, like you can now look and put a rational price in front of the company. The seed stage has remained in the US persistently stubborn in terms of high valuations, and I think it's because there's just a lot of capital, a lot of people, a lot of ideas that are there.

Frank Rotman
And there's this crazy thing called standard dilution, which I don't understand. Like, every business is actually different, the capital intensity of every business is different, the potential outcome of every business is different. But for some reason, the venture community talks about standard dilution. You say in the seed round you should take 20% dilution. Like, I don't understand it, but it is what it is, because a lot of these early stage founders are asking for a certain amount of money.

It's created a world where the valuation expectations really haven't changed when they should. The seed stage, especially domestically right now, is still persistently stubborn. It doesn't mean you can't find great companies and it doesn't mean you can't fund things. But, you know, that's a place where it's a bit high right now. Frank, you've invested in many companies from many funds over a number of years now.

Nick Moran
I'd like to get your thoughts on portfolio construction in reserves. I think it's easy, on one hand, to run a theoretical portfolio construction in excel and determine that no reserves is the best approach. I've done that. I think it's also easy to do some hand waving and just say, a one to one ratio is best. So that's the right strategy and we're going to use that.

And I've had many gps just advise me on that. Just stick with one to one and don't put too much thought or science into it. Yeah, I guess. To start, can you just highlight the key reasons why reserves are important? So the world of startups, by and large, is about investing in companies that are turning over cards to get answers before they actually can capture economic rent.

Frank Rotman
That means you're burning money. That is shorthand for these are companies that need cash. So when a founder is really allocating capital, they're allocating capital to one of three things. They're either building product, they're testing product, or they're scaling product and capturing economic rent. And only one of those three things creates enterprise value.

The other two burn cash. So if you think about the concept of turning over cards in order to get proof or anti proof that the business is on track or off track, it costs money. And there's only so many cards that you can turn over in a disciplined way within a certain amount of time and for a certain amount of money. So when you think about reserves or you think about making an investment, you're actually investing in a learning agenda, right? A lot of investors don't think about it that way, but that's what you're investing in.

You're investing in a founder, being an asset allocator, allocating it to one of those three things, and a learning agenda around those three things, and you're going to get answers. But a lot of those answers might be good answers that then require capital to actually put money behind the fact that you have a good answer so that you can scale the business or eventually capture economic rent and build enterprise value. If you put money in a company like you, you have to ask yourself what's on the other side of that. Is it another round of capital that they need to raise, or is it the last capital that they're going to need to get to profitability? Because you're investing in scaling, but you have to understand what you're investing in.

And if you're investing with no reserves, the assumption is that the company is going to turn over cards and it's going to be a set of good cards that someone else is going to see it as a chapter specialist that comes in after you and says, I'm willing to invest in the next part of the learning agenda. The more non consensus, the ideas, the less that's true, the more it's a consensus idea, the more you can just stare at the answer and you can say yes, like this is a good business, and we can see how you can construct the pieces and put them together. And if I put this money in, this is what happens. But guess where most of the return is? You would know better than me.

Investing in non consensus ideas, like ideas that you aren't guaranteed for them to be true. And that's in fact the case based on your track record with QED. Yeah. I mean, you have to be investing in something that has risk associated with it. And for that risk not to materialize, that's how you get paid.

So the more risk that you end up taking, the more things that have to come true or not manifest themselves as risk, the riskier the idea is. And if you end up turning over good cards, you should get paid for that. So we have a couple very large multi hundred x type outcomes out of the few hundred companies that we've invested in over the years. And they were very non consensus oriented investments. Got it.

They were investments that other venture firms didn't believe in. The market timing was wrong. Other people believed the incumbents were going to eat their lunch and there's no way they were going to be able to build the business. But credit karma was an example of one where they were turned down by pretty much every investor in Silicon Valley. Before we ended up investing in the company, it was coming right out of the global financial crisis.

And people are like, they're really going to build this business now? Are they really going to do this? And we were there very early, and it ended up being a $7 billion plus outcome to Intuit. And by the way, intuit loves owning credit karma. It's a great business.

Nubank was another one. It's a $50 billion public company now. And we had been talking to Daveed Velez when it was just an idea in his head. When he worked for General Atlantic, I was flying around south and Central America with him when he was a venture capitalist looking at a lot of things in south and Central America. He had this idea that someone was eventually going to build a credit card company in Brazil because of the industry structure and how it worked, and they were going to dominate the market.

