The Exchanges

Q That first group of the creators of the stablecoin They're not paying interest. They're earning a lot. What is the structure that allows that to stay in place?

A There's been a bit of path dependency. So back when rates were zero, you saw a lot of growth in the stable coin economy because people didn't really care about earning yield. And also you could take those stable coins and deposit them on chain in one of these DeFi lending protocols and earn some sort of rate that roughly matched the real economy. But today, you're seeing a whole host of different firms focus on yield-bearing stablecoins or tokenizing treasuries, and actually BlackRock is a great example of this. They recently launched a product that effectively passes through treasury yields, less a modest amount of fees back to themselves in exchange for passing that yield back to users, and yield-bearing stablecoins as a percent of the total stablecoin market is growing. As it should. But look, if you're in Nigeria using Tether for B to B payments, you may not really care about the four percent opportunity cost of capital. One, you are likely outperforming your local fiat currency. And two, you're just using this as working capital, and it tends to be a pretty great user experience relative to using their traditional rails or cash.

AI assessment note: “There's been a bit of path dependency... you may not really care about the four percent”

Q Well, that's a wonderfully articulated why. I'd love to go through some of these INGs as you referred to them and try to get a sense of either in the actual protocol or what is working today and what maybe still needs to develop. And we can start anywhere, but what about lending and borrowing to start?

A So yeah, lending and borrowing is an interesting use case. So there are several like money market style platforms where lenders can show up deposit tokens and earn a yield on them. And they get paid a yield by borrowers. Borrowers may be deciding to borrow crypto because they have a low, a low tax basis. They don't want to sell to realize a taxable event. They may need working capital. They may want to go leverage long, and so these borrowers, they may post, like, say, two dollars of collateral for every one dollar they borrow, so these money markets are kind of governed based on different loan-to-value ratios that can get changed by these networks. Secured lending is a category within DeFi that is actually probably the most mature in terms of finding product market fit and having real volumes run through these systems, and very low If any principle defaults on the blue chip platforms, the one area that we're still figuring out is unsecured or under secured borrowing. So in a world of blockchain where code is law and these instruments are all bearer assets and there aren't the same rights and remedies you'd have as a lender or borrower in a chapter 11 bankruptcy, there's very little recourse. And so People have attempted to build reputation-based systems or other systems to pull in web two data to give people kind of a web three credit card or web three credit, and we're just s…

AI assessment note: “Secured lending is a category within DeFi that is actually probably the most mature”

Q What have been some of the most developed applications outside of the financial world?

A One of the really interesting areas in the space is called Deepin. So it's decentralized physical infrastructure networks. And the idea behind this is if you're a telecom company, you normally would have a lot of CapEx, but can you get people to set up hotspots to form a network and give them tokens so you don't have to spend that CapEx, but they do. And so this idea of externalizing CapEx is Has been what's underpinned the growth of networks like Helium, or other of these deep in networks. So you've seen a lot of experiments here, a company that is basically paying people to attach a dash cam onto their automobile, and drive around and collect data in real time, similar to what Google Maps would create on what's happening on roads. It's a company called Hive Mapper, and they've now mapped roughly 10% of the world's roads in just a couple years without spending a single dollar of capex. They're basically using their native token to incentivize this activity, and they're doing it in a manner that's much more efficient and quick than Google Maps historically was able to do it. It's this idea of, can you use a token to To incentivize people to perform some action in the physical world. This concept, I think, will eventually extend to other types of networks. If you think about something like, it'd be almost like if Uber paid drivers in the early days with shares of Uber, as oppose…

AI assessment note: “It's a company called Hive Mapper, and they've now mapped roughly 10%”

Q What are some of your favorite stories of something you learned from, say, trading that applied to something else you're doing?

A The best example of this was in 2019, the term DeFi didn't exist. We called it crypto finance. As part of our diligence process, we forced ourselves to use every application that we invested in because the best way to really understand something is to use it. We were early users of Uniswap, which is the largest decentralized exchange, Compound, which is a large decentralized lending and borrowing market, MakerDAO, fiddling around and using them. As we kept on using these products, we realized ourselves, wow, this is capital markets that moves at the speed of the internet. There are no transfer agents or fund admins. There are no credit agreements or bond indentures. Everything kind of just works and you click buttons and it moves. And just through using these products day in and day out, and we actually became power users of many of these early DeFi protocols, it produced a massive, massive flywheel for our business. One, because we understood what products were good. And we ended up investing behind those. Two, because we were power users of DeFi, we could help our private teams build towards where the puck was going and build great products because we were using these things day in, day out. We understood all the nuances and edge cases. And it also led us to develop different non-directional strategies in the space using stablecoins because stablecoins were the main like subs…

AI assessment note: “The best example of this was in 2019, the term DeFi didn't exist.”

