Michael Egorov: Yield Basis - Bringing Real Yield to Bitcoin

Transcript

You're now plugged in to the Deli podcast. >> Hi everyone and welcome to another episode of the Deli podcast. I'll be guest hosting this today alongside Yan Lieberman who is the managing partner at Deli Ventures and I'm Jose uh head of Deli Labs. So today with us we have Michael Agarov, one of the DeFi OGs, co-founder of Curve, one of the original DeFi summer protocols that has invented a lot of the primitives that we use today in DeFi. And he's here today to discuss his new invention, which is Yield Basis. Yield Basis has the ambitious vision of eliminating impermanent loss on LPing, which would

basically allow you to have a natural yield on Bitcoin at scale. So effectively the go the holy grail of like a yield bearing uh a yield bearing Bitcoin. So really excited to dig into how this works today. Uh how he's taking it to market and and how big this could be. And before we start full disclosure, Delphi Ventures are investors in in in yield basis. Maybe Michael to kick us off, do you want to sort of simply summarize what what yield basis is and then we'll we'll dig in to the mechanism and more. Yeah, let's um well, as you explained very simply, yield basis is a product on on top of curve crypto pools and it uh essentially removes

impermanent loss. So, you can LP just crypto uh single-sided and turn it into essentially a yieldbearing crypto. >> Very cool. Yan, you want to go next on a followup of how how it works? >> Yeah. Yeah. So would love to maybe dive into how it works and I think part of that goes into addressing what the problem >> situation is right now and and and how you go about solving it. >> Uh yes absolutely. Um so um I think yes this problem is not necessarily um not necessarily well known to those who are new to DeFi and I would explain it this way. Imagine that you want to

you want to earn some money on um crypto exchanges, right? And let's say um let's say on exchanges between uh between WBTC and uh USDT, right? So you brought your WBTC and USDT and put into some protocol whatever that is um be it curve be it unis swap uh doesn't matter and you hope that these funds will just passively sit there and make you some fees and indeed they do but it appears that if um bitcoin goes up

let's say bitcoin goes up by a factor of four, right? For example, everything moonmed and yeah, um it's just like it like it actually happened. Um if Bitcoin went up by a factor of four, then your deposits in in the simplest AMM, let's say unis swap 2, but I guess in curve that would be the same kind of behavior in this regard. It would uh you would have two times the deposits plus the earned fees, right? And if Bitcoin went down by fac by a factor of four then you would have two times less the deposits plus the earned fees. So uh deposits um they grow and fall proportionally to

square root of bitcoin price over there. Now think about it what happens if you don't put your funds in amm. So let's say you held your bitcoins in uh your hot well in your cold wallet and you held your USD in bank account and you started with let's say with the same um 50/50 split but they are these funds are doing nothing. If Bitcoin goes up by factor of four then your um the value of your holdings if you they don't do anything would go up by factor

um essentially two and a half and if Bitcoin goes down by a factor of four in this case your you have your funds which are doing nothing would shrink to 62 and a half% of the original value. So whether Bitcoin goes up or Bitcoin goes down, it appears that if your um if your crypto and US dollars are doing nothing, you actually have more money than if you put them in AMM. So it sounds like uh sounds like uh trying to earn some fees on market

making uh actually um actually leaves you with less money than if you don't do that which is strange but that's what um what people call impermanent loss and it is some inherent property of all the AMMs out there. Uh, of course, um, well, of course, you can try to get rid of this impermanent loss in certain ways. You can try to hedge it by essentially selling this risk to someone else, but that is usually hard to do in a scalable manner.

So this whole situation really limited um really limited how how much TVL is is making crypto liquidity. When I made curve I made it actually with um as like an AMM to primarily swap between assets of the same denomination like US dollars to US dollars, ethers to stake ethers and so on. And uh because they are the same denomination, they don't really exhibit impermanent loss because they don't really diverge in price. And after all, if you have more US dollars than US dollars, you still have US dollars. Uh so it's not really a

problem. But uh with cryptos with volatile pairs it's uh it's much more tricky because uh this um like a loss of well impermanent loss which you get in in the pool is actually significant because crypto prices change significantly and they don't like to come back. And if crypto prices came back to where you started, then of course you would have a profit. Only the fees you earned because actually the prices would uh the kind of base price of your AM would go back to um back to where it was when you started. Uh but

crypto prices for some reason don't really like to go back in the long term, especially for Bitcoin. So something is uh something is really needed to address uh this problem and um and yeah this is um this is why I came up with uh yield basis. Appreciate it. And so just before jumping into yield basis, it'd be I think just a brief summary, right? The idea is when you put money when you put assets into an AMM, you're effectively buying when price goes down and selling when price goes up. And and so that's, you know, if you have a one stable one one you like a stable coin and a and a