And no one believed him because of the power of the incumbents in the space. He ended up leaving to go to business school at Stanford and then joined Sequoia afterwards. And he was so obsessed with this that Duglione kicked him out and said, here's a million bucks, go get the thing off the ground. You've been talking about this so much, you should actually go build it. And we were there actually helping him in those early days and we invested in the company as well, the very earliest stages.

And that's an example of a non consensus company that check after check, people kept looking at what was happening in the market and they said, are you really going to take on, are you really going to take on the big banks of Brazil now? It's a $50 billion public company that's growing incredibly fast at its current scale. Like it's going to absolutely dominate Latin America, if not other places in the world. So what you're saying is you've had to continue to support these non consensus businesses when there were not other buyers at these subsequent chapters. You either have to do that or you have to de risk the business very strategically and make the money last long enough until it become.

You convert it from a non consensus investment into a consensus investment investment. So like the best investments are the ones where you are the last non consensus check. Yes. And then all of a sudden people can stare at it and say, this is working. For Eubank, as an example, they had two years of growing by seven x year over year, 50 times the size that it was over a two year period.

Like the writing was on the wall that this thing was working. Of the first 6 million or 7 million customers that they ended up booking, only the first hundred thousand were originated using paid marketing. Everything else was viral. So like, it was non consensus. And then you actually could see the thing working and scaling, and then it became much easier to actually bring capital into the company.

But part of the use of reserves is when a company isn't that obvious consensus investment. There might be some asterisks around the company where there are some flaws that they still need to figure out. But you as an insider, instead of just having a snapshot of the company, you have a video of the company. And by having a video of the company, so much more about how the team operates, how the market operates, what's working, what isn't working. Like, you have so much more information asymmetry than a new player coming in, that theoretically you should be able to deploy reserves in a very strategic way with a lot more knowledge than an outsider actually coming into these companies.

So if you actually think about the companies that end up doing 50 x or 100 x or 200 x, if that's the ultimate return on the first check, as an example, if the first check is 100 x check, the second check, it's probably a 30 x check. And the third check, it's probably a ten x check. And the fourth check. It's probably a three X check. If you think about the goal of venture capital is to produce three X Plus funds, fund after fund.

If you have winners and you have information asymmetry, and they still have massive room to run in their markets, deploying that capital is the smartest thing you can do. So, Frank, what does that mean about the math? What's the structure used to forecast the appropriate amount of reserves to play defense, play offense, optimize returns? I want to say it's a science, but it isn't. I mean, it's experience.

You try to put enough money aside based on market conditions and the type of companies that you're investing in and the stage that you're investing, that you will have enough money to be able to follow on into the companies that are doing well and also to be able to support the companies that deserve a right to exist in the market. But they need a little bit of capital to prove out a few more things. The extension rounds are actually quite important, and having the capital to do that extension, in order to get them that extra proof point that they might need to be able to attract outside capital, that's very important. A lot of companies would die without that capital. So one way of thinking about it, and it's very hard to actually execute against in practice.

But if you can get 50% of the money of your fund into the top 25% of the companies in your fund, you're going to. So if I look at our track record, over 200 companies or so, we have, I think, 225, 250 companies at this point, somewhere in that range, about 25% to 35% of the companies. You're flat out wrong, right. The business ends up failing, the market ends up telling you that it was much more difficult to build the business than you thought. You end up with a pretty bad outcome, usually a zero on somewhere between 25 and 35% of your companies.

There's somewhere between 35 and 50% of your companies that you end up building a decent company, but it was really hard to build. And the outcomes aren't venture type outcomes. You might end up with a strategic sale for 100 million, 200 million, $300 million to someone out in the world, but you've probably put a lot of capital to work, so it isn't a giant return. So you've got 35% to 50% of your portfolio that you're probably getting a one x or two x or three x return on. That leaves about 25% of the portfolio that you do quite well.

So only one in four times you're actually, right. With your investment thesis and the company that you backed, and those are your five x plus investment returns. And of the 25%, there's about 10%. Right. Where you're really right.

So one in ten investments that you make, you're really right. And the key is, can you concentrate capital in that top 25% where you're getting a five x or ten x or a 20 x or 30 x return, right. Because even though you might be reducing your return on that logo basis, right, if you only had one check, it might have been 100 x, but if you have two checks in aggregate, it might be a 50 x, but it's on more money. And concentrating that money in your winners, where they take that money and they're turning it into enterprise value. That's the point.

