Epicenter - Learn about Crypto, Blockchain, Ethereum, Bitcoin and Distributed Technologies - Yield Basis: Disrupting Defi & Bitcoin Yield

Episode Date: November 5, 2025

Brian Fabian Crain and Michael Egorov, Curve Finance founder, discuss Curve's origins: solving inefficient DAI/USDC swaps after MakerDAO borrows by creating a DeFi AMM for stablecoins and LSTs. It hit 1M TVL with a bonding curve concentrating liquidity at 1:1, more effective for pegged assets than Uniswap. Features grew to include BTC wrappers, stETH pairs, and crvUSD (a CDP stablecoin with reversible liquidations & a peg-keeper). Governance uses veCRV: Locking CRV grants voting power proportional to lock duration, a mechanism now refined in Yield Basis. Yield Basis solves impermanent loss in volatile pools (e.g., BTC/crvUSD). Users deposit BTC; the protocol borrows crvUSD, pairs it at 2x leverage (50% debt/equity), and uses LP tokens as collateral. This gives 1:1 asset tracking, while fees accrue from auto-rebalancing arbitrage. Simulations show 20%+ APY (may decline as BTC volatility drops) under a $50B TVL cap. It complements Curve by directing veCRV incentives to crvUSD pools, enhancing liquidity, fees, and DAO revenues. Key considerations: manual migrations, deterring forks, and dev support to scale. Topics Discussed 00:00 Introduction to Curve Finance and YieldBases 02:24 Understanding Curve's Unique Mechanisms 07:58 The Concept of veTokenomics 15:27 Lessons Learned from Building Curve 22:20 Exploring YieldBasis and Its Innovations 29:47 Understanding Yield Basis and Collateralization 32:25 Navigating Market Volatility and Liquidation Events 35:32 Metrics and Performance Insights of Yield Basis 38:35 Scaling Yield Basis: Future Directions 40:33 Yield Expectations and Market Dynamics 43:12 Potential Growth and Liquidity Challenges 46:18 Expanding to Other Chains and Governance Tokens 49:35 The Symbiotic Relationship with Curve 54:31 Upcoming Milestones and Future Developments Links Mentioned Michael Egorov on X Curve Finance Yield Basis Gnosis Epicenter - All Episodes SponsorsGnosis: Building decentralized infrastructure since 2015. With Gnosis Pay, the first Decentralized Payment Network. Start leveraging its power at gnosis.io This episode is hosted by Brian Fabian Crain.

Welcome to Epsenode, a show which talks about technologies, projects and people driving decentralization and blockchain revolution. I'm Brian and today I'm speaking with Michael Agorov. I thought, what if we have voting power not equal to number of tokens? But what if we lock the tokens and the voting power will be proportional to number of tokens and time left before unlock? So if you lock tokens for like the maximum of four years, you have maximum vote. But if you lock tokens for one year, your voting power is four times less. What do you think would be the kind of yield that people would be able to earn? Like 20%.

It can go down if liquidity in yield basis starts suppressing Bitcoin volatility globally. And the question is, where is this limit? Something around 50 billion TVL, I guess. So welcome to App Center, the show which talks about the technologies, projects, and people driving decentralization and the blockchain revolution. I'm Brian Crane. And today I'm speaking with Michael Agoroff, who is the co-founder and CEO of Curr Finance, which was one of the most influential, original and most successful, defy protocols. And more recently, actually, very, very recently just a few weeks ago, has launched a new protocol called Yield Basis. which is also very interesting and is tackling one of the most tricky problems

about AMMs, automated market maker, impermanent loss. So I'm really excited to talk with Michael about this today. So just before we get into that, I would like to share a few words from our sponsors this week. This episode is brought you by NOSIS, building the open Internet one block at a time. NOSIS was founded in 2015, and it's grown from one of Ethereum's earliest products. into a powerful ecosystem for open user-owned finance. NOSIS is also the team behind products that had become core to my business and that are so many others like SAFE and KOWSWAP.

At the center is NOSIS chain. It's a low-fee layer one with zero downtime in seven years and is secured by over 300,000 validators. It's the foundation for real-world financial applications like NOSIS pay and circles. All this is governed by NOSISDAO, a community-run organization where anyone with the GNO token can vote on updates, fund new projects, and even run a validator from home. So if you're building in Web 3 or you're just curious about what financial freedom can look

like, start exploring atnosis.io. Thanks so much for coming on today, Michael. It's really great to have you back on. I saw we had you on, I think, four years ago, over four and a half years ago already to talk about Kerr, so it's been quite a while. Yeah, yeah, I guess it's the time flies. So you said I'm a co-founder of Kerr, but there is no other founder. That's right.