volatile asset. As the price of the volatile asset goes up, you're selling it. And and so impermanent loss is basically your underperformance relative to a buy and hold. So these AMM when when people are putting assets into an AM it's effectively a strategy where you think price will oscillate but effect but but mean revert and and your idea is just to capitalize on fees. And so you have a lot of these uh individuals that effectively would like to get some form of yield on their assets but at the same time they don't want to then sell that asset as it goes up. Right? And so that's kind of the the issue with the AMM component is you are effectively selling when it goes up, but in a world where you don't want to and you want to maintain maximum upside, but at the same time still benefit from this asset being

used as liquidity for for trading, that's where the product that you're creating comes in, right? >> Mhm. Yes, that's um that's a very good description, I think. Um but anyway um I guess it's probably time to explain how how exactly I'm solving that in yield basis and I think uh well the idea is this um well first of all uh we work with curve crypto pools which as I said have impermanent loss similar to um to impermanent loss of unis swap 2 and that That means that if you put

let's say bitcoin denominated and dollar denominated asset then um well curve pool will have pricing proportional to square root of bitcoin price and um what really creates this impermanent loss is the square root function. So this nonlinearity uh which we don't like and we want to get rid of. So how to get rid of square root and well obviously uh if you when you want to uh get rid of square root you need to probably um square it put it to the power of two. So we need to make some

position out of um out of the pool's liquidity which uh which actually makes square of the like price proportional to square of whatever you put in that sort of position. So and does this position exist? Apparently yes it does exist and this position is uh compounding leverage. So not just the normal leverage where you have price proportional to the price essentially changing two times faster than the original price just at uh when the price is close to the current one. But um uh

it's um ch changing twice as fast as the price of original asset at all prices. And when you do that um then the price of the position will be proportional to square of the price of the asset which you put in there. So um you put uh liquidity uh crypto pool liquidity which had uh the price um proportional to square root of bitcoin price into um compounding leverage or by leveraging by factor of two. this compounding leverage would

always keep uh well would essentially borrow against the liquidity and buy more of this liquidity right so and or it could it would de leverage if the price of bitcoin goes down so it will sell a little bit of liquidity and repay some of this loan all to keep loan to value ratio uh equal to 50% at all times and when you do that then your um essentially your leverage will always be factor of two and it will be autoco compound in leverage.

So it will delever on the way down and uh leverage more on the way up essentially like that. then price of uh then price of all this position will be proportional to square of the liquidity price which is proportional to bitcoin price then so um well plus the earned fees and it sounds like it should work for unis swap too as well but actually it it's not quite doing that um as at least simulations show if you take unis swap to liquidity and apply compounds and leverage to it, then indeed impermanent loss is disappearing. But in order to keep the LTV ratio constant, you

constantly need to buy and sell some of the liquidity. And that is inherently a lossy process. And um when you do it um all the time, it appears that you're losing something. And this is a permanent loss. It's not permanent. So it appears that this loss rate appears to be about the same as earnings of unis swap 2 are. So you would probably um if you leveraged unis swap 2, you would probably lose uh money on um on fighting impermanent loss with about the same speed as um as this pool would earn money. So you would be left with your position

proportional to Bitcoin price um which earns nothing which is not necessarily interesting to be honest. So um we want like some real yield here which is coming not from incentives but from uh from the actual fees and well good news is that we are leveraging not unis swap two pools uh we are leveraging curve crypto pools and curve crypto pools they do have concentrated liquidity which is automatically managed. So in curve crypto pools you have liquidity concentrated around a price which is close to current price and when the price of the asset changes it tries to

follow this price of the asset with concentrated liquidity but only do that if uh if you have a budget to do it. So it automatically manages concentrated liquidity in such a way that the pool is making some money but also leaving some money to to manage this concentrated liquidity. And it appears that this uh configuration actually earns more money than unis swap 2 while having the same impermanent loss.

So when we remove impermanent loss from that we have loss rate pretty much similar to what we what happens when we remove impermanent loss from unis swap 2 but we earn money with a faster rate than unis swap 2. So it appears that the difference is positive and um when we actually simulate that it appears that indeed that's true. Um I think uh probably well that's kind of in a nutshell explains it but it was not not very easy to to make simulations which actually exactly match uh the reality right that it was very very important to

do >> right so basically you know the way you're you're kind of addressing the impermanent loss issue is is you're a you've designed a mechanism that's able to borrow the other side All right. So instead of providing 50/50, you just provide 100% Bitcoin >> and then >> the mechanism borrows the cash component. And so what that means is as price goes up, you're not actually selling your Bitcoin. You're just borrowing more and and and kind of and and making sure you're still 50/50. And so the >> That's an excellent way to describe it.

>> Yeah. And and I think that's huge, right? that allows people to maintain their their quantity of bitcoin that they started with and they're just exposed to the price of it and then on top of that it can be used for liquidity without actually being sold as it goes up. So I think maybe uh diving into how the mechanism works right how so the the LP positions uses collateral and maybe where does that borrowing come from and and how the rebalancing would all work would be yeah great to understand. >> Mhm. Uh yes, I think um maybe one thing to mention which is not immediately obvious um the this other side which is borrowed is only matching the dollar

side on average. It's not matching it exactly at every single time. So the pool uh composition fluctuates a little bit around this number. So the dollar part of the pool could a little bit fluctuate around this loan size and that's apparently important because apparently if you match it exactly exactly at all times not necessarily you would get uh you would get a profit. So you uh it's important that asset has some volatility profile which is I would say returning to some um to some sort of trend line essentially and if there was zero volatility and asset would just

linearly change price maybe this wouldn't actually uh work profitably but anyway the this keeping this leverage or keeping the loan size which borrows the other side is made by a special automatic market maker which swaps between LP token and the debt and that is interfaced as a normal swap. Uh so arbitrage traders can use this swap and even um even DEX aggregators can use it.

