A&T View: Existing Issues and Solutions of NFT Lending Protocols

Author: Liam, A&T Capital

I. Abstract

TL; DR:

• Assets encapsulated in the form of NFTs were one of the major breakout points in the last bull market, but the development of other NFT-Fi related projects, especially in the lending market, has lagged behind spot trading. In contrast, during DeFi Summer, the rise of DEX and Lending Protocols was almost simultaneous.

• From an absolute level, the total amount of NFT lending is not high mainly because NFTs are still long-tail assets; from a relative level, the low penetration rate of NFT lending is not due to a lack of supply and demand in the market, but rather a lack of lending protocols that can efficiently match supply and demand tailored to the characteristics of NFTs.

• Lending protocols mainly address three issues: first, efficiently matching and facilitating the supply and demand of funds; second, securely holding collateral; third, disposing of collateral as agreed in the event of borrower default. Existing Peer-to-Pool and Peer-to-Peer models have not effectively solved the first problem, as their matching efficiency is low, resulting in either high implicit funding costs or high time costs.

• The advantage of the Peer-to-Pool model lies in its low time cost, enabling "Instant Borrowing," while its disadvantage is high implicit funding costs and reliance on oracle pricing. The Peer-to-Peer model has the advantage of low implicit funding costs and does not require oracle pricing, but its disadvantage is high time costs.

• Based on the lessons learned from the Peer-to-Pool and Peer-to-Peer models, we can envision a Peer-to-Orderbook model that combines the advantages of both. For example, orders with the same collateral, loan amount limits, and terms but different interest rates can be aggregated into a single order book, allowing both parties to bid at different interest rate levels, thereby reducing implicit funding costs and time costs, achieving higher matching efficiency.

II. Main Text

Assets encapsulated in the form of NFTs were one of the major breakout points in the last bull market. The total market cap of NFTs was less than $70 million at the beginning of 2021, but by August 2021, it had skyrocketed to $42.7 billion. Even in the second half of 2022, when the market turned bearish, it remained above $21 billion. The booming spot trading of NFTs gave rise to unicorns like Opensea, valued at over $10 billion, but the development of other NFT-Fi related projects, especially in the lending market, has lagged behind spot trading. In contrast, during DeFi Summer, the rise of DEX and Lending Protocols was almost simultaneous.

Marketcap and Trading Volume of NFT (source: nftgo.io)

So, what are the reasons for the inactivity and low penetration of NFT lending?

From an absolute level, the total amount of NFT lending is not high mainly because NFTs are still long-tail assets; for individual collections, the total market cap and trading volume are not high, and there is insufficient immediate liquidity depth.

From a relative level, the low penetration rate of NFT lending is not due to a lack of supply and demand in the market, but rather a lack of lending protocols that can efficiently match supply and demand tailored to the characteristics of NFTs.

The fact that NFTs are still long-tail assets compared to FT is evident. Even the top projects, such as BAYC, have a total market cap fluctuating around 1 million ETH, which is less than $1.5 billion, and even lower than the FDV of Ape coin.

This fact is something we cannot change in the short term, but as Web3 investors, we see the potential of NFTs. In the next bull market, it is highly likely that more types of assets will be encapsulated in the form of NFTs, and the total market cap of NFTs may see a tenfold or even hundredfold increase. Therefore, from the current standpoint, we can explore lending protocols with higher matching efficiency and capital utilization efficiency, as such projects will have greater potential to explode in the next NFT bull market.

Before evaluating the existing NFT lending protocol models, it is worthwhile to clarify the essence of collateralized lending and the role of lending protocols.

The specific process of collateralized lending involves the borrower providing a bundle of assets as collateral, reaching a consensus with the lender on key parameters such as the loan amount limit, interest rate, term, and liquidation conditions and methods, and then obtaining liquidity from the lender while agreeing to repay the principal and interest. During the duration of the lending relationship, if the borrower defaults or triggers the liquidation conditions, the collateral will be liquidated according to the agreed method.

In the above process, the role of the lending protocol can be considered from three perspectives/stages:

• 1. Before the lending relationship occurs, the protocol needs to efficiently match the supply and demand of funds, i.e., facilitating borrowers and lenders who can reach a consensus on key parameters such as collateral, loan amount limit, interest rate, term, and liquidation conditions and methods, helping both parties establish a lending relationship.

• 2. During the duration of the lending relationship, the protocol needs to securely hold the collateral.

• 3. During the duration of the lending relationship, if the borrower defaults, the protocol needs to dispose of the collateral as agreed.

Having clarified the essence of collateralized lending and the core value provided by lending protocols, we can begin to evaluate the pros and cons of existing models.

1. Peer-to-Pool Model:

Advantages: Enables "Instant Borrowing," low time cost for matching.