Yeah, yeah, the only founder, right. So I guess, yeah, I guess co-founded with my laptop. Yeah, if your laptop, yeah. I should also mention a disclaimer that through my company, course one, we are an investor in yield prices as well. Right. So let's, maybe we can, I mean, I think mainly we want to speak about yield basis. But for those who are not familiar with curve, maybe just a few, can you give like a very quick

background on, you know, what curve is and what are your biggest learnings from curve? Oh, absolutely. So a curve is, I would say probably best known. the venue for my exchanging and providing stable coin liquidity to swap between stable coins and between I would say between LSTs as well I started it back in 2020 because well with one reasons one of the reasons being because I was a big user of MakerDAO borrowing dye against ETH because I didn't want to sell pressures, ethers. And, uh, but I needed some funds. So I borrowed against ETH. Um, and, you know, you could borrow sell dye on, I don't know,

somewhere and withdraw, uh, USDC as USD in Coinbase or something. But I found that actually exchange between stable coins was very, very suboptimal everywhere. on centralized exchanges on dexes like, yeah. And at the same time, I was thinking about making liquidity for what we now call LSTs, liquid staking derivatives. And I thought, well, actually the same algorithm is applicable to stable coins. So I launched Curve with making the ability to make stablecoin liquidity. and already at 1 million TVL, it outperformed everything in existence in this area.

And, yeah, it just grew since then. And of course it since then expanded to Bitcoin wrappers, staked ethers, even volatile payers, and actually what powers, yield basis is some modifications of crypto pools for making liquidity between the volatile pairs. And after that, Curve actually created its own stable coin, Curve USD, as well as landing protocol

with some very interesting mechanisms like unliquidations. And also, of course, Curve is governed by Curve Dao, for which I invented VE tokenomics. And I iterated further on that in yield basis, because I think it was

well, it was nice to somewhat update what token economic designs do. Yeah, absolutely. I think there's a lot of stuff there to go into. Just to zoom out a little bit, right? So I guess the thing in something like a Unisop or traditional AMM, right, is you basically, let's say you have like a USDC, USDT market, and then, you know, people provide liquidity on both sides. And then as you put in, let's say, USDC,

then the price of USC goes up, right? So it becomes more and more, like, it's a little bit, no, no, down. Yeah, yeah, yeah, right? You're putting it down, right? You, you're putting in USC, taking on USDT, and then price keeps going down and then very quickly

because the sort of protocol thinks that means it's losing value, right? Because it doesn't have, it's sort of an internal price predictor, right? Based on the trading volume of the exchange. So then for, especially

for stable pairs, right? It doesn't make any sense, right? Because you know they're basically worth. There is pricing. So price is not set to one. in curve, actually. It never was. But curve has what we're now called concentrated liquidity.

The very first concentrated liquidity was curve. But it's not that concentrated liquidity where you hand craft your liquidity distribution. It's a concentrated liquidity where it's distributed according to some special bonding curve, which concentrates liquidity around the process. price 1.0. So because like you know that maybe maybe having liquidity at price 0.5 is kind of useless in stablepoint world, right? So you probably want everything to be around 1.0.

And apparently the bonding curve, which is in curve, that's hence the name, is actually very efficient. So I don't think I don't think I don't think anything in existence provides significantly or bigger efficiency than that. Yeah. For stable coins, in case. Yeah. Maybe let's talk about the V tokenomics, because I think V tokenomics is something, you know, that you've seen come up in other places.

I think a lot of people ended up copying it. But can you explain what was the problem you're trying to? solve and like how do we tokenomics work so the actual the original idea was aligning the governance with long-term success of the protocol and the problem is this imagine that someone controls a protocol with tokens or shares or whatever so the more tokens or shares they have the more of the voltage power they have. What if they, for example, well, let me just bring me one example, which I think what have happened. Maybe it's not that, but maybe it is. So it's still a good example. You remember how back in the days there were several crashes with Boeing Max aircrafts, right?

And, well, the reason for these crashes was essentially low quality of production because the company was trying to save money on everything. But why did they try to save money on everything? Well, it appears that they switched a bunch of revenues to share buybacks and didn't spend. much money on improving quality. So, well, shares went up, but they didn't have much money

for the company to operate, so they had to fire some good engineers. And, of course, quality of aircrafts deteriorated. And, well, it led to a few catastrophes.

