How Spark Is Rebuilding DeFi Lending | Sam MacPherson

Transcript

The short answer is we uh for for the emote is we're just not on boarding anything except for lido. Like we just the goal is actually to make this the uh you know the safest possible construction for ETH yield because I think this is actually what most people want. They have their ETH and they want to get a yield on it. They want to remain liquid and so they just want like to toss it into the lowest risk yield generation. Nothing said on Xerox Research is a recommendation to buy or sell securities or tokens. This podcast is forformational purposes only and any views expressed by anyone on the show are solely our opinions, not financial advice. Pocachio, Ryan, and our guests may hold positions in the company's funds or projects discussed.

>> What's up, guys? Welcome back to another episode of Block ZeroX Research. Today, I'm joined by Luke from the Blockworks Research team as well as Sam from Spark. How you guys doing? >> Doing good. doing great. Happy to be here. >> Thank you guys for coming on. Um Sam, you have done quite a few Blockworks shows already and I assume majority of um people are familiar with you, but maybe before um we get started, you want to give a very quick intro of yourself and then of Spark as well and then we can dive into the questions. >> Yeah, sure. So, uh I'm Sam McFersonen, co-founder CEO of Phoenix Labs, which is the main contributor to Spark. Uh so you could think of the relation like the labs co and the uh protocol. Um so spark

is within the sky ecosystem uh formerly maker dao. Uh I've been involved in the project since 2021 back when it was Maker and D and there's been a rebranding now it is Sky and there's been a splitting out of the governance structure into uh Sky as a wholesale uh credit um issuance platform and uh Spark as one of the subdows which exists along a few other subdows as more on the commercial lending side. >> Okay. And then um maybe as a bit of background on that because the uh as you mentioned maker shifted to sky um apart from I guess like this reorganization um what are some of the main changes

that have happened um such that with sky now? >> Yeah. So I would say there's two main things that happened. But one was the rebrand and the reason behind this is that um you know I think a lot of DeFi there's there's a big love for D the legacy brands there but as um you know the space is maturing and uh moving into tradi it was um decided that we would take the short-term pain of uh having to establish a new brand in exchange for getting a USD in the in the ticker uh because it's just like it's one less thing you need to tell people already um learning about stable coins. So, you kind of want less things that people have to learn about. So, having USDS uh is just a much stronger brand uh for

people to be aware of, oh, it's a stable coin, right? It's it's quite obvious uh what what USDS is. Uh so, that was one thing. And then the other is this uh governance restructuring and uh that was largely due to uh iteration on the way the DAO uh Dows are structured. uh maker like uh there's been a few atscale Dows um but it was I actually worked within the the old DAO structure this is like the uh token plurality voting um token holders are involved in basically every decision um it just like we tried it for a while it just it doesn't scale at all so what would happen in practice is there would be like politicking and um

you know lobbying for various efforts based on you know if you like people or don't like people or you know it just it wasn't like a robust uh governance structure for um you know running sort of atscale lending operations and so um there was a fork in the road and it was like we could either kind or the Dow could kind of dissolve and switch into more of a classic corporate structure or we could continue down this uh decentralization path and attempt to do uh governance at scale in a um as decentralized a way as possible. And so this Sky restructuring uh was was deciding that Sky is a sort of a unopinionated wholesale credit issuer,

stable coin issuer and uh all Sky defines the rule sets and this is largely based around uh like basil framework that banks use just define the rules at which the um subdows are able to uh lend with and then the subdows are able to pursue whichever growth opportunities that they see fit. Um and and then as long as they're within the guidelines of the the Sky risk framework, they can deploy capital um that is uh uh that that has a junior capital uh skin in the game element as well. So that anything that they're deployed into, they do have this skin in the game element. And then be once a competitive market forms with all the subdows, they all have the equal access

to the sky balance sheet. This will be a decentralized uh mechanism at the whole in the way that free markets are decentralized. >> So Spark sits under Sky as a prime agent subdow. So how would you describe that relationship in practice? What does Spark control independently and then where does Sky still set the terms and and how does this interaction work overall? >> Yeah, so um [clears throat] I've been in both governance structures. So previously in uh the maker dowo structure uh to do any initiative um you would you would have to do like a whole governance process for two months to do even the smallest things and it just

