Development History

The concept of DeFi emerged in 2018, with decentralized exchanges (DEXs) and lending protocols evolving through continuous iterations. This expanded the possibilities of DeFi, drawing inspiration from traditional financial systems, which led to the creation of derivatives, contracts, and other applications. DeFi products offered a platform for capital flow during this period, but the overall liquidity and user base remained limited.

However, 2020 is widely regarded as the true beginning of DeFi. The ecosystem demonstrated tremendous growth potential, largely driven by the explosion in the value of the lending market. One key factor behind this was the liquidity mining craze sparked by the lending protocol Compound. Another driving force was the composability between DeFi protocols, which facilitated the vertical development of the lending sector, spurring diverse applications and fostering a strong demand for lending among participants.

After 2021, the rise of new blockchain networks further emphasized lending as a foundational component of the DeFi financial system. Lending remained a core building block for on-chain ecosystems. Over time, lending protocols matured, with Ethereum leading the charge by providing a well-cultivated environment for iterative upgrades and innovation. The fastest and safest way to rapidly develop a blockchain ecosystem was often to replicate established protocols from Ethereum directly. These protocols attracted users with high liquidity mining rewards. As a result, there was little differentiation in lending products across different chains, with competition focusing primarily on liquidity acquisition.

In recent years, as congestion and rising fees on the Ethereum network became prominent issues, many lending protocols sought cross-chain and multi-chain solutions to enhance efficiency and reduce costs. On the one hand, protocols expanded by deploying on newer blockchains and adopting cross-chain bridge tools to enable interoperability of assets and data. On the other hand, as traditional finance and real-world economic sectors began to take an interest in DeFi, lending protocols explored the possibility of bringing real-world assets (RWAs) on-chain to expand the scale and impact of the lending market. Some protocols have already begun experimenting with tokenizing assets like real estate, vehicles, and bills, offering corresponding lending services.

At its core, all financial systems revolve around liquidity services. As the blockchain ecosystem evolves, the lending market, a key part of on-chain infrastructure, will continue to innovate and introduce new models. The future of lending is set to become increasingly diverse and dynamic.

Lending Principles

For users, the traditional lending process involves four main operations: deposit, borrow, withdraw, and repay. The participants in this process are depositors, borrowers, and the bank, with the bank acting as a third party to facilitate transactions between the lender and borrower. During this process, the bank fulfills three key roles:

1. Record keeper: Tracking the times when users deposit and repay funds.
2. Liquidity intermediary: Matching borrowers and lenders to fulfill borrowing needs.
3. Debt collector: Pursuing repayment from borrowers in cases of default.

In the DeFi lending market, users still perform the same four operations—deposit, borrow, withdraw, and repay—but interaction with a bank is replaced by interaction with smart contracts, eliminating the need for a third-party intermediary. For DeFi protocols, the core challenges that need to be addressed can be broken down into three areas:

1. Record keeping: Since blockchain fundamentally serves as a distributed ledger, smart contracts can use the Ethereum block height as a timestamp to update relevant user data, thus solving the record-keeping function.
2. Liquidity matching: The liquidity pool model addresses the issue of liquidity by enabling quick matches between lending and borrowing needs. Depositors place their funds into a shared liquidity pool, while borrowers draw directly from it. Since there are no fixed terms, borrowers can repay loans anytime, allowing flexible borrowing and repayment.
3. Debt settlement: Ensuring the repayment of debts is crucial for the survival of lending products. Without a centralized debt collector, DeFi protocols must rely on alternative mechanisms. First, over-collateralization allows borrowers to secure their loans with collateral, such as crypto assets, that can be liquidated in case of default. If a borrower defaults, the collateral can be sold to repay the loan. Second, anyone can step in to repay the debt on behalf of the borrower by purchasing the collateral at a discounted price. This mechanism creates arbitrage opportunities, incentivizing others to clear the borrower’s debt.

By addressing these three components—record keeping, liquidity matching, and debt settlement—DeFi lending protocols can function effectively without traditional financial intermediaries.

Variable Interest Rate Lending

Lending markets can be divided into two categories based on the interest rate structure: variable interest rate markets and fixed interest rate markets. The variable interest rate lending sector has matured significantly. The “big three” in DeFi lending—Maker, Compound, and Aave—serve as the primary liquidity hubs, steadily expanding their presence across various ecosystems and continually broadening their business horizons.

Focusing on Aave as an example, it originally launched as ETHLend in 2017 through an ICO. Initially, ETHLend provided a peer-to-peer lending model. However, due to inefficiencies in matching borrowers and lenders, the project shifted in 2019 to a collateralized lending model, offering products such as flash loans and credit delegation. Flash loans gained significant attention after the 2020 bZx hack, bringing Aave into the spotlight. Toward the end of 2020, Aave introduced changes to its economic model by adding a safety module and launching liquidity mining. The protocol saw explosive growth in 2021, driven primarily by its deployment on Polygon, Avalanche, and Optimism. Over five years, Aave has cemented as one of the “big three” in DeFi lending.