Q As we dive into what's happened, particularly with these use cases, really curious about the talent piece, because since in the last three years, first you had this explosion of interest in the metaverse, and then of course, AI. A couple of years ago, crypto and blockchain was the thing that every young programmer wanted to do. How has the talent moved around?

A I think there are like two ways to look at talent. One is quantity of talent and another is quality of talent. Many people often look at number of developers or number of GitHub commits, ways to measure how much new code is being written in the space. If you look at those metrics, there's been some organic growth over the last three years. In absolute terms, but there's been a slowing of growth. That's on the quantity side, but on the quality side, you've seen pretty material uptick, not only in the quality of developers, but also the quality of non-technical individuals in the space. So on the BD side, on the product side, on the go-to-market side at the C-level of these companies, because I think what you're starting to see is There's a surplus of block space, so there's a surplus of blockchains that you can use that are fast and cheap today. It's not really the bottleneck. The bottleneck is figuring out user experience, regulatory, understanding of KYC, AML, and making sure these products are actually built and can be used by institutions. For those types of dimensions, you actually need people that understand finance, that understand how corporates think about using blockchains. Net-net, the quality of talent in this space, the human capital is better than it's ever been. I remember when we started in 2018, 2019, people building in crypto were, with a few exceptions, not ne…

AI assessment note: “on the quantity side, but on the quality side, you've seen pretty material uptick”

Q So what happened when you went to launch Parify?

A The timing in the short term wasn't great. It was really at the beginning of a pretty precipitous bear market. I'd taken a long term view, and I had really high conviction that blockchain would be a thing long term. But in the short term, the markets were down fairly dramatically in the back half of 2018. This was kind of the hangover following the ICO boom. It was very difficult To scale our business, to raise capital, to get institutions to actually care. I think majority of institutions wrote off the asset class. I kind of think back to like the early days as being both challenging and exciting. I remember, so Henry Kravis was a day one investor in Perify, and is still a mentor of mine. And I remember getting lunch with him Say like a year into starting Perify, and we really weren't scaling. We were sub-twenty million of AUM. Our performance was great on a relative basis, but not great on an absolute basis, and I came to him for advice, and I remember him saying, look, when I started private equity, or when I started KKR in the 19 seventies, the term private equity didn't exist. They were called bootstrap funds, and the concept of Buying businesses with debt was kind of foreign. It was really a new asset class. And he went around and talked to allocators. Like, they really didn't even understand what he was talking about. And their eyes would kind of roll into the back of th…

AI assessment note: “The timing in the short term wasn't great. It was really at the beginning”

Q All of that just naturally flows into crypto. So how did you get from that traditional corporate analysis and different forms of the capital stack to crypto?

A In 2014, I was living in San Francisco. I had a roommate who was working at Google, who's a dear friend of mine to this day. And he had some of his coworkers over for dinner, and they were talking about Bitcoin. And I remember hearing the word Bitcoin for the first time. Someone explained what it was. I was extremely skeptical. I thought it was somewhat of a silly idea. And my friend said, hey, you shouldn't come to a view on it until you've actually done your work. So I said, okay, what is doing your work entail? And he said, you should read the Satoshi white paper. I said, okay, I don't know who Satoshi is, but I'll read his white paper. So the Satoshi white paper is an eight page document that was written by a pseudonymous individual that describes the idea for Bitcoin. I think it was written in 2008. And I had to read it a couple of times to really digest what he was describing. But I found it interesting because I had worked in finance my whole career, but I had never really asked myself, what is money? Like, it's such a fundamental question, right? Like, what is money? Money is this, like, lubricant for all capital markets and for the entire financial services space. And digging into the question of what is money, I found to be an incredibly intellectually stimulating exercise. And quickly realized that Money is a technology that humans use to communicate value with one a…

AI assessment note: “In 2014, I was living in San Francisco... talking about Bitcoin.”