So it it just looks like a normal pool to them but it's not. It's actually uh it's actually the smart contract which keeps the um the debt side uh the debt size appropriate to um to essentially to essentially cover the expectation value of the dollar uh dollar part of the uh of the pool. And have you guys ran kind of simulations or or um you know using historical returns what what potential yield profiles here look like >> right? Uh yes that's exactly what we did and um for simulations I have some

methodology which I use since since actually late 2019 which I first applied to stable swap but then I applied it to crypto pools and actually with cryptos I uh also managed to uh to make this methodology matching in uh very well real returns without adjustable parameters. Methodology is very simple. You treat everything as arbitrage.

So you have some external price feed can be from a centralized exchange or it can be uh it can be from Dexes. It doesn't actually quite matter. Um but it has to be quite quite a high quality data and you arbitrage your model of a pool with this price feed uh in such a way that um this arbitrage is profitable. So you assume that arbitrage trader buys in one place and sells in another place and also the arbitrage trader leaves some fee to the exchange right so let's say they arbitrage with

Binance for example they leave some fee to Binance or they leave some profit for themselves and if if this round trip is unprofitable then they are not doing that and you are not getting any uh any trading for so for example if your pool has I don't know 0.3% fee and the price is fluctuating within plus -0.1% no trades in volume will obviously happen and uh that's what simulator will show that no trading volume happens if fluctuations are this small so this high fee will not uh create any trading volume and any profit Well, that assumes

that all the trades are just arbitrage and there is absolutely zero uninformed volume, which is not necessarily true, but it is actually extremely close to reality. So, it's like a pessimistic assumption, but it's actually a very real assumption. So you kind of have to make an algorithm which works even in this pessimistic assumption uh saying that everything is arbitrage. Yeah. So that's uh that simulates u first crypto pool and then and then rele uh which re leverage which keeps the

keep keeps the loan size matching the dollar part of the pool. uh also with uh with the same arbitrary strategy we run it on historic data and actually there is another trick which I figured out to do just kind of more recently it's uh well the market conditions they slowly change over the time and if we apply the same parameters to to all the simulations and to all really running pool then we wouldn't really get uh good results. So we need to adjust these parameters all the time. Uh so what I'm

doing right now is taking let's say a year of data and then based on that data calculating the optimal parameters such as fee and concentration and applying it to the following month. So the month which was not included in the data which were used to find these parameters just so that we uh it matches the real situation when we take u historic data and calculate parame optimal parameters and apply it for for the month we didn't yet see and we repeat that every month right so if we do that then um it appears that returns well it would be I

would say they in this cycle um can range from from 10 to 20% something like that on Bitcoin and I would say that average since beginning of 2023 till now is well it's somewhere between 10 and 20%. But I would say at the bottom of bare market not much trading volume happens. So you probably would have just just few% APR. But when uh when market is kind of more alive where there is uh market crashing or market booming or like well I would say I would say 80%

of the market life or maybe even more then over there you have probably more than 10% APR on average >> and from the perspective of LPS right the the the yield and the fees are are the top line and then they have borrowing cost. So is that net of borrowing cost or >> it's net? Yeah, it's of of course it's net uh the fee APR minus borrowing cost minus the losses which you have on managing managing your u kind of your borrowing exposure, >> right? And and from the the borrowing

side, you know, um naturally you want to create the best trading venue, but but also from the return perspective, you want to minimize borrowing cost. So can you kind of touch on uh the mechanism by which the I guess the the borrowing capital is sourced and and how to make that as low cost as possible? So uh what we are what we want to do is we want to use um curve USD allocation and in exchange curve DA would get um some allocation of YB tokens. uh and probably the most efficient is not to distri to just distribute this allocation of YB tokens but to uh to use

it to essentially buy votes for liquidity of Curve USD uh stable coin pools. So that would additionally increase like expand the supply syncs of curve USD and make uh make a liquidity bigger uh for the curve USD stable pairs. So it kind of stabilizes uh curve USD price which is important for uh for yield bases to work. But there is another another thing which um which indirectly I would say charges a fee. uh here which is uh the fee to swap between curve USD

and USDC or USDT because like in most trading venues um WBTC is traded against USDT or USDC and if we use curve USD then there is some fee of course uh to exchange curve USD to uh to those stable coins and this fee would um you know well would most likely be at curve so that's also what curve gets and in exchange um yield basis gets the allocation of curve USD which is um I would say free of um free of normal interest rate but yield basis would pay in different forms