And, well, guess what? Shares started dropping, but apparently a bunch of shareholders already sold their shares by that time. And who knows, maybe they short it. But then U.S. government salvaged the company and they bought the deep. And, well, I mean, shareholders benefited, but it was not great for the company at all. Maybe it's not what happened, but that's how it looked. It seemed to me. And I thought, well, what's the fundamental reason for this to happen? Fundamental reason is that shareholders are not aligned with the long-term success of the company. So what if we do this?

What if we align token holders with the long-term success of the project? How do we do it? Okay, so and how did you use tokenomics then to align these incentives in the protocol? Right. So I thought, what if we have voting power not equal to number of tokens, but what if we lock the tokens and the voting power will be proportional to number of tokens and time left before unlock? So if you lock tokens for like the maximum of four years, you have maximum voting power. But if you lock tokens for one year, your voting power is four times less. Or if you lock tokens for, I don't know, for one week that your voting power is, I think, 0.5% of someone's voting power if they locked for four years. So that's essentially the idea.

So someone who has longer term alignment, hopefully maximum, has the biggest say in the future of the protocol. And that indeed worked. Indeed, this produced some long-term alignment, not necessarily of physical people, but also long-term alignment of protocols built on top of curve, like convex or stakedown or iron. And they are all passionate about moving curve forward. So this, in a sense, worked. But that was not the only thing. as you can imagine right because I guess the way I'd imagine

people would sort of like sort of work around it is if you if you can just then like you put that into some pool and that stakes for the maximum duration and then they issue like a liquid token right that you can like immediately that is

that is of course true but then you of course give your voting power to that protocol So that vertical decides how to use this voting power. That's one thing. Another thing, if people who do that decide to exit, then this derivative wrapper of the locked token will depot. And it will go down from parity and people will stop selling it

because there is not enough liquidity to do it. So you turn essentially, well, you turn this waiting for unlock into the market discount. Yeah, yeah. And then what was the impact or like what do you think were some of the positive effect of evaluating this tokenomic? Right. One positive effect, which I didn't necessarily think of. Well, it appears that locking the tokens removes them from circulation actually way faster than buyback and burn would do, around three times faster.

So if Curve basically collected revenues, which the protocol has, and bought CRV tokens from the market and burned, it would have removed three times less tokens from circulation than locking removed. So it appears that VE tokenomics actually practically is more efficient for decreasing their circulating supply, which is what, well, I had no idea that this will be the case when I started this. So it's a very interesting economic consequence. So, I mean, I don't know. And if you are now, as you kind of started working on a new protocol,

are there particular things you feel like you've learned or your biggest lessons from building curve where you felt like this is something I want to do differently? Yes, absolutely. So there were some details about how. VE toconomics works, which I wanted to change. Well, for once, VEanetoconomics actually wanted to resist this wrapping, and this wrapping happened anyway, but it created an obstacle of locking for multi-6, essentially. So I removed this obstacle

in VEYB. So even multi-6 can lock VEYB, that's fine. Another thing, In VE tokenomics, one of the function of VE tokens is boosting your token emissions. So let's say if you put money in the pools, take it to earn CRV tokens, then you can earn more CERV tokens if you have VECRV, which seems to be a nice function of VE tokens. That's right. but it encourages creation of natural monopolies in who hold VECRV. And that is not necessarily bad because they did a good function,

but nevertheless it's a little bit less decentralization. Well, they are decentralized within themselves, but it felt like, still felt like something around about it. So I, and also increases. boosting token production by having VE tokens is a little bit technically difficult, so I decided to remove that function from the EYB. So it's, and the YB is not doing that thing. Another thing, which is not necessarily directly VE tokenomics, which I wanted to try,

is splitting splitting revenues between staked and unstaked. So in Curve, there is no reason to not stake your liquidity. So if you provide liquidity on Curve, you earn, of course, fees of the pool, and if you stake, you also earn CRV tokens. And there is no reason to not earn CRV tokens because, like, value... If you don't do that, they just earn less, right? So why earn less?

And people who earn CRV tokens, they, sometimes they don't need SERV tokens, so they kind of sell them. And, yeah, so it's a complication for integrations. So I thought, why not do it differently? Why not make it so that there are two kinds of people. One people want to earn natural yield, right? so natural yield protocol is earning, so they have auto-compounding value just growing in their wallet.

And those who don't want natural yield, those who just want to hold the stable value of what they put in and earn protocol tokens, YB tokens. And that's what I did in yield basis. So you essentially split between two kinds of liquidity providers. and I think that is good for integrations because for integrations it's actually much easier to integrate something which doesn't earn tokens, something which just automatically compounds value. And it allows to split between two kinds of people who want to collect valuable governance tokens and not sell it, hopefully, and who want to just earn real yield and they, would have sold the governance tokens, but they just can earn real yield in the first place instead. So that's their idea.