In variable interest rate protocols, lending can be further classified based on the collateralization level into over-collateralized and non-over-collateralized lending.

1. Over-Collateralized Lending
   Currently, over-collateralized lending dominates the DeFi lending market. Based on the type of collateral, this form of lending can be further divided into stablecoin collateralized lending and crypto asset collateralized lending. MakerDAO, as a pioneer, paved the way for the over-collateralized stablecoin sector. Protocols like Liquity have rapidly grown since their launch, with stablecoins like DAI and LUSD as prominent examples of successful over-collateralized stablecoins. As the DeFi industry matures, many established DeFi projects have ventured into the stablecoin market, launching competitive over-collateralized products such as GHO and crvUSD.

2. Non-Over-Collateralized Lending
   This can be further divided into leveraged trading and unsecured lending. Unsecured lending protocols like TruFi allow loans without collateral, while Gearbox offers leveraged trading. Gearbox connects depositors and borrowers through credit accounts and interacts with multiple DeFi protocols, offering up to 10x leverage.

Fixed Interest Rate Lending

Compared to variable interest rate lending, the development of the fixed interest rate sector started later, with most projects emerging in the second half of 2020. Currently, the market has yet to reach a significant scale. Many fixed interest rate protocols primarily rely on leading variable rate lending projects for user acquisition, attracting users who are not necessarily long-term participants. Additionally, these protocols often maintain liquidity through mining incentives, which leads to user withdrawals as soon as rewards diminish.

Fixed interest rate protocols in the market can be categorized into the following types:

1. Zero-Coupon Bonds:
   These protocols mint and issue tradeable zero-coupon bond tokens at a discount with a fixed maturity date. Users can repay their debt at a set interest rate upon maturity to redeem their collateral. This type of protocol forms the foundation of fixed-interest rate lending. Examples include Yield and Notional.

2. Yield Tokenization:
   Similar to coupon stripping in traditional bonds, this approach separates a user’s deposit into two parts: the principal and the yield. Each part is tokenized into assets with fixed terms and fixed returns. Examples of protocols using this model include Pendle and Element.

3. Tranche-Based Risk Segmentation:
   This model divides yields into senior and junior tranches. Senior tranche participants receive lower returns but take on less risk, while junior tranche users can earn higher yields but often must provide guarantees for the senior tranche’s capital. BarnBridge is an example of a protocol utilizing this structure.

4. Interest Rate Swaps:
   In this model, users can swap floating interest rates with fixed interest rates, using the same principle. This allows participants to manage different types of rate exposure. Protocols like Horizen facilitate these interest rate swaps.

Long-Tail Asset Lending

Leading lending protocols typically use highly liquid mainstream assets as collateral. However, long-tail assets with lower liquidity, smaller trading volumes, and relatively low market capitalizations are often overlooked. As a result, new lending protocols are actively exploring use cases for these underutilized assets.

Currently, no dominant player has emerged in the long-tail asset lending space. Euler Finance was a notable protocol that showed promise with its permissionless listing of assets. It managed risk by categorizing assets into different tiers and adjusting their borrowing capacity to mitigate liquidity risks. At its peak, Euler Finance had a total value locked (TVL) of $300 million, but a vulnerability in its business logic led to a hack, causing the protocol to halt operations.

Silo Finance is currently gaining momentum in this space. It uses a novel approach by organizing its lending pools into isolated “silos,” where each pool is independent from the others. This risk isolation is achieved by using bridge assets as intermediaries, so if an issue arises in one silo, it does not affect the others.

Due to the high price volatility of long-tail assets, most protocols limit their borrowing capacity as a risk management measure. Borrowing long-tail assets is more attractive in bearish markets, as users who short these assets are betting on price declines. However, shorting long-tail assets in bull markets exposes users to the risk of significant price increases, making it less practical.

One reason users may be willing to borrow long-tail assets is that these assets could have more lucrative use cases, such as liquidity provision (LP) or yield farming, where returns might outweigh the asset’s price volatility. In such cases, the strategies offered by the protocol play a crucial role in making long-tail asset lending appealing.

In summary, despite the growing interest in long-tail asset lending, protocols in this space face significant challenges.

NFT Lending

The market for Non-Fungible Tokens (NFTs) experienced explosive growth in 2021, with a surge in both trading volume and user participation. This heightened activity, coupled with the need for liquidity for NFT assets, led to the emergence of NFT collateralized lending markets.

The core purpose of NFT collateralized lending is to improve the liquidity of NFT assets. The main framework involves assigning a valuation to an NFT and offering loans based on a certain percentage of that valuation. This provides liquidity to NFT holders, who can use their NFTs as collateral. Through NFT lending, NFT holders pay interest on loans to access new investment capital, lenders earn interest on their loans, and platforms facilitate these transactions for a fee.

