Re如何把金融市场最稀缺的收益来源之一——再保险——带到DeFi｜DeFi Frontier

That's kind of the bigger idea. It's Lloyd's of London, but scaled up for the internet age. Lloyd's, again, is just a marketplace. It's a pot of money. It has 103 different insurance companies that sit on top of it and specialize in everything from nuclear risks to aviation and space risks and property risks of every sort.

That's how you get to insane scale, I think. It has kind of decomposed this and decentralized it in such a way that's accessible to every other sophisticated reinsurer in the world.

Today's show features Karn Saroya, the co-founder and CEO of Re. Karn, thank you for joining us. How are you doing?

I'm great, gents. Thanks for having me.

When I think of Re, Re is a platform to give us all access to this gatekept yield opportunity that is reinsurance. If you've never learned about reinsurance, this will be a great walkthrough to better understand how it works, why the yield is real, and why that yield is so powerful when packaged through the Re platform.

1. Building insurance tech before Re

I think it's a great example of a regulated fintech that has the DeFi mullet setup, where DeFi is ultimately powering the back end and the whole global capital pool of lenders. Karn, why don't we get started with more about what you were building prior to Re? You've been in insurance or reinsurance for longer than the start of Re. I think this is part of the moat behind Re.

You can't just set up a DeFi protocol and start selling reinsurance. There's a regulated component to the Re protocol. Tell us more about what you were doing before.

Yeah, for sure. I don't think insurance is known to be the sexiest thing, but I've been in it for a while. I'm now coming up on, I think, year 12 of building technology into the insurance space.

Like other technologists and folks who have built in a variety of spaces where they perhaps didn't have expertise, I started off relatively naively. I didn't know anything about insurance. I was in my 20s, and I pitched Y Combinator on an idea for an insurance app that took a picture of something and got folks insurance, simplifying the entire consumer experience behind buying coverage for your house, your car, your pets—just about anything you could visualize.

My co-founders, who were actually the same co-founders who came over to Re, didn't have prior experience in distributing, underwriting, pricing, or working through the mechanics of matching risk to capital in the insurance market. We got hard and fast lessons. Within a couple of weeks of launching out of YC, Apple featured us as a best new app. We were seeing tens of thousands of people send us pictures of their wedding rings, their cars, their homes, and their pets.

In San Francisco, we saw pictures of tents that were being sent to Burning Man. Just about anything that you could think of was coming in the door. The consumer behavior was really surprising to us. We needed to figure out how to match those requests for insurance to people who could provide that insurance.

We were desperately picking up the phone, calling brokerages and agencies around the country, asking them to take on these customers who we clearly couldn't provide insurance to, and then routing the customers. What we quickly learned was that there was a significant disconnect between how people wanted to buy insurance and how they were ultimately receiving it.

Agents would take weeks to get back to folks. It was really, really slow, and drop-off would be significant. So we ended up building a national insurance agency from scratch. In 50 states, we distributed products for all sorts of insurance companies, and I got a good flavor for what direct-to-consumer acquisition looked like for insurance.

Again, we fell into the same kind of trap: The software that underpinned most insurance companies was pretty archaic. If you guys have tried to buy insurance on the internet in recent memory, oftentimes you're punted to a call center. There are very few opportunities that are seamless.

We saw tens of thousands of people coming in every day to buy insurance from us via our apps. We thought, "Why don't we go build an insurance product ourselves?" We were moving from pure distribution to figuring out what the guts of actually building an insurance company looked like—getting regulators on board and getting balance-sheet capacity, the money needed to write insurance, in place.

As part of that process, I got really good at building out the functions of an insurer. We hired folks who had run publicly traded insurance companies. Importantly, we interacted with reinsurance brokers and reinsurers so that we had the capacity to write insurance business.

I've been one of the few people on Earth who have stood up an insurer from scratch and then also made the bridge directly into DeFi. Of those few people, most are on our team at this point. All of this is to say that I've seen everything from top to bottom in insurance, all the way from distribution to this risk-transfer business.

2. What is reinsurance and why it matters?

I think that really resonates with me, just what you said about how getting insurance is not the smoothest thing that you'll do. From car insurance to home insurance to business insurance, it's never really fun. Most people listening to this probably understand insurance. It's just part of our lives; we've grown up knowing you have to do it, pretty much.

But I think very few understand the term reinsurance. I want you to try to explain reinsurance from a first-principles basis. Why does it exist, and why has it become such an important part of the global financial system without many people even knowing what it is?

For sure. I'll use a simple example. Let's say a tree falls over and destroys some part of your roof or your home. You have a home insurance policy, and you've been paying insurance premiums to your insurance company. That insurance company, in turn, has made a promise to pay if certain things happen.

As part and parcel of that promise to pay, there needs to be proof that they can pay. You've been paying in premiums, they've been investing those premiums, and they've been earning a profit. Is the money still around?