And I... That's kind of similar to what Pendle did, no? I don't know, actually. I think the encouragement for me was not Pendle. It was how you split yield in convex. You can stake CDX CRV and can choose whether you earn CRV token. or you earn stable coins. Yeah, no, absolutely.

I think that makes perfect sense, right? And it seems like a sort of win-win from both sides, right? Because on the one hand, somebody wants to use yield basis. They want to earn yield. Then they mainly just want a higher yield, right? And if they can get higher yield, that's fantastic, right? Then other people really want to bet on the token. And then if you can sort of segregate that, I think it is very,

it is very elegant. Right. But this design leads to even more consequences which you don't think of at the start. I think you imagine two limit cases. One is when everyone wants to, wants to earn governance tokens, right? Well, and no one wants to earn real yield. Pretty easy to imagine. What happens? Then, of course, token inflation is probably that it's added max. maximum but nobody wants real yield if nobody wants real yield where does real yield go and you probably can guess it it would go to all to admin fees in this case so admin fees would be very big if everyone wants to earn governance tokens or imagine the opposite imagine that everyone is i don't know a bitcoin maxi who don't really want to deal with those

shit coins, they want only Bitcoin. What happens? Well, they all want to earn real fees, real yield, so they all don't stake. They have value of Bitcoin growing, that's fine, but nobody wants to earn your governance tokens. Nobody's staked. So is there any reason to have token inflation if that happened? Well, apparently not. So inflation of YB token is dynamic, and depends on how many people staked. And if nobody stakes, then inflation goes to zero. Hmm. So that is all dynamic and all driven by free market.

Cool. So let's, we have actually haven't yet talked about what yield basis is. So maybe explain what is the vision or how would you describe yield basis? Right. So I would describe it as essentially, how would say, a vault, right, or a set of volts, which are built on top of curve AMMs, which eliminate impermanent loss.

So you essentially provide liquidity, and it is, for the user, it's looking like essentially a yield-bearing Bitcoin, if it's for Bitcoin, of course, or for wrapped Bitcoins. Right. Because impermanent loss, is a weird thing, right?

That I think people who have been liquidity providers in Defi, you know, often have sort of, you know, painfully experienced and other people, they have no idea what it is. But basically, right, it's in an AMM, right? So let's say you have a

pairing an AMM, something like EF, Bitcoin, and then if you put, let's say you put in $100 worth of EF when you put it in $100 worth of Bitcoin. Now, let's say, the Bitcoin went up by 50%. You know, if I just had $100 worth of Eve and Eif stayed the same, then normally I would

have $250. But if I put, I have both of those in the AMM, actually sort of as the Bitcoin goes up, I'm selling some Bitcoin and I'm buying more Eif. And so I end up with less than $250. I end up with $240, 240, so some number like that. I don't know exactly. Oh, actually, actually this is, uh, uh, he's a, to explain quantitatively with Uniswap 2 AMM, although it works the same with Curve CryptoSwap

AMM, right? So imagine you have the simplest AMM you can imagine, same as Uniswap 1 actually. So you put Bitcoin, 50K worth of Bitcoin and 50K worth of US dollar in this AMM, right? So if Bitcoin goes up, it sells a little bit of Bitcoin. Bitcoin goes down, it buys a little bit of Bitcoin or, well, to the traders. So how does the price of your liquidity behave? Apparently, it is proportional to square root of a Bitcoin price plus the fees it earns on the way. So imagine that Bitcoin went up by factor of four, right? then your value in your AMM goes up by square root of four by factor of two so your 100k turns into 200k

well sounds good but what happened if you did not put funds in the AMM well 50k worth of Bitcoin goes up by factor of four turns into 200k, but you have also these 50K of USD, so you'd have had 250k, and in AMM you have 200k, which is less. So sounds like you kind of missed out when you were in the AMM. And weirdly, if Bitcoin goes down, it's kind of the same thing. In AMM, you have less than if you have your funds sitting on a shelf. Of course, it's not exactly. like that because AMMs also earn fees and you get a little bit more in the AMM but fees usually don't cover this gap because Bitcoin tends to go up faster than AMM earns fees somewhat so this was a problem for people all the time yeah absolutely and I've definitely experienced that in the past

right where I've been in like a liquidity providers and AMMs to like liquid but I have I have anecdotal evidence from people who el-ped between CRV and ETH, and they've been very happy. They el-ped that in Uniswap-2 for several years, and they've been up both in terms of CRV and in terms of ETH. But it is one of the rare examples where FEE were actually bigger than the IL. Yeah, exactly. But so the challenge, like, let's say if you take the Bitcoin example and you guys focus on Bitcoin at least initially, then, you know, people tend to want to hold on to their Bitcoin. So they're kind of reluctant to put it into any kind of AMM because they're like, well, you know,

there's a good chance that they'll end up with less Bitcoin in the end because of the impermanent loss. Like, I don't want that. Right. So how does yield basis get rid of impermanent loss? Yeah. Yeah, well, as I mentioned, the price of your liquidity, like base price is proportional to square root of Bitcoin price. So it feels like it's actually not a real loss. It's kind of almost a mathematical loss.