like it was just totally not working. Um now in this new structure the the subdows are prime agent subdows are are free to um you know I don't want to say do whatever they want but because they have to follow the risk framework but whatever the uh opportunities that are available within uh the risk framework uh they can pursue and so the risk framework will just uh detail what the um it details a couple things I mean there's concentration limits um for you know counterparties collaterals etc. Um there is uh junior risk capital requirements. So uh for example uh if you have like DeFi lending with like say ETH on Ethereum, this is probably like one of the lowest risk uh activities you

can you can pursue. Um this is like this is weighted quite safely. You know the junior risk capital requirements may be uh even close to zero. Um [clears throat] u maybe a little bit more risky is something that has like a introduced counterparty from an issuer like say Lido or Rap Bitcoin um or CBBTC. U these are maybe one step further but uh Bitcoin is a highly liquid crypto instrument and uh I think lending against Bitcoin and and ETH and LSTs has proven to be quite a robust and safe yield sources as long as the collateral is kind of with kept within blue chips. Um, if you go out the risk curve a bit more, maybe you want to do uh private credit or um, you know, basis

trade stuff, there'll be uh, junior risk capital requirements attached to that. So maybe uh, you want to go into some um, private credit or whatever. And then the junior risk capital requirements are 10%. That means as the uh, prime agent, you can only get at maximum 10x leverage on your own equity. So um you want 100 million of exposure to this, you have to put up 10 million of your own capital that is first loss if there's some incident with this allocation. So yeah, >> so the past uh past month or so has been particularly impactful for DeFi regarding the kelp exploit and the illquidity of some of the largest

markets on a um wondering if you could talk us through the spark liquidity layer. um how it's built, what it's for, but then also particularly how it managed its exposures and allocations um over the past month during this um this DeFi incident. >> Yeah. So, the Spark liquidity layer um is primarily allocated to our internal lending market uh called Sparkland. I I believe there's about 1 billion uh allocated right now there in uh stable coins. Uh it is it is our preference to be in sparkland because precisely because we have a strongly we have full control over the risk management and as as a general rule spark does take a more conservative view than uh some of the

others out there. Uh we are uh sparkland is restricted to um what I would call blue chip collateral. So, you know, a couple of the highly liquid bitcoins, um, ETH, uh, Lido stake ETH, we have Etherfy ETH. Um, but in particular for this loop leverage stuff, uh, Sparkland, uh, only had, uh, uh, Lido stake ETH as as the collateral asset for ETH. We recognized uh back in 2122 uh there was this uh growing demand for yield on ETH and this was primarily going to come from staking yield. So we do recognize that the rehypothecating the ETH backing um unlike what MakerDow did. Maker Dow CDPs did not

rehypothecate the ETH and you know some people want that but there's more of this sweet spot where some people are okay with a little bit of a liquidity trade-off if it means you get most of the yield uh from staking. And so we positioned Sparkland as exactly that. Let's give the uh what we can construct as basically the safest yield we can provide to ETH. Uh the most uh liquid uh staking token is Lido. And so we restrict the efficiency mode uh to uh just that within Sparkland. Now for the broader um it's not just sparkland that spark liquidity layer goes into we also allocate to morpho and we have at times gone into the uh a core market although um we uh tended not to do that so much unless we really had to uh just because

uh we viewed the uh risk return on that uh market uh as as not uh particularly good for what the yield was. So yeah, any sort of allocation there's a question. It's not like is it good, is it bad, is it too risky. It's like what is the risk adjusted yield? This is always the calculus that you're going for. And so this will be taking the absolute yield and then discounting it based on the perceived risk that we have like the expected uh losses over time. So we run this calculus for everything and so uh it's it's just a matter of uh doing this. So we've had exposure to [clears throat] um various higher risk stuff. I wouldn't consider it super risky, but um and we would wind this up and then as soon as it the yield dropped

below our threshold, we wind it down. Um so yeah, we'll use a lot of like morpho infrastructure for this stuff because they have the isolated markets. You can unwind the positions quite easily. Um but yeah, I mean uh this is our general philosophy just slow compounding uh growth over the years. Don't be chasing after things the next shiny thing all the time. um and just reducing the attack surface. uh this specific incident with kelp I mean we did have it on boarded into uh sparkland granted not in e- mode uh we never on boarded in e- mode and we are actually very careful with an emote is basically like if people are not familiar with the avi market style it's basically like high leverage so 92 93% LTV which makes like