What reinsurance is is actually a financial backstop and guarantee that the insurance company is going to be good for the money that it owes to its policyholders. The insurance company itself is going to have a balance sheet. It'll have a pot of money that it pays claims from. If those claims exceed that pot of money, either for a specific program or across the board, the reinsurer steps in and is now part and parcel of this transaction.

As a result, the insurance company is actually paying premiums to the reinsurer. At the end of the day, we are an insurance company for insurance companies.

Ultimately, you're telling us that the major retail-facing insurance companies that we all might have worked with, such as State Farm, have a fallback plan if their losses exceed whatever their model predicts that coverage to be.

That's correct. It may not purely be the case that it's an excess-of-loss type of deal. It might just be the case that you're an insurance executive and you've made the determination that you're too concentrated in a particular region, or you want to grow more quickly and need more money to be able to do that.

Reinsurance is effectively a capital market for the insurance market. You could go write the insurance business, write the policies, and then offload those insurance policies—or, rather, the associated risk of those insurance policies—to a reinsurance partner or panel of reinsurance partners. It's additional diversification and a means by which the insurance company can grow.

3. How big is the reinsurance market?

Karn, how big is that addressable market for reinsurance?

It's massive. I mean, it's a trillion. There are 2,000 admitted insurance companies across the United States. There's a step function more than that across the world that support, you know, a trillion—there's $7.5 trillion of insurance premium written across the entire insurance market.

Then there are all these other collateral markets that are basically predicated on whether you have money to fulfill the promises that you make out in the world. These are letters of credit and trusts that are pledged to all sorts of commercial purposes.

You end up looking at nearly 15% of the global economy being related to showing people that you have the collateral necessary to make the promises that you've made out in the real world. It's a huge part of the global economy.

4. How Re brings reinsurance capital onchain

Okay, so we know what reinsurance is now, and we know how big the market is. I think you're saying there's a trillion dollars of premiums done a year for reinsurance.

I think the next question is, why bring it on-chain? We've seen Treasury bills get tokenized and private credit get tokenized. Now it's happening with reinsurance, but why is this attractive? You've toiled away at insurance for years, so why bring your business on-chain, and why now?

Yeah, we were talking about Nexus Mutual and other folks who had tried this in prior instances to bring insurance on-chain. The reason is actually—if you step back as an insurance executive, it slaps you in the face really quickly how analogous the insurance business is to the basic function of DeFi and blockchain. We are literally saying, “Here's a contract that pays out under specific conditions. Here are the triggers. Here is the money that pays out under those specific conditions.” You can analogize that to smart contracts, or to the basic staking of crypto-economic assets against adverse developments out into the future.

In some sense, a lot of what happens on-chain is actually insurance, and it made sense to me to find a vector of attack that bridged the gap between the two. I think where folks have stumbled, or had at least challenges in scaling, is that rather than trying to bring an archaic, conservative business like insurance on-chain and forcing people who've been doing business the way they've been doing business for 70 years, 80 years, in some instances hundreds of years, onto this new architecture, why not bring the specific things that DeFi offers—composability, transparency around the underlying capital flows, and capital verification—directly to the insurance market?

These guys still require capital to run their businesses. There's capital on-chain. That seemed to me like the path of least resistance, and that's proved out to be the case.

What we're seeing now is that we have 51 different reinsurance treaties across the United States that are using digital assets to support insureds across the country. There are 30-ish insurance companies, half a billion in business, going on a billion in business. It is not a startup; it's not a toy anymore. It's really at the inflection point where this zooms to tens of billions in premium over the course of the near- to mid-term.

That means the implication is that digital assets are going to be supporting tens of millions of business owners, people trying to get to work, people trying to buy homes, people driving across the country, and essentially the rest of the world. That's massive, and it's plumbing, right? It's not super obvious. It's kind of just now starting; it'll end up being foundational.

5. The DeFi mullet and capital efficiency unlocked

Re is such a good example of what we refer to as the DeFi mullet, and I think what you're describing—a permissionless capital pool that is ultimately powering Re—is the operational efficiency unlocked. That's capital efficiency unlocked. Can you just talk more about what that means for the Re business? Are you this much more profitable because of the way that you're able to raise capital from lenders on-chain versus the way that traditional reinsurance companies would ultimately raise money?

Yeah, for sure. So, in 2 parts, I think, 1, we are very privileged to be able to start a de novo reinsurer from scratch without legacy systems and legacy business. We don't have clients that go back 150 years and send us reports scratched onto a tablet that we need to make sense of. Everything is kind of new.

We don't necessarily need the same sort of headcount to do all of the back-office work and the back-office reconciliation of all this information that's coming from hundreds, if not thousands, of customers with all sorts of different reporting. That's a really big drag for traditional insurance, for traditional reinsurers and insurers, that we just don't need to deal with.

The second part is that we get to build this at a technology inflection point that basically squashes and solves really cleanly the problem of taking unstructured data in every form, structuring it, normalizing it, drawing inference from it, and using thinking computers to help price risk, sit within an underwriting box, do the work, assign the capital, and do the reporting. All of that stuff has now been made possible by LLMs over the course of the last couple of years. We sit at the frontier of that, and we get to take full advantage of it.