So you really want to get rid of that square root somehow. And apparently it is possible. and you just need to construct some position which inverts this square root function. And what position could that be? Apparently this position is leverage, but it's not a typical leverage, it's compounding leverage. So leverage where you keep loan-to-value ratio constant all the time. To have leverage of a factor of two, you need 50% LTV, And if you maintain that, then your leverage makes your position being square of the price of your collateral. And if your collateral is your AMM position, then you have square with square root, so it cancels out, and that will be proportional to Bitcoin price.

Plus the fees it makes minus any sort of expenses or losses you have on the world. about this would be not impermanent plus. So it's so you can get rid of impermanent laws this way, but then you have to have to check are you actually net positive or no, right? And it appears that you are actually net positive if you use the right AMM under the hood. Right. So basically, right, so we said before I'm trying to, let's say you have like the 50k USD and you have the 50K Bitcoin. That's kind of like the traditional way, right?

But now we want to somehow get rid of the impermanent loss. So instead of a person putting in USD and Bitcoin into, let's say, the curve AMM directly, they put just Bitcoin into into, uh, into, the yield basis. That's right. And so let's say they put in 50K worth of Bitcoin. And actually I think better to explain it like

100K worth of Bitcoin. Okay, 100K worth of Bitcoin. And the protocol borrows 100K worth of Curve USD pairs with that Bitcoin and uses this 200K position to collateralize your 100K loan.

So the net the value of this position is 100k. Yeah, so exactly. So you're putting in 100K worth of Bitcoin, the protocol borrows 100K of stables, and then it has this LP token. And the LP token basically represents the 100K work of BTC

and 100K worth of stable. That's the collateral. And with that, of course, it. Yeah, yeah. And this debt kind of cancels out on average the Curve-USD part. The tricky part is that this AMM is actually is having concentrated liquidity. So it's not XYK.

It's XYK only on average, but in different moments of time, it can deviate from ideal balance. So it can be not 50-50, but, you know, 45, 55 or, I don't know, 60-40. something and so it fluctuates like that and the loan is actually the loan is 50% and actually that well on average AMM is also doing 50%, but it's fluctuating around this ideal value and that is what is making it work because that's like inevitable property of concentrated liquidity that you have you have it going out of balance a little bit. Yeah, yeah. So I'm curious, under which scenarios,

and what's interesting, right, is you guys launched yield braces just before we had this, you know, massive liquidation event, you know, I guess it was a week ago, was it, or two weeks ago? It was October 10. Yeah, so two weeks ago, right? Which, you know, cause lots of chaos and some, at least in some centralized exchanges, some deep-partopause. How did yield bases handle the volatility and liquidations that happened there?

Right. So I actually studied this quite a bit. So first thing I checked, there are actually two amends in yield bases. One is curve AMM and another is a re-leveraged AMM, which keeps the leverage constant. And there is a spread between these two AMMs. When things are very calm, spread is actually huge. It's like, I don't know, 2% or something. But when things are volatile, spread can go down to almost zero.

But the weird thing, if arbitrage is not happening, then spread can go negative, because it's not worth the gas for arbitrage traders to trade. And that is what happened on October 10. And I thought, oh my God, this should be inherently lossy when the spreads are negative. Because like AMM gives out value. And that's happening because gas was very high and it was not worth the trade for arbitrage traders.

And at the time, TVL was still a little bit small. so bigger TVL is better. For bigger TVL, gas doesn't matter. But for smaller TVL, it does matter. And the TVL was smaller, so I thought, well, maybe TVL is too small for these gas prices. And Malay tweeted about that. But then I actually studied it more. But did the system actually fundamentally make money or lose money at that time?