you know 10 15x leverage um this is high so more it's kept at this like 70% number and you it's really targeted at more like boring stable coins So we offboarded kelp in January as just a general policy of reducing our risk surface. Um but even if we didn't um the damage would have been quite minimal because Sparkland has rate limits. We've had this since pretty much the beginning. Um and the amount of uh like there there's no way to exit in size through ETH because of the it's not an emote. So um yeah, I think the damage even if in the in the counterfactual where we didn't offboard it would be minimal and and probably even no damage

to the protocol but easily coverable if there was some losses. >> Makes sense. I have uh two follow-up questions. One is how do you decide what asset you want to put in e- mode and what asset you don't? And the second one is um why did you guys decide on offboarding RSE back in January? The the short answer is we uh for for the emote is we're just not on boarding anything except for lido. Like we just the goal is actually to make this the uh you know the safest possible construction for ETH yield because I think this is actually what most people want if they're they have their ETH and they want to get a yield on it. They want to remain liquid and so they just want like to toss it into the lowest

risk yield generation. And so we're gearing uh I think a has been that in the past but uh now like we're really differentiating sparkland with just we really want to produce just sort of the safest possible yield for ETH depositors. And actually this is why I think we're seeing a lot of the growth now. Sparkland ETH deposits have have doubled in the past uh past month since the incident. uh precisely because my thesis is that people just like they want this uh you know low as possible but you know we'll still capturing most of the of the yield from staking. Um so yeah we uh and so for the um yeah the second part uh which oh why was it offboarded uh the the kelp down RSSE

um yeah so this this is just a general sort of offboarding uh policy we have so during bull markets some opportunities present themselves uh and they're generating a reasonable term to justify onboarding um at limited size like we're always very careful about the concentration limits of particular counterparties collateral. So, uh it was onboarded but yeah in a very constrained way and only like for borrowing stable coins and then that uh return was not justifying the uh the tail risk of of having this collateral on boarded. So, uh just as a general like monthly checkup on the collateral we we decided um it was decided to offboard this.

>> Makes sense. I have um I also want to ask about Ephina and specifically you guys winding down um your Ephina exposure. Um I'm curious what the decision um decision- making process behind that was. Um and I'm curious how you guys managed to wind down exposure um quite quickly. >> Yeah, I mean I feel like my answer is going to be similar for this. It's like there was a calculated alpha on the risk adjusted return. Um so uh we deployed uh quite heavily into Athena up to a limit.

It was uh I think it got it peaked at about 1.2 billion or something like that. Um which was I think always less than the uh than 20% of the balance sheet. Uh but that was only like for a week or two when there were there was I think this was like shortly after the the DAT announcement or something like the ENA price spiked and there was just like massive uh returns coming from this. I think upwards it was double digit percent even in a senior position on on Morpho as the lender. Um so this was just quite a good return. This wound up very quickly and then it we wound it down just as quickly when this opportunity ended. So yeah the answer is like the riskadjusted return uh from our perspective dropped below into a negative alpha territory. So we just

unwound uh the exposure and the mechanism and actually this is a the strong a very strong argument for isolated markets. Uh when we were onboarding this we never onboarded Athena into uh into Sparkland precisely because we this looping uh use case is actually stronger more more strongly served in isolated markets because of the uh better pricing uh controls that isolated markets offer. um in spark lender poolled lending markets in general uh you can't differentiate on the rate that you're charging. So if somebody is, you know, has ETH and they're borrowing stable coins, they're getting the same rate as somebody who's deposited SUSD and is also borrowing those same stable coins, but this is not

the comparable risk. The ETH the ETH uh uh borrower is actually far lower risk in our view. So we when there's these higher yield, higher risk uh products, we only spin them up in in isolated markets and we actually don't do the looping use case at all in Sparkland. We we want to differentiate it from that. It's mostly surrounding uh volatile crypto assets borrowing stable coins with the one exception of ETH because I think there's enough demand for this getting this base yield on ETH that it warrants an exception. But as a general rule, we don't do any looping in Sparkland. We keep this on our partner Morpho and uh their their tech uh we've worked with them like right after they launched Morpho Blue. We were one of the first ones into this market. Uh we had

our internal teams audit it 600 lines of code. I re I read it myself and I'm like I have extremely high confidence in this. We can go into size fairly quickly. And Athena was actually one of the first use cases where we were deploying there at size. This was back at the beginning of uh 2024 I believe. So yeah, just also as Athena was was winding up as well. >> Yeah, perhaps a logical followup on the conversation of um pursuing like the highest risk adjusted returns and how you go about adjusting for that risk. Um particularly because like the nature of DeFi lending has very distinct risk surfaces from what we might see in uh Tradfi credit. um whether it's uh bridge

exploits, multi-IG compromises, um AMM logic, flash loans, oracles, things like that. So yeah, what's what are the like the main pillars of how you go about um building a risk framework for um these credit instruments? And we also see you recently went live with Credora um with risk ratings across six different Spark savings vaults. So how do you actually see users and depositors like using these risk ratings in practice?