Q How does the big movement in AI intersecting with blockchain?

A There are a few different areas of overlap. One of the big areas that we're seeing is this idea of AI agents. The end state of AI is not an LLM that you put data into in a call and response format. But it's really this idea of an agent that can take actions on your behalf. Blockchains were almost purpose-built for agents. We often talk internally that people have asked, when are the users going to come to blockchains? When are the users going to come? Well, we think the major users of blockchains just may be AI agents themselves. The reason why blockchains are really useful for AI agents is because they are And AI agents cannot open a bank account. You have to be a human or a company to open a bank account. But if you're an AI agent and you want to send micropayments to other AI agents or micropayments to humans, that's really not possible on traditional banking. You really have to go on chain. If you look at the way this AI agent economy is evolving, you have flows between humans and AI, but the biggest category is AI to AI. That type of activity from first principles, I think makes the most sense on blockchains. And so you're starting to see a lot of AI agents use blockchains, use DeFi, send stable coins to do things. And while blockchains are difficult for humans to use because the UX isn't necessarily intuitive, that's not an issue for an AI agent. They can figure it out.

AI assessment note: “One of the big areas that we're seeing is this idea of AI agents.”

Q do call it venture equity investing. You're actively engaged in kind of arbitrage type trading. You buy tokens. In the traditional markets, you often think of these as very different skill sets, right? There's a venture investor, there's a arbitrage hedge fund, or whatever it is. Is that the same or is it different as you are participating across the spectrum of kind of traditional activities in the crypto landscape?

A I think the lines are blurred between tokens and companies. Sometimes we'll invest in a company and they'll end up issuing a token. Sometimes we'll invest in a protocol, but a lot of value ends up accruing to a company that's really helping build products on top of a network. The actual exercise of understanding those two is very similar. The liquidity profiles can be very different. The risk return can be different, but I think the muscle groups that you're using are In terms of evaluating teams, products, markets tend to overlap quite a bit. On the credit side, in terms of using these products, market neutral returns, that is helpful in the sense that we end up becoming users of the networks that we're investing in, and when you use something, you learn a lot, you understand it more deeply. It's a different type of investing. It's more quantitative. It's more technical. But it does help inform how we make decisions and how we build theses across the board.

AI assessment note: “I think the lines are blurred between tokens and companies.”

Q So we've had this big sell off in the prices of tokens, crypto assets, and I'm curious for those that are generating cash flows as you're looking at them, What are the range of multiples if you're comparing it to equity multiples that you see in the DeFi space?

A We've seen things that have traded at like one times earnings or two times earnings. We've also seen things that are trading at a hundred times earnings plus, and all of them have very differing growth rates. I think that it's helpful to be mindful of fundamental analysis in crypto, but you also don't want to rely on it too much for a specific reason, which is in crypto, things are changing very quickly. You can have a protocol that's growing, you 10 X year over year. And then all of a sudden earnings fall off a cliff because of some like dramatic shift in the landscape. Because of that volatility of earnings, you have to ascribe at a lower multiple. Because if you're looking out like five years, 10 years and saying, how much cash flow will this protocol do? Well, we can make assumptions, but we know we're likely going to be wrong or our confidence interval is much lower. So you have to ascribe a higher discount rate to those cash flows very far out. So because of that uncertainty, I think DAOs probably trade at lower multiples than companies, at least in the near term. Now offsetting that, these networks tend to not have any capex. There's no balance sheet. Some of them have very high cash flow conversion. Some of them are effectively a hundred percent margin because they'll just clip a little fee of every transaction. And they kind of run on their own. So there's a wide range…

AI assessment note: “We've seen things that have traded at like one times earnings or two times earnings.”

Q What's an example of something that others commonly in the space thought to be true and you found wasn't by nature of using these applications?

A I'll kind of give one example. I mean, there's so many ways to answer that question. I think a lot of time there's a quoted yield in yield farms where it say, hey, if you deposit this token, you'll earn a hundred percent APR. And what I think the market has done a bad job at historically is understanding the risk associated with that return. The capital tended to aggressively move towards the highest return, almost agnostic to risk. When you really dig into these stated returns in yield farms, one, there isn't always the right disclosure around the risks that you're taking, but also a lot of time the return is coming in the form of a token that may not have liquidity, and there may be lockups associated with it. It's really messy, and so like just taking a return at face value and not thinking about all the other factors that go into it is, I think, something that the market Wasn't doing well for a while and will hopefully start to do a better job of.