The way that manifests is that, right now, today, we're still subscale at 500 million, but we write roughly 10 times as much premium per head, per employee, as a traditional reinsurer. That's probably going to end up looking like 50x over the next little while.

If you think about what the expenses associated with provisioning reinsurance or insurance are, they're going to get squashed. It's going to become an increasingly small part of the overall economic stack. The next question is going to be: how much granular control do you have over your cost of capital? That's the other part of this product, right? You need collateral to write insurance business.

Now what you have is these dynamic capital pools that can grow, they can pay, they can be tranched, they can be tokenized, and you can create secondary-market liquidity in the manner that we have. Those are just far superior to what existed in insurance today, which is a static pot that's governed by a couple of insurance executives who say, “Here are the risks we're going to attach to it. And, by the way, my bonus is predicated on how much of this capital I can sit around, even if I can't deploy correctly.”

You now have an audience and depositors who demand a particular yield, and they want the verifiability and transparency around what's happening with that underlying capital. That's just a higher standard, actually, than exists for the traditional insurance market.

So, fundamentally different architecture, I think, and way, way, way higher operating leverage. If you zoom back a little, this is just like Tesla being able to stamp out Model Ys at a lower marginal cost than any other auto manufacturer. We're literally just able to stamp out the product of reinsurance at a lower marginal cost than any other reinsurer in the world.

What does that mean over time? It means that if you just survive, you win the market, right? And that's the sufficient condition, actually.

6. Growing revenue with just 12 employees

I just want to linger on this point for just a second, because I really think this is the whole ball game right here—what you're talking about. I think this is why things move on-chain: this efficiency you're talking about, this low headcount. I think, before we move on, you're under 12 employees at Re. To scale your business to billions, I don't think it requires you to hire hundreds of people, right?

What kind of revenue per head is this right now? And I guess, to grow your business to billions, maybe answer that question: do you need to scale up a lot, or does this stay lean and efficient?

Yeah, we're at 12. You're right. It's about 30 million per head today, so we're profitable, which is great. I think that probably understates the efficiency. In the near term, it's probably going to be double that, maybe more.

Where I focus my energy is actually—there's clearly a need for some redundancy. Twelve is pretty thin still, and so we're working on making sure we have redundant capabilities on the technical, product, and design front. A lot of my energy is now spent on identifying reinsurance and insurance folks that have the muscles that are necessary for us to scale into the rest of the market.

Right.

I'm not thinking about a billion because we're going to get there, right? I'm thinking about how we get to 5 billion, how we get to 10 billion. Who are the folks with the expertise locked in their heads that we can pair with technologists to enable this crazy growth? That's where I spend my time.

Even that doesn't mean more than a dozen people, actually. It means you look at the top 100 people in the reinsurance market, filter for energy and enthusiasm around technology, and look for a vision for the future that involves digital assets. You start to whittle down that number pretty quickly.

Once you've got the core, you pay for that core, monitor performance, and go on from there. I don't think the reinsurance piece gets to more than a couple dozen people. I think if we expand the scope of what we're doing—if you have this programmable capital layer and all this software that's built on top of it for insurance companies and underwriters around the world to be able to access—you have a clean segue into, hey, there should probably be an effort, and maybe it's us, maybe it's not us, to go build an AI-underwriting-compatible stack.

There should be AI-enabled insurance companies sitting on top of programmable capital. That's my vision for where this is going to go. Now, what does that mean, very specifically? It means Einstein is literally going to take in all the information that's necessary to get you insurance.

It's going to come back and say, “Here's exactly what your coverages are. Here's your house, and here are your assets that need to be covered. You've got kids; you need life insurance. You've got a small business; you need an umbrella policy to go with that. Here's exactly what your risks are. Here's how you need to hedge those things.”

“Here's the package I put together. I provisioned the capital from an insurance company, printed exactly the policy that you needed, and provisioned the capital on-chain to be able to support that.” So, you have comfort that if anything does happen, you're good for it. The insurance is good for it.

It's not going to be Marsha or Kevin or whoever it is who's thinking about this every year and not having a fulsome view of who you are and what your business is. It's going to be in real time: what are the risks that you face, how are those risks hedged, and what are the instruments that have been provisioned to hedge those risks for you? That's the future.

7. Re’s strategy for winning new business

I'm curious: for the reinsurance buyers—your customers—are these folks ultimately buying coverage from you? Are they similar to, I guess, how some of us would operate with mortgage rates, depending on reputation and a bunch of other factors, knowing that they can do business with you as a newer reinsurance company?

If you're able to offer rates that are just 10 or 20 bps lower, is it that much of a magnet for new customers to Re?

We actually have not started to compete on price. A lot of the excess economics have just been internalized for the benefit of Re, depositors, and the rest. That piece is certainly a switch that we could flip at some scale.