>> Yeah, so I I'll start with the risk framework. Um you are exactly right. I think RWAS is like is very different to underwrite versus uh like DeFi credit. Um yeah. So how our risk framework is oriented is a is it's split into four pillars. We evaluate um the collateral asset uh the uh custodian like who is custodying the assets and for smart contract land this would be uh like governance. So if you're on sparkland or uh you know spark savings this is all ultimately custodied by sky governance as the admin authority. Um you'd similarly like with if you're on a A DAO is the custodian of the assets. Um if

you're in qualified custody um so for example with with say uh Anchorage we do uh OTC lending through Anchorage. Anchorage is the custodian. Um another pillar we have is uh so who is the entity facilitating the uh lending? Um and then um so so this is like uh this is more for like an off-chain thing like what is the entity that is like uh like managing the loans facilitating the loans. Um and then [clears throat] the fourth pillar is the borrower. So this is something that's actually uh uniquely different in uh in tradi versus uh defi.

In defi you actually uh it's like completely anonymous sometimes. So you actually don't even know your your counterparty. A lot of the big counterparties are kind of known from chain analysis, but like uh you know if if your lo if the position becomes insolvent on some lending market, the borrower can just leave and like they've they've won, they get free money basically. Um this is completely different than like TRDFI venues where the borrower is legally obligated to repay the loans no matter what. So counterparty becomes a much bigger uh part of the calculus if you're if you're engaging with more traditional uh players. So uh yeah, we so we use these four pillars as kind of like a a a general uh structure for the the way

that we do credit underwriting. Um and this is at the sky level. And um yeah, basically uh [clears throat] there's questionnaires that will be filled out uh by uh usually it's the the prime agents that will do this. And then at the uh end of this there'll be like a a credit rating generated um from the uh Sky team. Uh [clears throat] and this will correspond to like S&P or like Moody's uh traditional finance type credit rating that you would receive you know A+ B these types of numbers and then this would translate into uh junior risk capital uh requirement for the uh prime agent.

Um so yeah that's how risk works for Sky. Um yeah so for the for the credora ratings yeah this is this is something we increasingly feel is necessary uh we see the way the market has been evolving over the years and uh I think without this sort of like independent check on uh who is doing the risk curation um you kind of have this tendency due to competition for uh people to just keep risking on um and usually the lenders are um who who often associate with the brands and will be like, "Okay, this is a big brand. I trust this brand." They may uh they they often are just operating off of brand trust and not necessarily performing

sophisticated due diligence themselves. I think this is different for institutional uh LPs, but for retail LPs, it's just like I trust the brand, you know, I trust these guys. Go ahead. And the incentives are not good there because it kind of like the brand and it's not even like a fault of their own necessarily. there's there's competitive pressures to keep risking on while not taking any sort of noticeable loss and I am I was seeing the way this was going and I was worried about the like sort of entering into like a structural uh risk for the DeFi as a whole that may kill the industry and so we've been making a really strong effort to try and get uh ratings uh attached to all the DeFi

lending products that are available including our own so uh we want to you know uh do this ourselves. Uh so Credora is a is a great um independent ratings agency within crypto but we are also like and so they've delivered the ratings for savings and uh the the ratings are uh quite good. We think we're doing a good job but it shouldn't be up to us to be telling everybody this. We want somebody else to be saying that okay yes they are doing a good job or importantly maybe you're not doing a good job and just being made aware of like where a a fault is. I think this is tremendously helpful. So, uh, an example, uh, S&P did a credit rating for, uh, USDS. Uh, I think this was like a year or so ago, and, um, it was a

decent rating, but it's like a trady rating. So, uh, you know, it's it's not like the highest it could be, but they they they also like detailed where some of the the concerns they had. And so, this gives uh Sky a direction, a way a way to go for uh to improve this rating. and not just for improving the rating sake uh but for actually building a safer product. And so this this really aligns the incentives of you know we want a good rating and so this what this means is we're going to be focused on improving the safety and the quality of the product which is a very consumer aligned uh thing and so yeah we're we're very uh uh pushing this DeFi ratings thing and in fact we've had some good