>> got it thank you >> and on on the protocol side you've released uh new pool types like Curve V2 as part of as part of Curve. What was the reason for for doing this as kind of a separate protocol? And did you consider >> uh sort of just releasing releasing it inside Curve? Well, in principle it could work, but usually it's good to have um usually it's good to have a new token for bootstrapping a new protocol, but in addition, I wanted to try something new with token economics and um that's actually what I'm doing here. So I'm doing some new things in within the

framework of VE tokconomics but it it is also applicable to any um any token economics where native token is distributed for providing liquidity and that's actually an interesting one. So I probably can focus on that for a little bit. Um so of just like in curve uh liquidity providers get natural yield and also they get um they get uh protocol token. But there is a difference as a liquidity provider. You need to choose whether you want to earn um real yield from fees or you want to earn YB tokens. But you

cannot choose both. So either one or the other. Um, of course, if you chose to earn only YB tokens, then those who chose to earn real yields would get some of your of your yield, which you chose to uh not earn and earn YB tokens instead. But at the same time, this part of LPS who earn YB tokens is smaller than all the LPS. So they probably get actually a higher APR than if YB tokens went for everyone. And this makes it easier to integrate because uh for example lending protocols like Avi don't have to care about YB token

distribution. They can just uh integrate the real yield version uh of um of YB uh liquidity. And there is another interesting um interesting consequence of this. Imagine that everyone decided to earn YB tokens as an example. If everyone chose to earn YB tokens, then no one wants to earn real yield. And of course, YB inflation is maximum then. But um where would real yield go if no LPS want to get it? Well, the answer is obvious. they would all go to as admin fees to um to DAO which is represented by everyone who locked YB for VYB. So

VYB would get really many fees in this case but it could could happen a different way. It could happen that um we've got I don't know many Bitcoin maxes and Bitcoin maxes want to earn only real yield. they don't really want to touch any coins which are not Bitcoin or maybe it's a bio market and nobody really believes in tokens and anyway you can imagine that and then all 100% of people would uh try to earn real yield. Of course, there is some small admin fee going uh uh going to the system in this case, but uh the vast

majority of the fees earned by um by the system would go back to liquidity providers as deposits which increase uh which increase in uh in value expressed in terms of Bitcoin for example. But they all don't want YP token. And if they all don't want YB token, then inflation of YB token should be zero at that point. So inflation of YB token really depends on whether people choose real yield or whether they choose to earn YB token. And that I think should depend on market condition. But how exactly it will behave I think we will

see. But I think it's uh um it's going to work as some dynamic system which is choosing a dynamic equilibrium between how I I don't know traditional stocks work for example when you have dividends going to uh shareholders and how um how drone works for example when uh when uh inflation is distributed to liquidity providers and all the fees go to the D. So, um maybe neither this nor that is optimal, but maybe the optimum depends on on the market conditions and um I think that this um sort of design can uh

automatically search for this optimum. >> Yeah, we Yeah, I like the design. And I'm curious like when you as we're launching this the the kind of current meta is points and specifically, you know, having a point season where you bootstrap TVL, you don't communicate how many tokens are are going out. >> Um, and people farm it in the hopes that the airdrop is big and the token trades are high FTV. >> And that's worked pretty well for for some protocols that are similar to to to yield basis in terms of what you want to incentivize. like I guess Athena being being the the closest one. Um and in a sense like the the sort of Pendle meta on top of the points creates a dynamic

that's somewhat similar to to what Yield Basis has at the protocol level. So I'm I'm curious how you see the points meta and like why you decided to go with this design versus sticking to I guess what what other people are doing with with points. Well, honestly, I can tell for just as myself. Um, I never farm points. I I just don't do it. Maybe I missed out on many uh on many many point farms, but I just cannot uh force myself to uh farm something value of which I don't really know.

And I don't know maybe there are uh many people who are not like that but um well I think it's it's important that everyone is informed what's going on. Uh but it's interesting that you mentioned Pendle because Pendle could be fairly useful here. The yield which you are making from fees it's on average pretty high but it can change over time. It depends on what the market is doing and I think uh that's where that's the original original purpose of pendle original purpose of pendal is not to um not to convert imaginary point yield to real

yield but to actually stabilize yield which is variable. And here I think there is a lot of need to stabilize uh the yield which could be variable so it can give some peace of mind to different LPS and yeah that's probably important but it's a uh but yeah it's interesting so we are not really doing the point route and we are also not going the route of um the root of high FD it is, right? But but Pendle is still probably going to play a good role here.

>> In in what sense are you not doing the the root of high FDV? Like why why? >> Well, I mean it's it's all about the perception how how users see it. If um if you're doing very high FDVS um I mean it's um many will question like is it a is it a fair valuation is it not fair valuation and um like is it what's the risk of um of getting these these tokens right and yeah so it's totally different game maybe maybe I'm a little uh a little bit too close to the mindset

of uh DeFi summer 2020 when um I remember when uh Wifey started from a very tiny valuation as a valueless governance token and then for some reason it um it was um kind of revalued crazily over the course of next couple of weeks. So, but uh yeah, it's I think um I think that's probably that's pro probably a more interesting one. And after all, point games are um well, they worked for some time, but I think people don't like them too much anymore right?