AI assessment note: “just taking a return at face value and not thinking about all the other factors”

Q So I'd love to go through aspects of each of these three categories. So if the fiat money backed stable coins, is there leverage in that system?

A So if you look at something like Tether, they come out with some transparency reporting. The last I looked, about 80% of the assets backing USDT were in cash. The other 20% were in short-term securities, T-bills, some commercial paper. And really the way I think about fiat backed stable coins is they're almost akin to IOUs or credit issued by the institutions that are issuing the stable coins. So if you own USDT, you're taking tether counterparty risk. How do you price that? What return do you need to earn to, to make that worth your while? If you own USDC, you're taking kind of circle counterparty risk, and you're making a bet on their risk management. It's no different than having a checking deposit with JP Morgan or Bank of America. You're ultimately taking the risk of the issuer.

AI assessment note: “about 80% of the assets backing USDT were in cash. The other 20%”

Q Which two people have had the biggest impact on your professional life?

A Henry Kravis is one of them. He was a day one investor in Perify and was a mentor to me and believed in me. I think when someone believes in you, it's very humbling and empowering. I think he believed in me. I'm not sure how much he believed in crypto at the time, but he respected the entrepreneurial journey. And so that means a lot to me. The second, I've had a number of bosses over the years that have been extremely challenging to work for, and extremely neurotic, and have held me to a very high standard. It was painful to work for them at times, but I don't think I'd be where I am today if it wasn't for being in the trenches with them.

AI assessment note: “Henry Kravis is one of them. He was a day one investor in Perify”

Q How do you think about the underwriting process and valuing tokens?

A So yeah, tokens are really interesting because they're very much a new asset class. They're the fifth asset class in the world. Pre-tokens, there were equities, fixed income, currencies, and commodities, and really every asset could be put into one of those four categories. Interesting about tokens is they're really a brand new Category. There are thousands of tokens now that exist, and what's interesting about this market is there's no gap financial statements, no 10 K's or 10 Q's, no investor relations department to call and ask questions to. You can ask people in a Reddit forum or on Discord or talk to core developers, but because there's no person calling the shots, You have to talk to a lot of people to really understand what's going on. But this market is one where non-standardized information, the information is very fragmented, and it takes a lot of work to really get your arms around what these networks do, why a token accrues value, and really what a roadmap looks like, and what the team around these projects looks like. So The way I look at our research process is we look at tokens not dissimilar from the way someone at KKR or TPG, the way we would evaluate a company. So we are understanding, like, is there a solid use case and product here? What do the competitive moats look like? Is this company or product gaining market share, losing market share? And why? What's …

AI assessment note: “we look at tokens not dissimilar from the way someone at KKR or TPG”

Q How far along are we in that adoption?

A This is a gradually then suddenly phenomenon because tokenization is a network effect technology similar to the telephone. If you're the only one that uses a telephone, it's not that useful. One other person, it's still not that useful, but the utility of a telephone scales exponentially. With the number of participants on the network. Metcalfe's law. Similar with tokenization, if you're the only one that holds a tokenized treasury, it's not that useful. But as more and more people do hold it, as you benefit from all these features of tokenized assets, it becomes increasingly useful. And that's why when we start to see these long tail assets reach meaningful penetration rates, that's a meaningful catalyst. Because if you're an institution participating in one of these markets, you're pulled on chain because it's more useful, because it's the place where price discovery and liquidity takes place.

AI assessment note: “This is a gradually then suddenly phenomenon because tokenization is a network effect technology”

Q In the sports betting world has become very big and regulated. How do you think about polymarket and the prediction markets relative to the regulatory regime around it?

A Regulation is coming very fast and furious to prediction markets in the US. There's been a couple like pretty important landmark cases. Of web to prediction market companies like Kalshi. But I think that it's going to be very difficult for traditional sports betting platforms to compete with prediction markets at scale. These sports books generally like set the odds internally, they'll have like an internal bookkeeper, and then they'll do their own risk management. And if they get too much action on Duke, they'll adjust the odds to be more favorable for UNC and vice versa. The beauty of these prediction markets is they're just software. There's no actual market risk that's being taken. There's no balance sheet at risk, and you're just putting these odds out there into the world, and there's a central limit order book with bids and asks, and people are trading these things in real time. The interesting thing about polymarket and prediction markets on blockchains is that they're global, so they benefit from global liquidity pools, 24, seven, three 65. Versus these centralized platforms where you have to, like, wire in money, wait for your money to clear, and you're typically only servicing users within a specific geography with artificially set odds. It's a more free, open capital market, which I think produces better insights and more liquidity.