If you spin off tons of cash and you want to send it, and you're willing to pay the insurance company—or at least lower your rates such that the insurance company takes more of the pot—you're going to win more business. There's no question about that.

There are channel-conflict challenges because a lot of this business is brokered, which makes that less effective than we would otherwise have assumed in a perfectly efficient market. Because it's a brokered market, there's an agency problem and a whole bunch of other stuff that comes out of that.

That will not change my viewpoint. Where we really win is that we post capital to the insurance company itself. We're fully collateralizing our obligations, so if things happen, the insurance company knows that the money is there. It's there.

A traditional, rated reinsurer signs a piece of paper and says, “We're good for it.” Its auditors, on an annual cadence, say, “We're good for it.” We don't necessarily have real-time attestation to our ability to pay, and you're not going to have a very good understanding of what the rest of the risks are that we take.

What we've argued is that cash in hand is way better. Complete verifiability around what risks this capital pool is taking is better. Speed is way, way better.

If we can get back to you in half an hour when someone else would take 2 or 3 weeks to get back to you, we're going to win way, way more deals. You speak their language, use the same rails to conduct business as they've always used, find the highest-performing folks in the space whom they respect and want to transact with, show that the capital is there, and go out from there.

That's been the winning strategy. Don't innovate too much by forcing consumers or businesses to change their behaviors. That's not an errand or battle that you want to fight. You just want to meet them where they want to be met and do everything better.

Right, that's literally it. I want to reiterate a few things that you've called out. Just listening to you talk and trying to conceptualize how Re is winning in the reinsurance game, it's clearly efficiencies: way lower head counts and things like that.

8. How Re differs from competitors

It sounds like another differentiator is that traditional reinsurance outfits don't always prove that they have the capital ready at hand. They kind of give you a “Don't worry, we got you.” That's another one: you can sit on this capital base that's ready whenever it's needed.

And then another one is that everything you're doing is fully auditable in virtually real time, and people get that peace of mind again that, hey, this money is going to be ready when you need it. What else would you fill in the gaps here on anything else that you see as how you're disrupting the status quo of the reinsurance model?

Again, we don't want to make our customers uncomfortable. They've done business a particular way. Everything else is just the internal machinations of a reinsurer being refactored.

If you think about what this business does, some information is captured up front. People are trying to make a determination as to, hey, what should we charge you to cover claims and expenses should you make a claim, and how do we deliver a profit margin? At the end of the day, that's what this is. There's no tangible good; it's an intangible product and a contract.

Computers are going to be able to rationalize basically every single step of that, with the exception of potentially the human relationship involved in selling insurance. Some people are just going to want to talk to a person, and it's not going to be a chatbot.

That will become an increasingly smaller part of the market as you move forward into these future decades, but all of it's getting rationalized. Our point of view is that we just build for that inevitable future.

I don't think there are that many insurance executives in the world who are saying, “We should just set up on top of a verifiable capital lake, with endpoints for eight agents to be able to tap capital. Send us information, tap capital, and get it provisioned.”

It's, “Hey, build for the architecture that you think is going to happen, that you think is going to be inevitable, and be prepared for it.” That's basically it. Do it with as few high-leverage people as you possibly can, and you get further than you expect.

That's the whole lesson here.

9. The capital stack behind Re

Karn, can you start to walk us through what the capital stack looks like? There are 2 offerings that we've referred to before, which are reUSD and reUSDe. Then there's actually a third pool of capital, which for me was the lightbulb moment when I was learning about Re: understanding that I'm not the junior tranche in either one of the first 2 I mentioned.

Talk us through how that all works—who's senior, who's junior, and so forth.

For sure. The critical thing in getting all of this to work is incentive alignment. Everyone talks about incentive alignment, but for our purposes, we've got all of our assets tied up as first loss.

We've got this massive safety pool, which is our assets. It's earned premium. It's our equity that eats any sort of volatility in the underlying insurance portfolio first. So, we can credibly say that if we win, you win; if we lose, you're probably still going to win.

This layer ends up being the shock absorber behind all the stablecoin tranches that we've finally set up. It's a massive pool that will inevitably be larger than the size of any sort of insurance or safety pool for any stablecoin or DeFi product, period. This is going to be very, very large.

In fact, I think right now it probably rivals, if not exceeds, the size of the safety pool for Athena and for the Sky ecosystem. And we are pretty subscale.

Above that, we have a mezzanine layer that has quarterly redemptions that are actually gated. It takes on some insurance risk, and you make the trade-off between liquidity, the yield that you earn, and risk remoteness.

Above that, we have reUSD, which is a senior overcollateralized tranche that has at least a 50% holdback for redemptions at any given moment. That’s the product that is getting most widely integrated across protocols, with approvals from curators and the like. As you move up the stack, you get increasingly remote from insurance risk and increasingly liquid.

10. Duration mismatch and how Re navigates this risk

All of this is to say, the way we cracked this nut is that we’re the duration. At the end of the day, you couple us being the duration with these liquidity sleeves, and you’ve now got a machine that makes sense for current DeFi architecture.

