Why Aave V4 Is Built Like a Central Bank for DeFi | Stani Kulechov

In the past 12 months, I don't think I have met a bank that doesn't have a digital asset team and are thinking already all the way to DeFi. And banks and asset managers are actually thinking of building products and what to launch.

Really good. Really nice venue and great to be here.

Stable coins, regulation, liquidity, you know, all the things I love.

Yeah, I think the context for every iteration of the Aave protocol has always been the existing context of where the state of DeFi is at that point. So we built Aave as a lending protocol. The way it works, if there are people new to the DeFi space, is that you deposit cryptographic assets, stablecoins, or tokenized assets, and you can earn interest, you can use them as collateral, and then you can borrow as well. Typically users go to Aave to earn yield on stablecoins. That is the biggest use case. I would say 70% of the users using Aave are using it for yield, and the rest are the borrowers that create the interest rate that then goes to the depositors. Obviously DeFi has grown significantly over the years. When we started back in 2017, the space was really small, even we had less users than this room today, and it was very experimental, but the vision was big. Aave had multiple generations. In 2020 we launched the first version, and a year later we followed up with V2, and V3 was our state-of-the-art protocol, focused mainly on capital efficiency, building a really good onchain lending infrastructure for digital assets. Aave V4 was a creation of the vision that in the future we might have a very diverse set of lending activities on chain directly, involving native crypto assets like Bitcoin and Ethereum, but also more tokenized assets, or lending without tokenization, lending to the real world effectively. So we wanted to build something extremely modular with good risk controls, but that allows extending the protocol to any use cases in the future. That means if we have a new use case, we don't need a new version of Aave because we can extend the protocol with this hub and spoke architecture. Spokes are effectively lending markets, and hubs are where the liquidity is stored. The idea is that it is a lending protocol that can be extended to any use cases for institutions, fintech, and more traditional use cases beyond what we are doing today in DeFi.

Yeah, I mean our existing V3 infrastructure has seen multiple market cycles, market or credit cycles. Aave has been really resilient during turbulent times, like if we look into October 10th of last year where we had a lot of unwinds in the crypto space, also during FTX and a lot of centralized lenders going underwater, and the early DeFi days as well. So over the past years, a lot of risk management experience has accumulated for Aave quite extensively, and for DeFi generally. When it comes to development, there are very few issues in DeFi protocol smart contracts; generally they are third-party dependencies related to more traditional security that might have impact across the DeFi space, like we saw recently. In terms of Aave V3, it allows creating any kinds of markets, has risk management features, allows isolation or collateralization where collaterals aren't borrowed. So it really depends on where the demand is and what kind of markets to create for different audiences. V4 in return is building to the existing DeFi audience, providing capital efficiency to native DeFi use cases, but at the same time recognizing that the assets we are going to onboard into decentralized finance are more issuer-based assets. That means not just stablecoins but tokenized assets, and tokenized assets have different risk characteristics compared to native assets like Bitcoin and Ethereum. For example, you might not have the liquidity on chain, but you need to sort out unwinding and liquidations. With V4, we introduced features like risk premiums where you can price the risk directly into the collateral without affecting all the borrowers borrowing that stablecoin. It also allows creating very flexible market structures. We have now deployed a prime market which is low risk, core is risk adjusted, and plus that is risk return. Within these markets, you can decide how to share liquidity with each market. So you have a hub and spokes structure where each spoke has a composition of assets, and the hubs give credit lines to these spokes up to certain points. You can think of V4 as a model where you have a central bank that allocates capital and a commercial bank that has a loan book. It's a similar relationship, and that creates novel risk mitigation features but also allows sharing liquidity to avoid the liquidity bootstrapping issues we see on an ongoing basis in DeFi with new use cases.

No, it's extremely hard to build DeFi protocols and really resilient infrastructure. I think the talent in general is very scarce in our space. That's why we have so much consolidation and also the reason why it's better to use something that is already built, battle tested, resilient, and has a strong community looking after this infrastructure. I think the big difference between what we saw back in the days with FTX and with DeFi is that traditional models and centralized lenders act like black boxes where you don't have much visibility and the trust factor is quite high. There is tooling that improves that, like proof of reserves, that can be applied and makes it easier to understand the collateralization in a qualified custodian. That world has improved significantly over the years. But in DeFi, all these code bases and protocols are reviewed by countless people, users, and industry members. I think that is the key difference: when you have a completely auditable and public system, anyone can inspect the code and do different kinds of risk analysis based on that. That is the key to building resilient software. In DeFi, we are learning on an ongoing basis, and we are far in a better place than we were last year, and far better than years ago. I think that is a big difference. I do think there is going to be a trend where a lot of the risk moves away from the protocols and onchain into offchain, into the actual issued assets, and risk management will happen more on the asset issuer level, for example with tokenized assets. So traditional risk management tools apply, but we need to figure out how to communicate that transparently back into the onchain and be able to adjust parameters or safety functions with those assets in mind.

Yeah, I think Aave serves a couple of purposes. It is a treasury management tool in the sense that you want your treasury to yield interest on an ongoing basis. Where onchain lending comes into place is that you have the ability to go above the traditional risk curve with collaterals that might be highly liquid and can be liquidated fully to the last debt position. But at the same time, it opens up a completely new world of building financial primitives or lending markets where you can go higher on the risk curve but with transparency and auditability that is unprecedented. That becomes really interesting for treasury managers. Also on the retail or commercial side, Aave has been recently integrated into a platform for creators to sell digital goods online, and they have huge traction. One of the products they launched was Treasury, where creators can use their earnings and put them into a treasury product powered by Aave, which is a vault manager, and Plasma network where this Aave market is deployed. This was the first biggest fintech integration of DeFi at that scale, and it showcases how easy it is to plug into DeFi. The benefits are that the more fintechs, asset managers, and treasuries use and plug into DeFi, the network effects scale. That means more liquidity to withdraw, cheaper cost of capital for borrowers, and compounding effects that you don't find in traditional finance.