>> Interesting. And I I'm remembering a a point with the design that you have this admin fee that increases as more people stake. What are the what's the reason for that? What are the economic risks I guess of everyone staking? >> Uh yeah. So I think uh one of the one of the reasons is this. It's well on one hand on one hand we have the dynamic emissions right and if you don't make them dynamic then uh you have the emissions kind of diverging in a way that you have zero people uh staking and getting nonzero emissions

and the same the same thing applies to talk to like real yield to APRs from fees. If you don't make admin fees dynamic, then imagine a situation when um very small number of people um are earning real yield, right? So, everyone chooses to earn tokens and almost no one chooses to earn real yield. Then if that happens, then all the liquidity is earning fees. But close to zero um close to zero LPS are trying to earn these fees. So you have kind of

finite number of fees divided by zero, right? Divided by zero people earning it. This is not a great situation. You really don't want um uh something divided by zero. You want kind of um like zero divided by zero situation over there. And this is why we we need to create some admin fee which goes to 100% when uh when nobody is earning real yield. And the other way to to describe it is like I said imagine that nobody wants your um uh real yield. So where would this real yield go? The only way

it can possibly go in this case is to um is as admin fees. So admin fees have to become 100% uh if the number of people uh who are earning it goes to zero or or they go nowhere. Right? So um and then it's just a matter of which function you put in to uh well how exactly it goes to 200% when people who want to earn it go go go to zero.

>> Appreciate. Yeah, makes sense to, you know, not overreward when there's a low stake rate and and then and then at the same way in the same vein, you know, reward well when the stake rate is high and and kind of um create a balance and and reduce concentration of yield to a few people. Um was curious how you're thinking about the the go to market, right? I think that the product itself has a a lot of natural appeal um especially at a time when when prices are volatile. So the the the the organic yield from trading fees is going to be high at the same time, you know, high high activity and and all that lends itself to a higher borrowing rate, right? So that that does kind of offset it. But I think broadly uh in in a this

is pretty a pretty ideal market for this type of product where people want to maintain their exposure but but also try and earn in any ways on top and and there haven't really been a lot of compelling Bitcoin yield products. I think most of them you're either trading off some some economic upside uh from Bitcoin whether it's you know like a it's sitting in an AMM or you're selling calls which comes with its own risk or or you're farming yield from a a new protocol that's launching and meaning that's not really a sustainable way. This is I think from what I've seen really the one of the the first sustainable ways of maintaining exposure to Bitcoin and earning an organic yield.

So yeah, curious how you're thinking of going this to market in terms of, you know, where you want to launch and and and the uh the early days of of of everything. >> Mhm. Yeah. Um I think there are really multiple uh multiple potential users here and of course there are defi DeFi native people who want to earn real yield on many things and there are plenty of plenty of people who u really trust Bitcoin here but there are um I would say there are many uh funds many who hold Bitcoin, many miners in

principle, there are institutional investors who hold Bitcoin as well. Um, but for institutionals, uh, I think it's important to think what regulations allow them to do. So I think uh more like a little uh well not really little like medium funds and uh Bitcoin miners would be um would be a good first target but we should keep an eye on uh all these uh treasury treasury companies as well. So maybe um maybe they have um some uh good needs here. But also um at

some point we definitely should go beyond Bitcoin and beyond Bitcoin. Well, the first obvious choice is Ethereum. So you could earn yield this way on top of staking it. And it's not really uh not really small. It's just Bitcoin is probably probably having the best uh volatility profile for this, but Ethereum is also not bad. So, um that's u another another side of like where where to go.

>> Yeah. And and I think you know what's particularly important is or one of the things are uh that you want to maintain this position if you're bullish on the asset because you're not selling on the way up but also you know in an environment where um you think maybe it's toppy um and and you don't necessarily you're not as bullish on it. This is also better than a strategy than an AMM because if price comes down in an AMM you're actually accumulating more of the asset, right? And so um here you're you're kind of I think the the the fixed exposure in terms of quantity of of ETH or Bitcoin works well in both directions and and is a very uh unique exposure relative to what is a relatively

commoditized exposure in crypto which is the the AMM where you're you know buying on the way down and selling on the way up and and so um I think this does have broad appeal to in that sense too. And then in terms of like the the the locations you're going to launch or what chains or what networks and and yeah, how are you thinking about the I guess the the security of Bitcoin component, >> right? Yeah, I think uh we first should start with Ethereum which is um well which is the most conservative smart contract chain you can you can imagine today and conservative is probably important given that uh uh holders of Bitcoin are conservative.