Karn, you mentioned the different durations of these different assets, and something we’ve seen in the space that’s been killing protocols is this duration mismatch. I feel like it stems from when a protocol maybe doesn’t educate users well enough: “Hey, look, these are duration assets that have certain redemption criteria.”

We’ve seen some protocols sell a bit of a bill of goods to people, and then all of a sudden everybody wants to redeem, and they’re like, “Wait, I can’t.” Or there’s no DEX liquidity. How have you been navigating this with reUSD? I believe that’s a quarterly redemption vehicle. Speak to this a bit, because I think it’s a huge issue in our industry, and I’d love to hear how Re has been navigating that.

Yeah, for sure. Our own assets are predominantly the duration. They’re locked, and they’re the ones absorbing most of the insurance risk. reUSD is a mezzanine layer that’s still protected by all of our assets and by our equity. The reUSD product is intended to be made available and released as the actuaries determine that collateral can be released.

As you move up the stack toward reUSD, you get increasingly higher levels of liquidity, and commensurately, you get paid more. Depending on where you are on the stack, you get paid a certain spread that reflects the illiquidity risk.

You’re right that you can look at what cash we have sitting around, our ability to service redemptions, and what we can pay. Some part of it is going to be encumbered—the vast majority of our own assets—for the purposes of meeting something called a minimum regulatory capital requirement. That’s the number that ultimately we have to meet. It’s what the regulator says we have to have and what the customer says we have to have as collateral. Theoretically, everything else is redeemable above that stack.

We’re doing our best to make sure that, whenever we enumerate every single deal we’re party to, we understand what the capital requirements of those deals look like. Folks can formulate an opinion about where they want to sit. Ultimately, you get paid a higher spread for the illiquidity and the quarterly redemption cadence.

I think you guys had pointed out, “Hey, folks could simply step in and purchase these products if they ever trade at a discount.” There are huge hedge funds that already do stuff like this. There’s Stone Ridge, which makes an entire business of this, knowing that there’s significant capital protection sitting downstream of that. It’s a timing thing, but of all the RWAs, I think we’re probably the best positioned to solve for this.

We spin off tons of cash. This is a business that generates a lot of free cash flow. You can see all of it coming in, and you can see all the deal flow coming in. There’s a predictable pattern of collateral release. We know that if you sell an insurance policy on January 1, it earns over the year. As it earns over the year, the risk is gone and collateral will come out the door.

If you sell a policy on December 31, it’s going to be earned over the following year. It’s going to follow a similar pattern. You can approximate what the collateral needs are for this in a way that’s not disconnected from what the underlying business is.

Before we saw any capital formation, we spent nearly a year educating allocators on exactly how these cash flow patterns worked in reinsurance, what the underlying risk was, and what protections were in place. Importantly, the only way we got this to work at all was, “Hey, we’ll put up our own money”—and a lot of it—in order for the rest of this machinery to work.

11. Re junior tranche and what’s it earning

That’s been our solution at the end of the day. We know what the minimum capital requirement is. Theoretically, the rest of it is releasable, starting with reUSD, then the mezzanine, and finally our own capital, which is not moving anywhere.

Karn, when you mention your capital, I think you’re referring to that junior pool. How much money is in that? Do you know the approximate yield that the junior pool of your capital is ultimately earning, versus reUSD, which is the most senior and is earning, I believe, about 6.3% as of this recording, while the mezzanine tranche is earning about 12.3%? I’m just trying to understand the difference here. Is that all the fees that Re takes ultimately? Are there other fees for doing all of the business development involved in getting reinsurance clients, doing the underwriting, and so forth?

The only fee that’s charged today is a redemption fee of about 6 bips. We’re probably going to end up getting rid of most of that, too. There aren’t any other fees. There’s no rake on the spread or anything like that. We have a defined fixed spread for the senior layer and a defined fixed spread for the mezzanine layer.

That junior capital tranche, right now, with earned premium and equity, sits around $80 million, something like that. The return on equity for that sits at about the market return on equity. It fluctuates depending on where the market is, but right now the reinsurance market is in the mid-teens to low 20s. That’s where it is.

One thing we spoke about before we started rolling is that, when you talk about that massive piece of junior capital that you have sitting there, to me, that’s a moat. If somebody were to say, “Hey, this reinsurance business on Chainlink looks really good. Let me get into that,” or, “I’m going to fork a protocol,” there are a lot of impediments to that.

The traditional behavior we’ve seen in DeFi is that, if something’s working, you see 10 or 20 of them crop up. But that massive pool of capital you have, I think, is a big impediment to that. I want to get into this because there aren’t many other protocols doing what you’re doing. There’s another one out there, but I’m curious how you think you’re differentiating yourself from other competitors that we might see come into the space or that are already in the space.

What separates Re from other people who are seeing this and want a piece?