talks with the EF about this. uh they've also been increasingly uh interested in this and so uh yeah I think I think it's it's a lot in line with I think what Blockworks is doing you know you guys with the token transparency framework um and you know exposing the data like I I think like this is we all want the industry to mature and be you know institutional grade and these are kind of the steps to get there. Yeah, very good. Organically brought up. So, Yano is going to be very very happy with [laughter] um I guess uh I I want to ask a bit about the Spark liquidity intents which you guys launched a few months back. Um and you're allowing users to withdraw above the target liquidity buffer in one

transaction. Um, was this asked for by um, I don't know, certain demographics or is this something that you guys were always planning on including? I'm curious if it's like an institutional requirement so that people can come in or if it's um, just something you guys were planning along the entire time. >> Yeah, sure. Um, maybe I'll just explain it briefly. It it is a little bit technical but um the the idea is that so a lot of like uh lending markets they'll have this amount that is uh deployed into uh you know it's usually loans if it's specifically a lending market but this applies to vaults as well. They have this amount that's allocated and then they have the amount that's idle.

So the idle amount is like this 247 liquidity facility where you can atomically as a user withdraw that amount. The problem with this is if you keep a large amount in idle, it drags the yield down and especially if um there's also idle mounts in like the further deployed markets. So with a lending market, it's not quite as important, but I would say at the vault level, it's extremely important because you may have idle liquidity on uh so we're deployed in Sparkland and Morpho for example. uh you may have an idle liquidity on Sparkland, you may have idle liquidity on Morpho, but like really you don't want like all this like siloed idle liquidity all over the place. So um the idea was to basically instead of having the uh requirement having a synchronous ability to uh pull

the funds at any time, we do an asynchronous uh request where the user says um gives their intent where they're like I want to withdraw 200 million and then the uh automated allocation system can read this and then send them in the next block usually uh the their full liquidity withdrawal. This is just a much better user experience. like like professional teams uh they're they could deal with like uh so that the cap for spark savings was 10 million in this 247 liquidity but as the product grew now USDT is at like 1.2 but it peaked at like 1.5 billion just like a week ago um withdrawing [clears throat] 10 million at a time if you have a uh

600 million uh deposit is super annoying and like maybe funds professional funds can put up with that but it's just a super bad UX. So uh this this feature is just like nice for uh yeah and it is mainly institutional depositors. They can uh asynchronously signal how much they want to withdraw and then that will be fulfilled. And again our goal with this is to cover like um as for spark savings with USDC and USDT product we want uh it's meant to be like a full liquidity product. So even the largest depositors uh should be able to uh exit immediately or like within like the next block sort of time frame is what the goal here is.

>> Sam, you guys have buybacks um that you are activated back in 2025. Um and you're holding the Spark SPK um in your treasury instead of burning it. Can you um walk us through a why you guys decided to implement buybacks and b why you are holding them in the treasury as opposed to burning the tokens as well. >> Yeah. So for the first one, why were buybacks uh initiated? Um I mean to me it's like there there clearly needs especially with the way tokens have gone in the past there clearly needs to be a value acrruel attachment to the to the revenue that

comes in from the protocol. I think a lot of people have been burned by this in the past and so signaling uh to the market that like we're serious about this. Uh you know we view the tokens as like something that is deeply attached to the protocol and um you know there is a value acrruel mechanism. Uh I think was incredibly important. Um and so yeah for me that's why the buybacks were proposed in the first place. Um and again to be clear it's not us deciding anything. there is a spark governance process. SPK token holders approved this. Um now and why uh there was a buyback and hold is that um there is um just more optionality with the way that

the protocol can spend. So to be clear about the exact mechanics uh the buybacks are uh based on the amount of uh excess uh profit that is uh the protocol is making some proportion of that will be according to a formula will be sent to buybacks and then the SPK is then sent back to the protocol. So the SPK token holders are the ones that receive the SVK and then there that can be decided what to do with that afterwards could be a burn. Um but there's also you know buyback and uh distribute. uh there's lots of different mechanisms and so uh yeah that the we really want to get this in as like a the proposal was intended as like a a first step to get this going but how to use the uh SPK tokens uh still needs to be

determined. So with the framing of Sky being the central bank and Spark being the the commercial bank, um I'm curious who you see as like the primary competition um within those those verticals for both Sky and Spark. Is it other onchain money markets or increasingly do you seek to have like the SUSDS product being uh in competition with um perhaps offchain or traditional uh yield products?