AI assessment note: “Regulation is coming very fast... But I think that it's going to be very difficult”

Q We've had a series of crypto winters, even just in the years you've been running Parify. As you look out, are there pockets where you are concerned about risk?

A Always. And I look at the crypto space, I think, like, 99% of things in this industry are not investable, but one percent are. The amount of dispersion that you see and the amount of creative destruction is massive. There's nothing systemic that I see right now where I would say, like, this is going to cause an earthquake in crypto markets. There are a couple risks that I think people have been talking about for a while. One of them is Binance. One of them is Tether. The transparency of those organizations. In today's market, crypto is two and a half to three trillion. We're below where we were in crypto market cap in 2021, almost four years ago, despite all the progress in the space. And there's never been more real economic applications of blockchains that are relevant in the world today than there are right now. And there's never been more institutional interest in the space from the investing side. I think the question is flipped from, why should I pay attention? Today, it's more, why shouldn't I have exposure? Oh, why don't you have exposure to crypto? Because having zero exposure to the space is almost a view in and of itself. This has been the best performing asset class of the last 1510, and five-year periods. Not having a position today, despite the increased regulatory clarity, despite the institutionalization, despite the use cases, I think is becoming a bit more of …

AI assessment note: “There are a couple risks... One of them is Binance. One of them is Tether.”

Q If we need this permissioned DeFi in order to scale, I'm curious what that implies for the large percentage of the world you mentioned earlier that are unbanked. And does that somehow the need for KYC type things for certain types of trading come in the way of those people entering a banking system through DeFi?

A Potentially. I think there's always a trade-off. We've decided on a global scale that like money laundering is a bad thing, and we don't want criminals, terrorists interacting with the financial system. So there is a trade-off between just being totally open and letting everyone interact with Financial products and being inclusive. And it's interesting because sometimes when you think about this state of the world, many would say, well, what's the point? Isn't that the world we have today? The whole point of this is that it's permissionless. And the reason I'd push back is I would say that's one benefit of DeFi and blockchains is that they're censorship resistant. They're permissionless. Anyone can use them, but there are a laundry list of other really Important benefits as well that I would highlight. So one, you have instant settlement when you're using DeFi. So it's not T plus two settlements for equities or T plus 20 for bank loans. These are all bearer instruments. So there's no global working capital drag from all the unsettled trades in the world. It happens instantly. That's powerful. The second is that Within permissioned DeFi, there's still reduced platform risk. So we all saw what happened with the LME when they canceled nickel trades, or what happened with Robinhood and GameStop when they simply said, you can't buy GameStop. And there is judgment risk or platform ri…

AI assessment note: “Potentially. I think there's always a trade-off... but there are a laundry list of other”

Q So your firm, I think when we sat down three years ago, was probably roughly the size it is today with typical with the cycle, a big dip and a big recovery. How do you think about continuing to maintain and grow a business over the next couple of years?

A We have a super long range view on building this business. There's a lot of short termism in crypto, a lot of medium termism, but there's very little super long term focus. There's a cultural answer to that question, which is like every decision we make, every hire we make, we want to have that very long term alignment and build towards what we want our business to look like 10 plus years from now. But really the core of it is as more and more institutions enter the space, And want to generate interesting risk return profiles and also want to understand this ecosystem. We want to be the institutional partners for them. And then really the North Star is crypto is a secular trend. People think about crypto as a cyclical macro asset, but we really think about it as a long-term secular trend. And we want to invest in real use cases that are bringing GDP onto blockchains. That has to be the North Star. Everything really comes back to that. It all has to come back to what are financial applications or other corporate applications for this technology that actually make sense? Actually unlock 10 X better use cases because they're on chain. That has to be our North star as a firm. And I'm just a believer that in building a business to build something extraordinary, you just have to do something consistent over and over and over again. That Kaizen mentality of just get better and better …

AI assessment note: “We have a super long range view on building this business.”