Yeah, for sure. I think this is an incredibly difficult business to execute. You have to navigate DeFi, a regulated insurance market, and you have to get the customer to ultimately buy the product—the insurance company that’s buying the insurance from us—which is a dance.

Across all 3 of those things, you have to execute perfectly to scale. There’s no copy-and-paste version of this. You’re not going to be able to say, “Re did this, and now we’re going to go stand this thing up.” You have to execute on 3 different fronts and 3 different businesses in order to make this work, which is a huge moat. It’s just difficult.

We are the only ones in the market that have shown that we have the ability to scale. We’re whitelisted with every reinsurance broker of note in the world—all 5 of the top 5 and 8 of the top 10. We’re actively doing business directly with insurance companies across the country, those relationships are deepening, and the size of our deals is going up. The execution has been exceptional on that front, in a way that we haven’t seen anywhere else in this market.

On the DeFi front, the quick ramp-up to a couple hundred million in deposits is also significant. There’s really no reason why this shouldn’t be a multibillion-dollar TVL product over time. It’s tapping into a yield source that is near infinite, completely uncorrelated, and relatively predictable. As we mature through this market cycle and mature in our understanding of this business, it should be a no-brainer, in my mind.

Not financial advice, none of this, all right? But this is just from my objective chair, looking at this and looking at the rest of the overall DeFi market.

The other important thing here is that we’ve indexed aggressively toward conservatism to start. When we’re out there and we say we have all these assets as first loss, we do.

There's nobody else with assets that sizable in first loss in this market. When we say that we're low volatility, that is true. All of our business is low-limit, low-volatility lines of insurance business across the United States. This is auto insurance, home insurance, small-business commercial insurance, and certain parts of workers' compensation.

No floods, no fires, no hurricanes. If they are in there, they're de minimis and fully funded by the economics that we charge. And so there is no circumstance here where you have a binary outcome, right? If a hurricane hits Florida, we're not going to show up and say, "Oh, we lost all of your money," because that's not the type of business that we write.

And so, for our current state of the business, that makes complete sense. I think if we hit a multibillion-dollar run-rate business, it makes sense for us to titrate in more of the overall insurance market. You have these higher-margin, higher-volatility little pieces of the overall portfolio, as opposed to making that your bread and butter, which I think is difficult.

It's tough because I don't know that many people really understand the types of risks that they'd be taking if those are the products that they were buying. And so, as a differentiation: our tilt towards conservatism, first loss, complete incentive alignment, and then exceptional execution on just getting the insurance market to accept us. We're still one of one there.

I know we've made you restate this a few times, but the thing that has really been shocking for me to better understand is that the junior pool, acting as first-loss protection, has $80 million in Re that is ultimately there as a shield for me as an LP in, let's say, reUSD or reUSDe. So I've been thinking for some time—part of the reason we do these podcasts is that we're doing research out in the open—we are sometimes actively looking to deploy our own capital into certain DeFi protocols.

And, like a lot of lenders, I'm tired of earning T-bill rates, and DeFi has taken a lot of hits this year. This mashup of what Re offers as a regulated sort of fintech with a DeFi mullet backend, but offering that reinsurance yield, is huge to me. So, reinforcing: is that the best way to think about it? There's $80 million there that's acting as a first-loss cushion.

Yeah, it's first loss for insurance risk because most people will just be candid, right? Nobody—there are very few humans on Earth who know how to underwrite insurance risk. And so to ask an LP to figure out how this book should perform and fit, and think about where you want us in the stack, becomes a little easier if you say, "You're the one to take the first hit," right? And not just take the first hit, take the first hit to a significant extent.

And so that's been the innovation here. It's to get the market comfortable in scaling. There may be a future here where folks are able to take a junior position. It's very possible, but we just don't think the market is there yet because there aren't that many people who want to underwrite insurance risk.

12. AI and its potential impact on Re’s future underwriting

Another question for you. I don't know if I'm just injecting something here, or if all this AI hoopla and hype is bleeding into my thoughts too much, but I've seen countless instances on podcasts or interviews or what have you where some smart guy is just like, "Yeah, there's not even enough capital for the AI build that we need. We need more money. We just don't have enough money."

And I look at reinsurance, a capital-intensive industry, and I'm wondering if there's any bleedover from this massive suction of capital that's happening in AI—whether dollars are becoming harder to come by for your business, just seeing this once-in-a-generation buildout that's happening.

Yeah, we haven't seen that, actually. And so I think they're quite disconnected from a financing perspective. In some sense, they're both kind of like secured lending to some extent, but the performance of 100 million insurance policies moving in any given direction converges to predictably really, really quickly. It's a law-of-large-numbers business, right?

You have a relatively more nascent GPU and data-center financing opportunity, which is important and I think should be funded, but the market is going to clear on what the acceptable market price is. It's going to clear on how much capacity should ultimately end up getting built.

I think animal spirits are in play, and whether we are overbuilding or underbuilding, I'm probably not the right person to say. I think more computers are generally a better thing, but what the appropriate price and funding structure of that is, I think I'll leave to the machinations of the market.

