Written by: 0xGeeGee
Compiled by: Sissi
In traditional finance, lending markets offer short-term borrowing opportunities, typically for liquid and low-risk assets, aiming to provide safe and as high returns as possible. In decentralized finance (DeFi), this concept has evolved to refer primarily to the ability to borrow various digital assets in a decentralized and permissionless environment, with no specific time limits. These platforms allow users to deposit cryptocurrencies into the protocol, while borrowers provide sufficient collateral in exchange for paying interest to depositors.
Lending markets use dynamic interest rate models, which automatically adjust lending rates based on the liquidity utilization of a particular market or pool. These models ensure efficient allocation of capital while incentivizing borrowers to repay borrowed assets during times of liquidity tightness. A key feature of this interest rate model is the "kink point," where interest rates begin to rise significantly once utilization reaches a certain critical threshold to control leverage in the system: as utilization rises, interest rates may gradually increase, but once the kink point is exceeded, rates can soar rapidly, significantly increasing borrowing costs.
It is important to note that lending markets are different from unsecured loans: lending markets require borrowers to provide collateral to secure the loan, ensuring that repayment can be made at any time during the loan term; while unsecured loans (typically referring to traditional loans) allow customers to borrow without providing collateral (or only partial collateral or other guarantees), relying on credit scores and legal means to secure repayment.
Lending markets: Fundamental "LEGO" in the DeFi ecosystem
The importance of lending markets in DeFi is primarily reflected in their ability to help users earn yields and unlock liquidity from idle assets without selling their holdings. This function plays a crucial role in capital efficiency in DeFi. The ability to borrow against specific tokens is one of the most sought-after features in the industry and is often a key criterion for determining whether a crypto asset is considered a "blue-chip" asset.
This feature allows users to obtain leverage at a low cost, helping high-net-worth individuals (HNWIs) incorporate assets into tax planning, while also enabling asset-rich but illiquid teams to support operating funds by using their treasury and held assets as collateral to borrow and earn interest on the collateral in the process (for example, Curve and Maker in the past few years are two typical cases).
Additionally, lending markets also act as pillars for other DeFi tools, such as collateralized debt positions (CDPs), yield farming strategies (supporting many approximate "delta-neutral" strategies), and on-chain margin trading. Therefore, lending markets are one of the most important building blocks of DeFi, often referred to as "money legos."
To give everyone a clearer understanding of the scale of these lending markets, the total value locked (TVL) in crypto lending protocols has now exceeded $32.6 billion, as shown in the following chart.
Source: Defillama
Design decisions in the crypto lending market: shared liquidity pools vs. isolated liquidity pools
While cryptocurrency lending markets serve the same fundamental purposes, there are significant differences in the design of liquidity structures. The biggest difference lies between markets that use a single shared liquidity pool (such as @Aave) and those that adopt isolated liquidity pools (such as Compound v3). Each model has its own trade-offs, influenced by factors such as liquidity depth, asset flexibility, and risk management.
Isolated liquidity pools: Flexibility and risk isolation
In the isolated liquidity pool model, each market or asset operates within its independent liquidity pool. This approach has been adopted by protocols like Compound v3, and an even more extreme example is platforms like Rari Capital (before its collapse).
The main advantage of isolated liquidity pools is their flexibility in building sub-markets. This flexibility allows protocols to create markets tailored to specific asset classes or user needs. For example, isolated liquidity pools can be specifically designed to support a particular group of assets, such as meme tokens, or to allow only certain types of tokens to exist due to unique risk characteristics or demand.
This customization is one of the biggest advantages of isolated liquidity systems, as it allows projects to tailor solutions for specific communities or niches that may not fit into the broader framework of shared liquidity pools. This advantage is particularly evident with the rise of LRTs (Liquidity Reward Tools) and tokenized base trading (such as USDe): many users want to capitalize on high yields but are unwilling to take on the risks associated with new assets.