>> Yeah. So this is interesting. I think this is something that actually changed as of the kelp incident. So if you ask me three weeks ago, four weeks ago or whenever it was who is the main competitor to Spark, uh it is A. So we're both offering high liquidity products on USDC and USDT. Um so the Spark savings USDC USDT and A core USDC USDT was what I viewed as the primary competition. But since the Kelp Dow incident, um the liquidity on a is dropped to pretty much nothing. Uh I think it's like 400 million total. uh which may seem like a lot but compared to uh you know Skye's balance sheet on on USDC's side it's more on the order of like four to six billion in liquidity um

not just deposits um and then on the USDT side uh this is now 750 million in liquidity on the USDT product and then total deposits is 1.2 two billion. So this is like 70% liquidity on uh this is a this is a more liquid product. So I think that's changed. So now uh this is yeah when uh when I was thinking about this question I I was uh I was like okay who are the who are the competitors now? I I think it's it's moved on. It's now the uh I would say fintexs with large balance sheets. Um as they increasingly come into DeFi this is probably going to be the next competitive layer. I think we're mostly uh collaborative at least at this stage. uh where these fintexs want to get into DeFi um but they don't

know how to do deep DeFi. So it it's more of like a collaborative relationship and I I hope it actually stays that way. I we are this is a um this is a a better place for us to be I think. Um but uh we want to we think with the stable coin market cap growing as a whole uh we expect uh this you know two three trillion by 2030 I'm a big believer in this sort of uh numbers and I think will grow not only proportionally to that but I think also getting more market share and um even versus like non-yielding uh stable coins. Uh so that'll be the next phase and then you know the final boss is is the banks. they just have orders of magnitude more liquidity. So like

liquidity is a lot of where the mode is. Uh technology is also a factor but so we're we're a big fish but crypto is a is a small pond and as we keep scaling into tradi I think there's going to be new challenges there and um yeah but I guess yeah to answer the question right now I would say fintex with uh decent sized balance sheets. I'm I'm kind of curious, Sam. I think a lot of the other lending protocols slash um stable coin protocol slash like anyone involved in some form of vault or deposit structure are going for this consumer angle where they want to launch a mobile app that perhaps people can like I don't know use as a card or um just generally like this

DeFi savings account. Um from what I've seen, you guys have mostly been web app focused. Is this um for the time being or are you going to I don't know provide the fintex with mostly the back end or do you guys have plans to kind of dive into the consumer fintech space as well? Yeah. So, we we had a a decent think on this um [clears throat] like mid last year. Um and we really like uh you know from my perspective uh I think we should like we we just want to focus at what we're good at and we do not do consumer at all and I think it's you know we want to break our team is not particularly big like we're uh you know on the order of like 25 people. So, uh, consumer is is like

it's an extremely competitive market and we're not particularly toolled up for it. So, if we were to go into that, I think we would have to like have a very serious plan of how to tack it like how do we compete with, you know, Coinbase, Robin Hood, um, Revolute, all these different fintech apps like uh these they have huge numbers of users and so uh we just uh feel like we want to double down at where on where our strength is, which is as we say the back-end collateral. So we want spark savings and we we'll have spark savings for every major stable coin. We just want to produce this uh you know high highly uh very scaled up defy native uh yield. Um and so if uh exchanges are are now kind of coming on chain with these

earn programs and um and I think Coinbase being a big example of this integrating with Morpho they've got the earn program they've got the borrow protocol. So the D5 mullet thesis, we're big believers in that. So these consumer apps are going to hook into the blockchains through these lending markets or yield products. And so we don't necessarily have to be um like even uh directly um you know attached to the users there even as just like a natural backing asset that is provides a scalable yield on one side as well as on the other side. we can use our own balance sheet to uh seed lending markets if if that's what the um fintech or exchange is looking for. So we just cover both sides of the of the market

for yield generation and really just um we just want to be the the the natural best choice there. Mini the you know the smallest margins, the uh most efficiency, the the biggest balance sheet, the best insurance like these types of best liquidity. These are the types of properties that the um fintex and exchanges are looking for. And so already we are I I [clears throat] would say now especially with the incident uh we are the market leader here and so we're just going to continue to uh compound uh this lead here. >> So Spark is um one prime agent subdow for Sky um against a handful of I think it's eight others. Um I'm curious like