Because it could be either if arbitrators is not how. happening and it relies on arbitrage. It appears that actually it net earned money. So it was that crash was actually good. Probably it would have earned MoMA if arbitrage was happening better. But yeah, it was interesting. And what's even better after this crash, someone like turned volatility on like Bitcoin woke up, started being volatile. And volatility. And volatility is what earns yield in your basis. So it was actually very, very good. Well, I mean, not to say that,

I wouldn't call the crash very, very good, but volatility turning on was a good one. And by the way, this crash was, well, allegedly called by USDA depagging on finance. But on curve, USDA did not depag at all. It was super stable. It's just, I don't know, it was due to every, I don't know, all exchanges, all, how call futures or whatever, looking at price on Binance, but it was the wrong price. Thank you.

Your microphone is off. Yeah, that's great to hear that yield pay. It's performed so well in that context. I mean, let's talk a little bit about the metrics around the launch. Like, what have you learned about the behavior? of the system and I realize it's still early phase, right? It's still early phase. It's capped and that's good.

Our cap was going up gradually, right? So first the goal was to actually verify that everything is working as expected, although there were multiple audits and everything. You can never know, right? Well, it looked fine and the cap were raised and right now the caps are at

150 million total, and that is a healthy number to make sure that gas is not an issue, and system performs as it should. A few things, I found few things on the way while this was operating, so I would be rolling out in your version of the pools with liquidity migration. I think the biggest thing is how value is split between states, and unstaked. When you, it really depends on how you measure profit. And how you measure profit, this metric is inherently volatile. And this volatility of this profit metric, it goes into the value of staked tokens, right?

So you want this volatility to be as small as possible. So I started thinking about it after talking to multiple people. and I realized that I actually can reduce this value volatility by about factor of 10. And I implemented that in one day. Auditors are still looking at it. And, well, I think we are about to put this new version out and migrate liquidity to the new version of the pools to make this value volatility smaller.

because like otherwise you have like growth happening it's actually measurable what the growth is but the real value it fluctuates around this growth and you know it's like people saying oh my god i'm i'm down my one percent why and then oh my god i'm up two percent in one day what's going on and but it's really noise around the growth and growth is not that fast you cannot grow two percent in one day. And you cannot drop the fundamental value by 2% in one day. You just have jumping around happening around this growth. So I'm just, I just figured how to reduce this growth, sort of this volatility somewhat. And that's, well, that's worth a new version before scaling up more. because the eventual scale, which can be, is much, much bigger than it is now.

So we got to figure out everything before we scale. So regarding scaling, right now, I mean, yield base is why it's this vault, where I'm as a Bitcoin user, I can put my Bitcoin into the vault, and then that provides liquidity in Curve, right, for a Curve-UST pool with T-Cube. Bitcoin versions, of course, you could imagine scaling in a few directions, right? Like one could be maybe providing the credit also on, I don't know, Uniswap or maybe other types of AMMs or... No, no, you couldn't do that, I guess.

I cannot imagine how you can do it on top of your swap, really. Okay, okay. Like, it just wouldn't work. It wouldn't be... Maybe you can, but not really. I guess. Okay. So you think yield basis is really like,

we'll just serve on curve, probably a pretty young curve. Okay. Yeah. And then initially, right, we have to focus on Bitcoin. And of course, Bitcoin is the largest asset by far.

It's also something where it's hard to earn yield in other ways. Right, right. Well, I guess the idea is that I want to get most of Bitcoin background liquidity away from centralized exchanges to on chain. Just essentially like it happened for stable coins. Yeah. Yeah.

And what is your expectation? I think you did some like modeling with historical amounts. And I guess we have a little bit of real data now but what do you think would be the kind of yield that people will be able to earn well it's a good question because after october 10 the natural apr appeared to be super high like 20% and i don't know what sustains but so far it does for some reason so 20% APR in for just from the fees right now and I kind of really promise that it stands like that but that's what observations show so far

but 20% APR I mean I presume if you remove the caps then people are going to be like 20% is incredible let me put my bitcoins in there and then you'd end up the question is what's yeah it's not the token APR so it has nothing to do with tokens it's the APR from Bitcoin volatility So if it can go down if liquidity in yield basis starts suppressing Bitcoin volatility globally. And the question is, where is this limit? How high is the liquidity to suppress Bitcoin price fluctuations?

So the 20% is not something that primarily comes from the trading fees, but it's something. It is. Well, I mean, it is from trading. but trading he's come from volatility. Right, because people basically say, oh, you have a decoupling between the curve pool and like the... Yeah, I know it's just... So it's arbitrage, yeah.