But uh also of course it's good to explore other chains and other chains could be well what could they be? They could be um of course exchange chains like uh um well like taking BNB or base for example, right? Um or Bitcoin layer 2 is an interesting option. Um, of course with a caveat that these layer 2s are mostly operating as um as essentially multi today. So maybe uh maybe that will change over time and then it will be probably very good. Uh

but I think um the main objective with Bitcoin L2s would be to check if if Bitcoin OGs will trust them more than Ethereum, right? But in general, I think Ethereum is trusted more than any other smart contract chain. Well, and of course there is uh of course there is Salana, but uh Salana is probably having less trust than Ethereum. It's still still it's good it's good to look at and uh well the difficulty is that you need to write in a totally different smart

contract language for it. So I think we first test uh test grounds in EVM world and Ethereum first is probably the most obvious choice. And on the on the new chains you're launching on, I remember you talking about there potentially being separate instances of the of the token. Is that still the still the >> Yes, that's that's still the plan. I mean, it could be done differently. Uh but um we probably will make it so that YB which is going to be launched is the token on Ethereum which is connected to the system which is running on Ethereum

and if uh the system is launched on a different chain then uh they're going going to be um a different token u well like uh which is doing the same thing as YB token but on that different chain. So you well those systems they would operate like isolated from each other and the reason is this one reason is that we don't quite trust bridges as much as uh as the chain itself and maybe maybe it's better to avoid using bridges but also it increases the simplicity of the system

if you're not using any crosschain capabilities. The system is simpler and there are simply less things to break. And I think also this uh creates very interesting um BD opportunities because like imagine that the system is launched on two competing chains, right? If it was a monolithic system then like one chain helping the project to grow there kind of helps um for the system to grow everywhere else and they don't necessarily want that. So they probably

want only well something to grow only for them and maybe to maybe to market only uh only the product which is running on their chain and they really want to make to build their own world and uh the system is very friendly to this approach. Yeah, we we haven't really seen is there an example that you can think of where that worked and was kind of >> active on there is a good ecosystem on base I would say which is >> um well more or less base only >> there is some there is an e ecosystem on BNB chain which is

>> um well not really present outside BNB chain too much like for example Um there is Pancake Swap on BNB chain and Venus for lending and although they exist on other chains they mostly on BNB chain on um on base it would be um probably drone for trading and I would say that there are some other chains who try to repeat that that there is um Barra where the whole ecosystem system is um kind of branded in uh a little bit bearish terms just for fun. Well, that's

I guess for marketing purposes. There is uh Sonic which has its I would say very own ecosystem although they do have well they do have curve deployed right but uh they are I think helping more helping more the projects which are native to native to Sonic and even arbitrum or like optimism as well for example right so I I remember when optimism was given uh uh in op incentives They preferred they preferred to give more incentives to project which are projects which are really um like native to optimism not

like projects which uh which come from other chains. So looks like everyone tries to uh tries to do that. >> Yeah, I see what you mean. Um and I'm I'm curious maybe like last question on the on on yield basis then I'd love to ask you some just higher level like DeFi questions but what um what um what what part of governance is managed by curve versus yield basis token is it yeah what's what's the interaction there >> right >> and what things are voted on by governance >> right so um a curve governance would definitely define u what are um well how big are the limits for curve USD to use

by yield basis that's probably the most important part and also curve governance will uh will probably vote on parameters of crypto pools used and yield basis defines of course parameters of parameters of the AMM which releverages it defines which assets are um used in are added in yield bases and also it defines how the CRV USD allocation is split between uh different markets

>> makes sense and then on on DeFi generally you you spoke about DeFi summer and kind of the rise of of we remember those days fondly as Well, but DeFi has changed a lot since then in that I think in DeFi summer there was really this idea that we were building a new completely decentralized financial system. Uh like everything was supposed to be on chain. uh like we people cared about governance and and having actual kind of direct control of governance and the new sort of I guess the DeFi that emerged 5 years later and and the big winners look a lot different and they made very different trade-offs to optimize for for actual for like growth and and and almost treating it as a

business rather than these these very decentralized protocols that could could you know spin up by some anonymous dev in in their basement and and grow to to billions of dollars. Like when we look at Athena, right, that's using centralized exchanges and and and hyperlquid that was very few validators for for a very long time and closed source. These these look very different than than the original DeFi. How do you think about this new world and whether you want like are you changing anything in the way you do things to adapt to it or do you are you still kind of an OG DeFi principles believer? I'm still closer to OGD5 principles, although maybe I'm not um not sitting in the

basement. Uh that's probably the difference. Other than that, I think yes, it's still uh fairly decentralized governance. Um with VE tokconomics it it looks like it's uh working very efficiently to bootstrap like when you have one token one vote without VE you often get quarums like I don't know 3 5% and on curve quorums for votes are typically I don't know 50 to 70%. Something like that. So it's a big difference and uh it just means that V tokconomics works as

like or as something which organizes um the voters together and also I don't quite I don't quite think that uh this um CDI approach that it is the right approach. I think it will still backfire. Maybe not today, but maybe it it could happen. Although I'm like still I would say I'm very bullish on Athena and Hyperliquid and I totally believe in what they are doing.