The companies that are winning, you can actually point to the source of compounding value. If you can point to the source of compounding value and it requires capital, you can be the fuel to generate the enterprise value that then delivers the return that you're looking for. So the key is how quickly can you identify the companies that are going to be in the 25%, and then how good are you at identifying the 10%? And do they need money, and can you make it easy for them to get money, and can you offer them a fair value for that money to make it easy for them to take the money? Have you earned the right to put that money in the company?

By helping the company? So there's a lot that goes into being able to concentrate 50% of your money into your top 25%, which, again, it's very difficult to accomplish. Most funds come to the conclusion that reserves are a bad thing, and it's because they're poor at deploying their reserves. Or even their initial checks. If you're bad at the initial checks.

If you're probably also bad at the reserves, you have bigger issues. If the logos are bad at both. Yeah, yeah. If you're bad at the first, you're going to be bad at the second. But there are.

There are firms that are actually quite good at picking, but they're not disciplined, where anytime a company has a round that's coming together, they just automatically do their pro rata, or they see if they can do super pro rata in their companies, and you have to be very focused to make reserve strategy work. Frank, if we could feature anyone on the show, who do you think we should interview, and what topic would you like to hear them speak about? So one of my favorite people in the space is Matt Levine, he writes a newsletter called money stuff. He is just hilarious. He is so smart.

You could ask him anything about any topic in fintech, in banking, and he'll just make you laugh. Like his sense of humor. Maybe it's just me, I get his sense of humor, but you actually get these incredible analogies and perspectives on things that are happening, like, in the world real time. He writes it pretty much every day, and it's one of the only must reads. So if you're not following Matt Levine, anyone who's listening to this, you should be following him.

If you just want to laugh, but also learn at the same time. So you could ask him about anything, you should have him on the show. Frank, what book, article or video would you recommend to listeners? So this might be a little bit different than the answers that you get from other people. I always recommend people think about the things that they hold strongest in terms of opinions and then try to read the other perspective.

So as much as it might annoy you, irritate you, whatever word you want to use, and you might feel like you're wasting your time learning about the opposite perspective, I find it extraordinarily valuable to kind of challenge your thinking and make sure that you're seeing clearly. So I'm a first principles thinker. Like, if you were to ask people to describe me who know me well, like, they're going to, I would say 90% of them, that those are the words that they're going to use. Like, I collapse problems into atomic units and then I build things back with frameworks. And for me to do that well, I need to actually understand the opposite perspective.

So I think there are some interesting books that can really challenge how people think about things. One of them is like, the new capitalist manifesto is a really interesting book by a british economist, Umair Hake. I think you pronounce his name, and it's just a different way of looking at capitalism. And I'm a capitalist. Like, I actually do believe capitalism is a good way of running things.

But when you read what he's written, it does make you really think about it in a very different way. So that would be an example of it's not the opposite side, but it opens your mind to thinking about things. And steel manning your argument in different ways. Frank, do you have any habits, tactics or techniques that are a secret weapon? My particular habit is that I write a lot.

So anyone who wants to follow me, I'm at fintech junkie on Twitter, and I write a lot. If I can't write about something, I can't explain it. So a lot of times when I'm writing, it's for me, it's about sharpening my thinking. It's usually a result of having conversations over and over with founders about a topic, and I refine it every time. And when I can write about it, that's when I know I understand it.

Nick Moran
And finally here, Frank, what's the best way for listeners to connect with you and follow along with QEd? Like I mentioned in tech Junkie on Twitter, qedinvestors.com, as a way of contacting us as well. But you know, we are very open and easy to find. He is the legend Frank Rotman the firm is Qed Frank, thanks so much for joining us today and sharing your insights. I really enjoyed it.

Frank Rotman
Thanks so much.

Nick Moran
Alright, that'll wrap up today's interview. If you enjoyed the episode or a previous one, let the guest know about it. Share your thoughts on social or shoot them an email. Let them know what particularly resonated with you. I can't tell you how much I appreciate that some of the smartest folks in venture are willing to take the time and share their insights with us.

If you feel the same. A compliment goes a long way. Okay, that's a wrap for today. Until next time, remember to over prepare, choose carefully, and invest confidently. Thanks so much for listening.