Yeah. And it matches, the numbers pretty much match the arbitrage model. So my simulations simulate arbitrage and it's almost arbitrage. It's a little bit bad more than arbitrage because you have... natural swaps like by a cow swap or khyber or even curve ui but i think around 80% of the trades at least is arbitrage and i counted on 100% arbitrage when i was simulating if you have natural volume being bigger than zero right then of course you have more revenues than just arbitrage but even just arbitrage is giving you good numbers so do you name you think um

yield basis can like how how big can it accommodate but i mean can it can be billions can it be like ten billion dollars worth of a big or you know like good question i tried to estimate when does it start affecting bitcoin price and like when will it affects its own yields by its own liquidity, by the means of making Bitcoin more stable. I think it's something around 50 billion TVL, I guess. Well, I mean, it's not unimaginable. It's comparable to Avi TVR, right? Right.

So it's not like something we've never seen. It's just, I mean, it's just making really good fees, but TVL is not something. unimaginable. Yeah. Yeah, that would be enormous. Yeah. That's the goal.

Well, of course, we have not only Bitcoin, we have also Ease, and we can make liquidity for ETH, and who knows, maybe something else, but we shouldn't really get too crazy, because I'm pretty sure it cannot work with very volatile coins, like, I don't know, like, Defy tokens or meme coins, I cannot think it will work. Because they tend to sometimes change the price too rapidly by factor of two or something in one block and stay down or I don't know, stay up. You cannot really work with that. Right, because you need to rebalance like quickly enough. Yeah, you have the, you have to have the price being more or less smooth.

Right. Even with ETH price, when I was modeling, I've noticed that, for example, if you look at ETH prices during FTCS collapse, its price collapsed fairly rapidly, like instantly to, I don't know, I don't know from what to what, but it does cause some losses in the algorithm, like minus 5% on the spot. And that's if. So imagine what happens if you use some meme coins. so please don't do your basis on NIMcoins Yeah yeah yeah yeah and then I guess even on October 10th right we saw like

I mean I don't know what it looked like for Ethereum there but it seemed to drop like super fast Yeah I will include that in simulations when I'm simulating parameters for Ethereum so before all an Ethereum out all the parameters are really carefully simulated like if you have the range of parameters different from like seriously different from what it is now than instead of having profits you will have losses right so and then in in the theorem example of course there is also the ability to stake if is it would you be able to somehow get the staking yield as well or not I think so yeah might not how

would you do that? Well, I think the simplest way would be to just use, I don't know, steak ifth instead of it, right? Then, by the way, then you, well, what then happens? Yeah, you essentially would earn a stake in yield on top of your yield basis yield, which is, I don't know, but not bad, I guess. Yeah. Yeah, yeah, yeah. not much, but honest word, I guess. Yeah, yeah. Then do you... What do you think about expanding to other chains?

You know, like, I don't know, like a Solana or other EBM chains? Yeah, absolutely. It is possible. And I think... Well, at least that's how I planned it. it's possible to do it differently, but I plan it so that each new chain should get its own governance token, right? Of course, all the investors will have their distribution honored in all the new tokens, but otherwise the token split can be a little bit different so that you align with

these chains, I don't know, marketing or whatever, whatever the rules are on that chain to make sure it's well aligned with the local ecosystem and whatever. And also you don't really have to deal with breaches. So you can actually have native, I don't know, native yield bases, which is dedicated to this chain, and you are not relying on some, I don't know, some DBNs or whatever. So it's like really leaving, there and you like you don't introduce more risks in the system essentially and so the way you

enforce this is because i guess somebody could just sort of copy it without permission or or there's some kind of license all there there is of course license but apart from that it's uh well it requires some very particular knowledge to now figure out the right parameters and You need to do that from time to time, and I don't really think that people yet figured out how to do it properly. So I would be fairly surprised if someone successfully forks it because it's not really, well, dumb forking wouldn't really work here. Yeah, right. But at the same time, at the same time, it is totally possible to, to, to, make some deals with talented people to to maintain something running and whatever like that's that's a

possibility but the course the team can also do that depending on capacity so yeah so i'm also curious about the you know the relationship here between yield basis and curve going forward i mean of course in some ways it seems very symbiotic, right? So yield basis kind of becomes a big liquidity provider for curve. Curve presumably has more liquidity, more liquid market, maybe earns more fees. Are there ways in which there are like conflicts of interest there too? Or do you see challenges in that kind of relationship? I think challenges are usually, they usually come with public not entirely

realizing how big of the return would yield basis do for curve outside the token allocation part. So there is a token allocation to curve, but the best use for it is to incentivize liquidity in curve USB pools, which is what will be happening very shortly. For yield basis, it's absolutely required that liquidity for curve USBD will be big. And that's essentially why yield basis makes. Since yield basis shows CurveUSD, yield basis makes this allocation of YB tokens to Curve. So they will be used to buy votes for Curve USD stablecoin pools,