But um I would say even even when you do everything right, you have you may have high risk on um on the regulatory side at some point. Again, not today. Right now, environment is actually fairly favorable, but it's still could be the case uh could be the case down the road. >> Yeah. And and how do you think about like for example working on these digital asset treasury companies um like the the listings the the kind of business side of building a protocol these days that in in DeFi summer >> sort of the founders were were somewhat

like above this to some extent or at least the core founders really really thought their themselves as building decentralized protocols whereas the modern founder is is much more of a is much closer to like a typical startup founder I would Okay. How do you think about that as your like is it different building yield basis? Do you do you do you put a different emphasis on like the BD side and the integrations versus initially we needed curve? Yeah, I think it is probably good if treasury companies uh get launched if they if they would be able to actually um to actually put their Bitcoin to work with yield basis rather than hold um some raw bitcoin that and it does make sense to

work with them. But at the same time um at the same time I don't think uh Satoshi runs Bitcoin anyhow differently today. >> Makes sense. And and I guess what do you see as the the biggest next hurdle uh for you guys as you're building this out? Well, uh I think uh first we need to we need to see um how it works in in real life. Although I um kind of did it so the simulations match u reality perfectly but uh well uh

still uh running it in in production in real uh sizes is probably different and what's more important is that um everything should be essentially essentially pre-programmed right that changing parameters should happen from time to time as well. But anyway, so um I think first um first challenge is to actually um have it working in real life. Maybe it's not a challenge, but it's still something to um something to see and something to feel.

And next um I think um well I think next challenges would be totally uh totally BD challenges like um how to how to grow right how to increase number of um number of people or well number of amount of deposits amount of TVL in uh in yield basis. Um, so how to uh make it uh make sure that everyone um everyone who wants to potentially use it knows about it that uh they feel comfortable with it. Make sure it's uh

it's very safe. So that's uh all all of the things we need to care of. >> Yeah. And I mean we often talk about like the killer apps of crypto. I think Bitcoin at this point has definitely is definitely one of them. Um like the the sort of digital asset store value, digital gold narrative has taken hold. And then I think stable coins and dollars with a yield is is is definitely another of them. But I think Bitcoin with a yield is a product that has been like kind of almost prophesized in crypto for a while and tried in different ways, but has never been never really been achieved. What do you think the addressable market is of of a

Bitcoin with a yield? And do you think because historically like Bitcoiners have never really wanted to to take their Bitcoin like offchain like the the the OG Bitcoiners or move it from from the Bitcoin network? Do you think a sufficiently high yield that feels sustainable that's provided in this single-sided LP format could do that or or do you think it requires also some kind of like Bitcoin L2 on top? And I guess how do you see what do you see as the addressable market for this? Like what are your what are your goals your your stretch goals for for for yield basis over the next few years?

>> Right. Well, I think the addressable market is not really going to be limited by people who don't want to put their Bitcoin in on a different chain. I think it's going to be limited by um by how yield basis itself affects volatility of Bitcoin because like it it it extracts yield from price fluctuations. But obviously if you do that then price fluctuations becomes become smaller and they have less yield.

So yield will saturate. It will saturate not at very small numbers but it will and my estimates would be that um well today today this saturation point would probably be some somewhere between 50 and 100 billion u US dollar worth of bitcoin but if you think about it it's uh still like what 5% of Bitcoin market cap, right? So, um not necessarily you really need to uh take uh well to uh to get to everyone who holds their their Bitcoin

and they don't really have to put all their bitcoins in yield basis because if they start doing that the yields will start saturating and of course it's still quite a big market but The limit is I think the limit is not necessarily people but the limit is uh volatility. Well, volatility is still produced by people after all. But um but actually the limit is not um not the people who who want to hold their Bitcoin in their cold wallet in a very simple form and maybe >> well maybe it's good if they continue

holding some of Bitcoin in that form, right? >> Well, those people who want to >> Yeah. phrasing it as um the way the way you phrase it is really interesting. Basically that there will be enough of this BTC onchain that it'll actually suppress volatility enough that the fees that these people earn won't be compelling enough to make them do it and and that being the ceiling. Um, >> yeah. >> And and right, that's a huge sum of money because, you know, we've seen uh that that large BTC sell order, right? The 80,000 that got it was like 50,000 that got sold in in less than two days and we've seen the market absorb that and absorb other sums. So, it's it's a very large number. So, you know, using that as the north star, what you'd

imagine being a hurdle ahead of that is likely the size of the lending market for the stables needed to to pair that Bitcoin with. Well, >> so curious how you see that that element evolving right to to satisfy a market that large that you like as large as you're describing for this asset. >> Yeah. Well, we are not taking funds from like public lending markets. We are using premium of curve USD and then uh well the question is uh how that would scale. So uh if you you are using more curve USD in principle the pool which exchanges between um curve USD and WBTC itself acts as a supply sync for curve