But all this is to say there isn't that much crossover. To the extent that there is some crossover, these data-center build-outs and GPUs are looking for insurance—builders' risk. From what I've seen, folks look at us as any other shiny new object. There's a ton of capital from a reinsurance perspective that is now moving to secure these risks.

Whether it is appropriately priced or not, we do not have a position because we don't do that business. I think if I were a guessing or betting man, I would argue that because it's a shiny new thing, the risk is probably underpriced. This is my guess.

13. Re’s role in the growing market of RWAs

Yeah, sounds like you're not in the reinsurance of AI data centers or the insurance policies that back them. I do want to talk about another major thesis underpinning the growth of DeFi, which is real-world assets. We're seeing lots and lots of new protocols and exciting headlines about tokenizing real-world assets. Where do you think Re fits into the broader RWA movement?

Yeah, I mean, it's an asset class. Insurance is one of the largest financial markets in the world, right? There's no reason why the capital tied up in every insurance company and every insurance policy in the world needs to be in little pots and little silos. I think all of it could end up being represented on-chain, and it improves this market massively.

And that is trillions of dollars, right? This is an order of magnitude greater than the overall stablecoin market. But that's to say, as we grow up, the rest of the market's going to grow up, and you're probably going to still end up with a similar ratio of where the dollars end up, right?

Could insurance and reinsurance be 10 to 15% of all of DeFi at scale, when everything else is scaled? I think so, right? There's really no reason why that wouldn't be the case.

The unique feature of insurance and reinsurance is that it's actually a natural capital-formation vehicle, right? As opposed to requiring retail demand or institutional demand to generate stable digital assets, insurance and reinsurance ends up being an engine for that formation.

To give you an example, you stick $1 of collateral in an account, and you can now write $4 or $5 of insurance business. Where does that $4 or $5 go? That could find its way back into stables that sit in a trust account, or it could find its way into other digital assets that earn a return in a transparent way.

It ends up being a force multiplier on the aggregate amount of digital assets that are out there floating in a way that very few RWAs can do.

14. Agentic operations within Re

I promise I don't just ask AI questions, but I have another AI question for you. I liked that segment earlier when you were talking about these agentic underwriters, and I want to get into that a little more. I'm curious how much you yourselves at Re are exploring this avenue.

We've been talking about this agentic wave on this podcast for years now. It hasn't happened yet, but I do feel like we're always on the verge of it, and I feel like we're even more on the verge of it now than ever before. But speak to a bit more about how that's going to change your business and ultimately probably your industry going forward here.

Yeah, we use agents internally to do all sorts of things: to take in data, to conduct actuarial analysis, to do write-ups on underwriting, and to make recommendations on capital provisioning.

A lot of that stuff is already happening with us at the reinsurer level. Where I think it's more effective is a step above us: the insurance companies themselves. My stylized example would be there's a CFO of a company—let's just call it mid-market. Maybe it's S&P 500. They have products out in the world, and they need product liability insurance. They have executives, so they need directors and officers insurance, errors and omissions, and all the types of things that you would typically need if you're running a business.

The way things work today, they would go to a traditional reinsurance broker on an annual cadence. The broker would come in and do a presentation and say, “Here are the risks to your business, here's your run rate, and here's all the insurance that you need to buy. We'll go out and shop it for you.” That happens again on an annual cadence.

The future is literally that CFO interacting with superintelligent thinking computers that already have access to payroll, their enterprise resource planning, and everything else that gives that computer a fulsome view of what that business is, how it's performing, and what the risks are. It's then going to figure out, “What do I need to hedge? What are the existential risks to this business? Do I need product liability in case my products are faulty in some way and I get a class-action lawsuit? How do I cover that particular risk?”

It's going to come up with all these risks and the coverage levels that are appropriate for the size and scale of your business as of today. Here's what it's going to look like as you grow according to your growth plan. I'm going to print you an insurance policy that stipulates exactly what these coverages are, and then I'm going to provision with an insurance company and reinsure the capital that's necessary to back it, and you're good. Not only that, you're good for the entire future because I'm going to reshop this at every interval where there's a meaningful change to the underlying business—all in real time.

That's a massive change, because what I described is like 30 cents on the dollar in insurance. Thirty cents on the dollar—the expense load, the origination load, regulation and taxes, and everything else that actually doesn't come back into the pocket of the policyholder—that's going to get squashed. My specific vector of attack is to build the layer that shows where the capital is and how the money moves, and then start to think about how to interact with these things that are going to pop up in the future.

I can already use these tools to make my operations much more streamlined. I think that the massive effectiveness is going to come upstream with us. That's kind of where my head is at on this.

15. Tracking the most important growth metrics at Re

So, Karn, you have a dashboard at app.re.xyz/metrics. This is a dashboard that we've been tracking just to understand: How is Re growing? Are you successful? Can you talk us through what metrics are most meaningful to you, and what is indicative of Re being successful as a protocol and growing?