like Curve USB USBsd, USBT, and so on. And that will create, apart from big liquidity for Curve USBUSD, It will create supply sinks for Curve USD, which means that eventually it will lead to more loans in CurveUSD and more borrow fees. But also all the trading which is happening in YB pools will create volume in Curve USD staple coin pools. And these fees generated by Curve USD pools will go to CurveDDAOys. Apart from that, there is peg-keeper mechanism in Curve-USD. So if Curve-U-SD slightly de-pegs up or down, then peg-keeper steps in and essentially buys ourselves Curve-USD,

and it earns money for the Dow on that as well, like essentially money on passively trading the DPEGs. They are micro-dpegs, so you cannot even notice them. But they do happen, and they do turn into returns for Curvedown. So that's another stream of revenue, and it's actually the whole streams of revenue. They are bigger than what I originally thought, and it looks like Curved Dow would earn,

I would say something comparable to what YB Dow earns from yield basis. So this is what it seems to me. Oh, okay. So it's actually fairly huge. It's just not direct driving use all the time. But, you know, when volatility of Bitcoin was particularly large, you probably could see that CurveUSD was the stable coin with the biggest trading volume on Curve in total,

already, with just 150 million TBR on build basis. Yeah, yeah. And so the way Curve USD works, it's similar to the MakerDA model? In a sense, you can say that. It's a CDP, so in that sense it's similar to MakerDAO. It has few unique features like reversible liquidations. So if price goes down and your collateral gets converted to Curve USB,

but if price goes up, your cholesterol gets converted back to your collateral, and you kind of didn't lose your cholesterol on the bottom. So you didn't sell the bottom. That's nice. That's one of the things why Curve USB is so cool. Another thing is stabilization mechanism of Curb U.S. is something we're called peckkeeper, which works, I think, better than anything else for keeping the pecktight.

Cool. So what are the, what's the timeline now in the upcoming milestones for yield basis? So you mentioned the protocol upgrade. Is it reduced so the volatility? What else is coming? Yeah, pools upgrade. I think it's going to happen, well, either

well maybe maybe on Monday is fine because I will have all the feedback or the will implement of all this in UI but it's actually ready in code so well maybe possible to do it on weekends but I don't know if people are up to that on weekends so anyway it's imminent for the upgrade here

but you need people to act for the migrate liquidity because in everything I build, I make non-upgradable contracts. So you cannot upgrade contracts which hold liquidity, and that's by design. But also, on Monday, I think, the vote on curve will finish the vote to spend YB allocation on incentivizing curve USD pools. And this this YB will be used for buying volts for Curve USB pools. And that's happening on Monday. That means that on Thursday, incentives for Curve USD stable coin pools will go up

and the pools will start increasing their size. And it means that after that, TVL of yield basis can be further increased. with the caps further increasing. So, well, how long after that that will happen is a good question, but technically in a few days it's ready. So that's, I think, the next step.

So we have this upgrade of the pools and then increase of the cap and then we really, really need to turn the fee switch on to connect the admin fees to VEY. so that the EYB earn those sweet admin fees. And after that, this system will be fully connected. After that, it should work for, I don't know, a couple of more weeks in this mode connected.

So, well, first of all, as I said, we would be ready to raise the caps after Curve Stable Coin Pools, get incentives. and the raise should be then to maybe half a billion. And then we need to connect admin fees to the EYB. After that is done, I think in a couple of weeks, we would be ready to raise the caps a little bit more than half a billion. I don't know, one billion or two billions or something.

We will see. And after that, we probably should watch and continue raising the caps. You cannot really raise the caps instantly because you don't want to affect Curve USB stability. You really grow two things together. Yield basis growing up and Curve USB grown up because Curve USB is used as the stable coin in yield basis. Of course I could try to make a deal with some of the established stable coins. I don't know, Taylor or something. But, you I created CurveUSD. Curve USB is making revenues for Curved Dow,

and I'm very incentivized to do it with CurveUSD. So, yeah. Cool, cool. Well, it's fantastic. I really enjoyed the conversation with you, and I think it's a super interesting, fascinating protocol, and I think this could absolutely become massive. So it's been really exciting to talk with you about it,

to sort of be a part a little bit of this evolution. And I'm really excited to see how yield pace is going to develop in the next, you know, in the next months and the next years. Absolutely. Absolutely. It's a fascinating thing. I'm very excited to see that it actually, actually works. And it actually works as predicted. So just the numbers are, they are pretty much matching the simulations.

and so that's so exciting to see. Yeah, absolutely. Cool. Thank you so much, Michael. It's a big pleasure having you in the show. Thank you.