to what extent these different subdows might be um competitive in the yield sources or products they're pursuing or to what extent they might be like complimentary um to have a more diversified uh service of yield for Sky. >> Yeah. So it's definitely more collaborative uh right now and I think that's the best approach. We want um Sky as a whole to first and foremost needs to uh win uh to like have a an ecosystem surrounding it. I think like the hyperco competitive uh long is like a long-term thing and it w it will eventually get there but even in even in the medium term I think what's going to happen is uh uh with the risk framework in place

uh the yield opportunities are going to be available to all the prime agents equally. Um so for example we have uh Spark prime uh which CD5 prime brokerage um not to be confused with the prime agent but uh so yeah this prime brokerage um and so this will enable uh collateralization of uh across uh exchanges qualified custodians as well as DeFi protocols to basically allow crypto hedge funds to margin across uh all of these uh platforms at once which drastically improves the capital efficiency versus DeFi lending market.

So, um you know, maybe it's changed a bit. Last my check on sort of the volatile crypto assets borrowing stable coins in DeFi lending this market was about 10 billion when I was looking at this back in February. Um and 40% of that 30 40% was was hedge funds that are borrowing to deploy across various strategies. So improving this uh capital efficiency through something like Spark Prime where you're basically covering all the yield opportunities through a competitive marketplace with the hedge funds will produce a a very a very good but it's like a it's a slightly higher risk yield but it's it is a in a riskadjusted way it is a good yield and so the prime agents will probably actually all be allocating to stuff like this where it's a rock solid venue. uh

we've done all the hard work of the you know the legal setups the um you know making sure the the tech text technical side O opsseack is good and um as we scale that up uh this is just going to be a good riskadjusted yield that the prime agents can plug into but not even just that that's just on our side the other agents are also producing new uh lending venues um for example uh like uh OEX has these uh more private credit instruments that are available and if it's a good opportunity um It's just like plugandplay for the other prime agents to get involved. And what constrains all of this is going to be the uh equity that the that the that the subdow has. So I think it'll just be equity constrained and there's going to

be more than enough opportunities going around for everybody. Um and then yeah long term I mean there'll be lots of this this is more like five 10 years like it's it's a while in the future a lot can change in the industry for then but I think that's more the time frame where this becomes like very much more like zero sum competitive but that that's a ways out >> makes sense Sam I'm curious about this governance shuffle that has happened in Sky um more specifically I'm curious about like what led to this idea of we have to change like was it just very slow or was there something annoying happening between the different subdows like what was the reason why um I know we said like we want to rebrand etc but

every time somebody talks a bit about sky they kind of mentioned that the governance process wasn't like as efficient or as happy as they like um that it should have made them. So I'm I'm curious if you have any um either examples or if you have like an explanation for why this um sentiment typically is the case. >> Yeah, I mean [clears throat] so unlike a corporate structure where there's kind of like a a leader there's a hierarchy and leadership the structure of the workforce within maker dao was was uh segmented by function. So I was in the core engineering um like just purely smart contract team um and then there was another team that was working on uh

like the front end. There's another team that did growth. There's another team that did marketing and there's no leadership on top. So every it's like choose your own adventure and then and then after that you just started having random teams that were on boarding like you know I'm going to make some videos for maker or uh you know whatever little task and I'm going to form here's my here's my team that's going to do this and we're on salary uh indefinitely. So like it seemed like maybe in a naive way this would maybe work at first everybody's just excited but then I think you could like see how this is going to break down quite quickly. uh there's no strategic direction or if there was uh it was just like in a completely informal construct

and so basically it's just like a paralysis you're not able to do anything and uh firing was basically impossible if anybody like I think there was only like a handful of like core units that were let go a lot of them didn't make sense and so um again without leadership it's like why would a delegate risk their own political neck over getting like over cutting spending somewhere. Uh it's just for many many reasons it just wasn't working. So um I mean Ruin was the one that proposed the the governance overall. So he he was away for a little while and he's like okay this is not

really working and uh and so we've moved into a more sane structure. And yeah, I must say like having been in both systems, I was inc I had to deal with governance so much and I'm like I'm a product person. I'm like strategy person like I just want to focus on this. I don't want to make proposals to and have it argue on the forum against like random internet anons uh for three weeks. It's like it's not a good use of of my time. Um and I found those was like we're I was like lobbying was like all where was spending all my time now under this new structure. Um it I it still is under like a subdow structure but it's like the people who are like uh politically aligned can can silo themselves into particular. So if

there's some like DeFi or decentralization maximalist uh team that wants to form they can form a subdow and they can uh they can compete just like everybody else and people can go into this. Um, so whereas in the previous system you basically have the at the top level you have the political factions forming and fighting over every little vote, now it's like, okay, let's just let the free market do it. Um, you have ideas for how to how to go in a good direction that people will will like, go ahead and do it. And there's an expectation that some subdows will fail.