Disadvantages: High implicit funding costs (low capital utilization, and significant interest rate spread), reliance on oracle pricing.

Peer-to-Pool Model

This model essentially imitates AAVE. Although the AAVE model has succeeded in the FT market, it is not without its drawbacks. The main drawbacks of the AAVE model are threefold: first, low capital utilization; second, significant interest rate spreads; third, reliance on oracle pricing to determine whether liquidation conditions have been triggered.

Due to the setup of the interest rate curve, the funds deposited by lenders are generally not fully borrowed out, and the actual capital utilization rate is often below 50%. This issue further leads to significant interest rate spreads, as the interest paid by borrowers needs to be shared among all lenders. This greatly increases the implicit funding costs for matching borrowers and lenders. For example, a lender may be willing to provide liquidity of 100,000 ETH to the market, but the borrower may only be willing to borrow 50,000 ETH (borrowing more would result in an unmanageable high interest rate); the borrower may be willing to pay an annual interest rate of 36%, but the lender can only receive an average of 12%.

BendDAO's current interest rate curve

When matching the supply and demand of funds, the protocol makes decisions on behalf of the lenders, who cannot determine which collateral is used for the borrowed funds or control the interest rate and term of the loan. Therefore, to control system risk and protect the interests of lenders, the Peer-to-Pool model needs to introduce external oracle pricing to ensure in real-time that the collateral can repay the borrowed funds.

However, since assessing a fair price for NFTs remains a significant challenge, the drawbacks of relying on oracle pricing are magnified in NFT lending. For instance, reliance on immature external oracles may lead to the protocol incorrectly estimating liquidity in the market, creating liquidity risks for subsequent liquidation phases.

In summary, the current Peer-to-Pool model is not efficient, with high implicit matching costs for both parties and risks associated with reliance on oracles, making it an imperfect model.

2. Peer-to-Peer Model:

Advantages: No need for oracles, lower funding costs (high capital utilization, small lending spread).

Disadvantages: Higher time costs for matching, high barriers for becoming a lender.

Peer-to-Peer Model

Essentially, the various deficiencies of the Peer-to-Pool model stem from the fact that the protocol makes decisions on behalf of the lenders when matching the supply and demand of funds. So, if the rights to decide key parameters in the contract are returned to the lenders, would these issues be resolved?

Indeed, in the Peer-to-Peer model represented by NFTfi, since the acceptance of which NFT to use as collateral, the loan amount limit, term, interest rate, and liquidation conditions and methods are all agreed upon by both the borrower and lender, the amount of funds provided by the lender directly corresponds to the amount the borrower can borrow; the interest rate paid by the borrower directly corresponds to the interest rate received by the lender. Moreover, as long as the borrower can repay the principal and interest before the due date, liquidation will not be triggered, and there will be no need to rely on oracles.

Although the Peer-to-Peer model represented by NFTfi addresses the issues of the Peer-to-Pool model, this solution also comes with sacrifices and is not a perfect solution.

The drawback of the Peer-to-Peer model is that the matching process takes longer, as reaching consensus between borrowers and lenders often requires several rounds of back-and-forth quoting; furthermore, the current lack of support for a borrower to borrow from multiple lenders (Peer-to-multiPeer) hinders smaller potential lenders from entering the market.

3. Peer-to-Orderbook Model:

Based on the lessons learned from the Peer-to-Pool and Peer-to-Peer models, we can envision a Peer-to-Orderbook model that combines the advantages of both.

In fact, the Peer-to-Peer model already utilizes standardized lending orders:

If these dispersed orders are aggregated into a public order book, it can reduce the time cost of matching while retaining the advantages of the Peer-to-Peer model. Because, before lending, both parties are looking for counterparts in a Pool (Orderbook), benefiting from the advantages of the Peer-to-Pool model; after lending, the actual lending relationship is precise and point-to-point, thus benefiting from the advantages of the Peer-to-Peer model. For example, orders with the same collateral, loan amount limits, and terms but different interest rates can be aggregated into a single order book, allowing multiple lenders to provide liquidity at different interest rate levels, enabling borrowers to withdraw the funds they are willing to accept from the order book at any time, achieving so-called "Instant Borrowing."

For example, the figure shows a possible order book. The header "BAYC-40ETH-90Days" indicates that the borrower in this order book accepts a loan with a maximum amount of 40 ETH for each BAYC provided as collateral, with a maximum term of 90 days (the same applies for lenders). The left "Borrow" column represents the unmet borrowing demand at different interest rate levels; the right "Lend" column represents the amount of funds that have not yet been borrowed at different interest rate levels.

I believe that allowing both parties to bid on a public order book will greatly enhance the efficiency of matching. Considering both the time cost of matching and implicit funding costs, the Peer-to-Orderbook model will outperform the Peer-to-Pool and Peer-to-Peer models.