The NFT lending market is still in its early stages, with most projects remaining small in scale and in the initial phases of platform development and business expansion. Two main models dominate the space: Peer-to-Peer (P2P) and Peer-to-Pool (P2Pool) lending.

• Peer-to-Peer (P2P) Lending:
  The leading P2P project is NFTfi, which matches borrowers and lenders individually. The advantage of this model is that it can cater to highly personalized needs, and lenders can exercise precise risk control on a case-by-case basis. However, its downside is low efficiency, as finding the right lender may take time, slowing the platform’s growth.

• Peer-to-Pool (P2Pool) Lending:
  The leading P2Pool project is BendDAO, which provides immediate liquidity to borrowers by pooling capital from lenders. The advantage of this model is higher efficiency, allowing borrowers to access funds more quickly and the platform to scale liquidity faster. However, the risk lies in potential bad debt accumulation if the collateralized NFTs depreciate significantly in value, which could trigger liquidity crises or pool-wide liquidations.

Most NFT lending projects currently favor the P2Pool model due to its efficiency and ability to meet user demand. Additionally, P2Pool models are more conducive to partnerships with DeFi protocols, further improving liquidity and fueling competition in the sector.

RWA Lending

With the prolonged bear market in the cryptocurrency space, combined with the Federal Reserve’s interest rate hike policies, the TVL (Total Value Locked) in the DeFi sector has drastically shrunk. The outflow of capital from the market has led to a contraction in market capitalization, leaving idle assets in search of new narratives to drive growth. Consequently, attention has shifted toward the vast traditional financial market—Real World Assets (RWA), which represents a multi-trillion-dollar asset class. This move aims to bring off-chain assets on-chain, marking the beginning of the RWA narrative. Blue-chip DeFi protocols like MakerDAO have already introduced RWA-related services, and the founder of Compound has launched a second venture focusing on RWA. Due to these leading protocols’ established user bases and robust liquidity, they hold a first-mover advantage, intensifying competition within the RWA space.

Maple Finance connects institutional borrowers with lender pools, allowing borrowers to obtain unsecured loans by providing credit reputation and business information. Lenders can earn interest by providing liquidity to different pools, each with its own risk and return profile. The protocol has partnered with Centrifuge to introduce RWA-backed loans onto its platform.

Core Aspects of Lending Development

1. Matching Mechanism Based on different matching methods, lending models can be divided into peer-to-peer (P2P) and peer-to-pool lending. In P2P lending, borrowers and lenders engage directly, requiring time and effort to find a suitable counterparty for the transaction. In peer-to-pool lending, both borrowers and lenders interact with a shared liquidity pool, following its rules to facilitate lending, thus improving efficiency. The peer-to-pool model is currently the mainstream approach, as seen in protocols like AAVE and Compound. This model effectively spreads risk, as all liquidity providers share the burden of bad debt. However, because a significant portion of the funds often remains idle, the utilization rate of capital for liquidity providers is relatively low.
2. Interest Rate Model The interest rate model dynamically adjusts deposit and borrowing rates based on the utilization rate of capital, which is the ratio of borrowed funds to the total pool size. When the utilization rate reaches 100%, it means all available funds have been lent out. In such cases, if some depositors wish to withdraw funds, a bank run could occur. Thus, the key to designing a proper interest rate model is to maintain a reasonable borrowing rate to ensure the liquidity of the pool.

Take Compound as an example. The borrowing rate would gradually increase as utilization rose, but the changes were relatively mild. As a result, when utilization was high, the model couldn’t effectively stimulate the market, potentially leading to liquidity exhaustion in the pool. In later iterations, Compound introduced the concept of an optimal utilization rate. Once utilization exceeds this optimal level, the borrowing rate rises sharply to suppress borrowing demand and safeguard the pool. Similarly, the deposit rate would increase to incentivize depositors to add more liquidity, helping restore the balance between supply and demand in the pool.
This interest rate model serves as the fundamental structure for most lending protocols in the current market.

1. Liquidation Mechanism When a borrower’s position surpasses the safety threshold, part of their collateral is sold at a price below the market rate to repay the loan’s principal and interest, ensuring the normal operation of the liquidity pool. The design of the liquidation mechanism directly impacts whether and at what price a user’s position will be liquidated. Several factors must be considered:
   • Collateral Type: Lending protocols typically choose highly liquid mainstream assets as collateral, as the broader application and higher value of liquid assets reduce the risk of “bad money driving out good.”
   • Collateralization Ratio: Protocols set relatively low collateralization ratios to ensure that the collateral can cover the debt upon liquidation. Generally, the higher the liquidity of an asset, the higher the collateralization ratio, and conversely, less liquid assets tend to have lower collateralization ratios, for the same reasons as collateral selection.
   • Price Feed Mechanism: Lending protocols rely on external price feeds for liquidation. However, price feeds from a single, shallow source (oracle) are vulnerable to manipulation, potentially triggering abnormal liquidations. To mitigate this, lending protocols need stable and diversified price sources.