USD which you borrowed. But the amount borrowed is not ideal ideally matching amount of curve USD in the pool. And you need some liquidity for well some sort of supply sync for potentially that difference. And this is why we want to have to have YB tokens which are used as a kind of grant like um IP uh kind of IP fee to curve to be used to incentivize CRVUSD uh liquidity so that these uh supply syncs for for CRVUSD are created

and I think they are uh those would be sort of equivalent ent to um to what you are talking about but you don't really need the size of um the size of lending market to uh to match the TVL. You probably need I don't know 10% of the uh TVL to be um to be absorbable somewhere from time to time. Yeah, I'm not sure I fully follow that last part that the 10% um >> Yeah. Uh well, um if you're using a traditional lending market, right, if you put 100k worth of BTC in the

protocol, you need 100k worth of USD from somewhere, right? >> Mhm. So in a traditional lending market you need 100k worth of curve USD to and someone should lend it to you. Here we would have some premium of curve USD. But like question is can you scale this up to infinity and what do you need to do to scale it up to infinity and the answer is if you try to scale it up to infinity then you like price fluctuations of bitcoin and mechanics of what's happening in your

system will um not not kind of create a price pressure up or down on curve USD on average but it will will feed into fluctuations of curve USD and you need some pools to absorb these fluctuations. These fluctuations would be probably exorting pressure I don't know 10 times smaller than uh than this amount of curve USD borrowed but still it would exist. That's what I meant.

>> Got it. Appreciate the explanation. And you said like at some point it's capped by, you know, Bitcoin volatility, which which makes sense, but like what's the I mean the the size on that must be multi- hundreds of billions of dollars. No. >> No. No, it's not multihundreds actually. >> I tried to estim I tried to estimate and I would say today it's a hundred billion. >> Interesting. >> Yeah. So it's not uh necessarily as gigantic as you uh can imagine. It's uh a little bit smaller than that. But but I don't know if uh more people trade Bitcoin. So who knows?

>> And 100 billion you calculated like to maintain a certain yield. >> Uh no no I calculated it this way. I looked at how much is sitting on centralized and decentralized exchanges in order books. and you have certain liquidity density certain liquidity sitting in plus minus 2% uh prices for example right so I matched matched how much would would the TVL in yield basis be to approximately match this uh amount of Bitcoin liquidity sitting in plus minus 2% well it's very rough but that's

uh >> essentially I'm hypothesizing here that yields will start saturating when liquidity and yield basis will start matching the sum of all liquidity in the world in within small range of prices. >> Okay, makes makes sense. >> But it's it's not well it's not trillions. >> Yeah, I mean there's a long way to go. Yeah. Anyway, but um but actually um if we try it with cryptos, why not try it with gold, right?

>> Yeah. So, you would have some onchain. Yeah, you could have onchain versions of of anything that that's doing this. >> Yeah. Well, I mean, I probably wouldn't dare to try it with stocks because they like to change prices instantly um sometimes and it's scary. But why not why not gold? That's everyone wants wants to have a brick of gold which grows in size every day. Yeah, makes sense. Um, and last last few questions now just just on DeFi generally like what what's the coolest innovation that you've seen maybe in the last cycle and and what's one or or a

few projects that you're that you're excited about in in in DeFi? It could also be in crypto generally if you're if if you're looking at crypto more broadly. >> Yeah. Yeah. I actually more no more like DeFi and more look into DeFi and I think it's pretty fascinating how um how Pendle started being used. I wouldn't have thought that that it gets adopted as much as it is. It's it's basically an ICO platform right at this point like to some extent >> in a sense. Yes.

Also I really liked how um how Athena grew up. So this is uh fairly fascinating and uh I mean I would I would have thought that this idea is a little dangerous in a way that sometimes the funding uh turns into not into your favor but on average it is on average it is right and they manage uh they manage the like where the funds are very efficiently and it works. So it's um well it's not really not really

decentralized but it's the concept is working and that's uh fascinating and um I pretty much also liked how Hyperlid uh grew up in in principle there is nothing super new over there like there they were per deckes before and plenty of them. But I think it's pretty pretty interesting how they managed to bootstrap um community like having some airrop which was of some unheard of size that was a good marketing twist without uh without

having uh VCs or well that's what they say but anyway it sounded uh fairly I would say fairly brave and it worked and uh yeah so it's more I would say not so much um not so much in technology although technology is uh very good but it's not like a super unique technology but the how they do business is uh very very very good. So that was the these ones were um pretty interesting and felt somewhat new in in different aspects.

>> Yeah, agree with all those. It's been it's been impressive to see the the the growth especially of Athena which I guess is the nicest comparable to to uh yield bases in terms of growth strategies right? >> Probably a lot of inspiration you guys can can take there getting integrated everywhere. But uh yeah, all those you mentioned are have been impressive comeback stories this cycle. Uh Michael, really nice to have you on to to chat through yield basis. Uh I think this is the yield basis is still pretty under the radar despite how exciting the the the potential mark like market size here is and and obviously your your credibility. So hopefully this podcast can can help people learn more about how

it works and and get as excited about it as we are. Um, and yeah, really appreciate you joining. I know you're traveling, so thanks for making the time. >> It's great to have you on. Thanks for joining. >> Yeah, thank you. Very, very nice chat. >> Cheers, guys. [Music]