Some will be wildly successful and overall this will just be a a tool that while maintaining decentralization can still uh adapt to market conditions. >> Yeah. I didn't know it was um the previous structure was like that. That sounds like a pain to deal with to be honest. >> Oh yeah. >> Um okay. To wrap things up, I typically ask um one of two. Um I think because I'm very excited about Spark, I'm going to do the future question, which is um what is something that you're very excited about over the next six to 12 months.

>> Yeah. Okay. So, uh what we are focused on right now, um maybe [clears throat] not the most exciting thing, but I think it's very important is growing the USDT side of the balance sheet. So um so far Sky has like largely been oriented around USDC only and I think this has been to a to a to a fault. um having a USDT market is is is quite quite valuable and so we just weren't competing there at all and so as part of an initiative in Spark we've just taken this on we launched Savings USDT at the uh end of last year and it's already grown uh in terms of deposits it's it's it's the second largest um probably will soon become the largest and it's already the most liquid product on the market um

so we're receiving a lot of interest in this um for like high highly liquid uh like at billion scale uh USDT product. So continuing to grow that and um I think that's quite good. Uh on the supply side, yeah, our focus is uh largely around the prime brokerage. We think this is a huge market opportunity. So uh we this is not going to be super big in bare market conditions because the yield opportunities are are kind of minimal right now. But during bull markets, it you you don't know what exactly it's going to be. But uh largely the LPS in the space are hedge funds that are just reallocating around whatever the points program or TG or

whatever it is. And basically we'll get coverage of this and this also includes uh the basis. So often they're um you know farming some token and then they'll they'll they'll need to be delta neutral. So they'll have like a short position that they need to have a hedge on like a centralized exchange or hyperlquid. Uh so this is going to give us very good direct market coverage. So previously we've gone through um like unsecured loans. Um it's it's not even really lending. So like we've gone through things like Athena in the past. Now we have a direct mechanism uh to go through uh where uh it's it's not like an unsecured loan to a hedge fund. It's a venue where hedge funds can borrow but these positions are overcolateralized and uh spark is setting the risk

framework. So we're setting what the collateral that is accepted. We're setting the LTDs um the composition um this can all be set in a very dynamic way. So um this will make the market efficiency much better for the borrower for Spark as the lender. There's much much more recourse because these positions can also be liquidated if they become too risky. So overall I think this is just a far superior setup than what's been available in the past. And yeah, this I think this will scale to several uh billion quite easily like at least sort of the um perpetual futures sort of uh um you know 10% of the OI is kind of like where the I think the threshold is and that was 220 billion at peak in October September of last year.

So two 10% of that this market size is about 22 billion in size quite scalable and like increasingly more so actually than than uh I would say like uh OTC or onchain uh cryptoback loans primarily against Bitcoin but both these markets are are you know tens of billions and by covering the entire market here I think this is we're we're very excited to scale this up. So this will this is still like we're in a scaling uh beta phase right now but this will be scaling a lot more towards the end of the year. >> Yeah. Also very underexplored not something that has been done um in DeFi successfully like a part brokerage.

>> Yeah. >> Okay Luke unless you have anything else you want to ask I'm going to thank Perfect Sam. Thank you for taking the time to uh join us today. Um, I guess before I let you go, for people who want to read more about Spark, um, where would you what would you want for them to first check out or first look at? >> Yeah, so you can go to spark.fi, um, and that will link out to the blog. There's where you can do some detailed reading. Uh, we have docs. Uh, you can go on our, uh, X Twitter, spark.fi, or follow me as well. I'm Hexonaut on X.

>> Perfect. Um, you can also check out the uh Block Works quarterly report um that went out >> um as of today, May 12th. Um, I think this episode will air later, but still check it out. Um, again, thank you so much for coming on, Sam, and thank you as well, Luke, for coming on. Uh, and thank you to our listeners for joining us. >> Yeah, thanks for having me.