For sure. You've got all the traditional DeFi metrics, so you can see the metrics dashboard and kind of where the cash is. But if you want to decompose this down to how this thing is actually performing and how it's going to be as a going concern, what matters are these 2 numbers: total premium written and, ultimately, what the treaty performance is.

You can see on a line-by-line basis exactly what insurance deals have been written. What you want to track is book loss ratio and combined ratio, which we publish. You want a combined ratio that's in the 90s or the 80s, implying that profit is being generated and that the entire thing is self-sufficient. In other words, depositors are going to be paid and the capital is going to be paid.

For me, that's the North Star. It's literally: Are we making money on the insurance, and is the insurance business growing?

When you say “treaty,” just for folks who aren't familiar with that term in reinsurance, what does that mean?

It's no different from your own policy. If you bought a home insurance policy, a treaty would just be that but on steroids. It would be tens of thousands of homes or hundreds of thousands of policyholders. It would just be a big block of risk, as opposed to an individual policy.

One more: Could you go back up to the top of your dashboard? There's a metric that I had been digging into before. I was on Claude trying to better understand this. Premium receivable—what does that mean, and how is that indicative of the growth of Re?

This is a number that's related to deals that have been inked. What you would see is, on a monthly or quarterly cadence, insurance companies sending us that premium. The reinsurance premium receivable is going to end up being earned and end up as capital that's available and sitting in trust. That's what's going to happen.

The receivable number is actually going to continue to grow because we're still writing more business, but you'll see more cash flow into Re itself over time. It's an important number as an indication that the business is still growing. The other thing that's really important to note is that we don't post collateral until we receive money from insurance companies, so there's no credit risk or credit mismatch there.

16. What’s needed for hundreds of billions to come onchain?

Karn, as we wrap up here, talking about growth, I'm curious what you think is needed for hundreds of billions of dollars of insurance capital to start moving on-chain. What are the bottlenecks to that future playing out?

This is a neat segue into the protocol function itself. Re Labs owns a regulated reinsurer. It provides capacity, puts up its assets and equity, and is the first client of a protocol that forms capital for this purpose.

If you want to move from us writing $5 billion or $10 billion to the entirety of the reinsurance market, what we need to do is find a way—and we are, because we're building it for our own client—to plug this capital into any underwriter, reinsurer, or insurance company in the world seamlessly. Just because we're one client of this thing doesn't mean that we can't have essentially thousands of other clients around the world.

That's the bigger idea. It's Lloyd's of London, but scaled up for the internet age. Lloyd's, again, is just a marketplace; it's a pot of money. It has 103 different insurance companies that sit on top of it and specialize in everything from nuclear risks to aviation and space risks and property risks of every sort.

That's how you get to insane scale, I think. It is about decomposing and decentralizing this in such a way that it's accessible to every other sophisticated reinsurer in the world.

17. Overview of the RE token

Karn, before we wrap up, I'd be remiss not to ask you about the RE token, which went live pretty recently. What can you tell us about the RE token? What is the utility behind it? Just a quick TLDR for folks who are trying to understand how it fits into the protocol.

It's intended to emulate the Lloyd's governance council. The Lloyd's governance council stipulates who the acceptable counterparties are, what capital they need to post, what economics they need to have at risk to be aligned with the network, and what fees are charged by that network to access the capital.

Ultimately, the network—the Lloyd's market itself—is the market of last resort, so it absorbs tail risk and is compensated for that. The idea behind Lloyd's is exactly what drives the idea behind Re. It's just that Re is built for the internet age. That's the difference.

Instead of a governance council made up of sirs and knights, it's folks who hold RE governance tokens.

And then just remind us, Karn, if you can: When did the RE token go live? If someone holds the RE token, is there any staking they should take advantage of, or any other recommendations to maximize their usage of Re?

A lot of the functionality—I mean, voting on counterparties and so on—is about to happen. There is a staking mechanism, but this is in phases. We've got a client that's going to scale this thing into a couple billion dollars in premium. We're going to bring on other counterparties and reinsurers and continue to scale it.

When we're trying to assess who those folks are, their acceptability, the collateral that's necessary, and what they're charged, that's when the RE governance token starts to really shine.

18. Closing

Karn, I think this is a great place for us to start wrapping up. We really appreciate you coming on. I think that Re is one of the best examples of fintech building that next-generation capital stack using DeFi. In a bear market, we badly need to hear these kinds of stories.

This is the foundation, the bedrock, of the next generation of on-chain finance, and it's really cool to see how Re is bringing that all to life. We'd love to have you back in the future. Keep up the great work building Re, and I want to give you the final word here before we go.

Yeah, for sure. I really appreciate the opportunity to talk about my life's work. I think this is one of the biggest opportunities in DeFi. This is civilizational technology. Right? It enables risk-taking everywhere by folks who want to take risks and build things. It is only natural that it ends up on-chain. We're excited to do it.