In addition to flexibility, isolated liquidity pools also provide better risk isolation. By isolating each market, the risk of any specific asset is contained within its corresponding liquidity pool. This means that if the price of a particular token drops sharply or becomes excessively volatile, its potential impact is limited to that market and does not spill over to other parts of the protocol.
However, these benefits also come with costs, as isolated liquidity is a double-edged sword, which also means fragmentation of liquidity pools.
For isolated markets, each market must face the "cold start" problem—this is not just a one-time challenge, but one that needs to be addressed every time a new market is created. Each market can only rely on its participants, and liquidity may not be sufficient to support large-scale lending activities.
Source: Solend — limited liquidity available on isolated markets
As mentioned earlier, some protocols have pushed the concept of isolated lending markets to the extreme, achieving permissionless market creation.
In these cases, such as Rari or @Solendprotocol, users can create permissionless markets themselves, deciding on whitelisted assets, setting risk parameters (such as loan-to-value ratio LTV and collateral ratio CR), and managing corresponding incentive mechanisms.
Shared liquidity pools: Deep liquidity from day one
On the other hand, a single shared liquidity pool provides deep immediate liquidity from the start. By consolidating all assets into a unified pool, shared liquidity systems can support large-scale lending activities with fewer liquidity constraints, even supporting newly added assets.
Lenders can also benefit from shared pools: a larger liquidity base attracts more borrowers, resulting in higher yields, which are typically more stable as they are supported by diversified borrowing demand.
This is the main advantage of the shared liquidity model, and although it may be the only advantage, its importance cannot be overstated. In every market, liquidity is king, especially in the crypto market.
However, the main drawback of shared liquidity pools is systemic risk. Since all assets are tied to the same pool, problems with one asset (such as sudden devaluation) can trigger a chain reaction of liquidations, which, if they result in bad debts, may affect the entire system.
Thus, these pools are less suitable for niche or more experimental assets, especially compared to liquid and mature tokens.
Finally, the governance and risk monitoring of shared liquidity systems are generally more complex, as the risks involved in any changes to the protocol are greater.
Model integration: Exploring hybrid models
The trade-offs between isolated liquidity pools and shared liquidity pools are significant, and there is no perfect solution. This is why, as markets mature, lending markets are gradually moving towards hybrid models (or at least introducing hybrid functions) to balance the liquidity advantages of shared pools with the customization and risk isolation offered by isolated markets.
A typical hybrid example is Aave's introduction of customized isolated markets, which collaborate with platforms like @LidoFinance and @Ether_Fi. Aave's system generally uses a single shared liquidity pool to provide deep liquidity for major assets. However, Aave also recognizes the need for greater flexibility when supporting assets with different risk characteristics or use cases, thus creating markets for specific tokens or partnered projects.
Another key feature of @Aave aligns with this trend, namely the eMode design. eMode aims to optimize capital efficiency when trading related assets. Specifically, eMode allows users to unlock higher leverage and borrowing capacity for assets that are highly correlated in price (thereby significantly reducing the liquidation risk for these assets) by isolating specific positions, greatly enhancing capital efficiency.
Other protocols like @BenqiFinance and @VenusProtocol, which traditionally fall into the shared liquidity category, have made significant strides by introducing isolated pools targeted at specific sub-sectors. In these cases, isolated markets are tailored for niches such as GameFi, real-world assets (RWA), or "ecosystem tokens," and do not affect the operation of mainstream pools.
Meanwhile, isolated market lending platforms like Compound or Solend often have a "main pool" as a shared liquidity pool, or in the case of Compound, they have recently started adding more assets to the pool with the highest liquidity, effectively moving towards a hybrid model.
Note: Solend initially adopted a shared liquidity model, which was later changed in its design.
The business model of the cryptocurrency lending market
The core business model of the crypto lending market revolves around generating income through various mechanisms related to lending and collateralized debt positions (CDPs).
1. Interest rate differentials: The primary source of income for lending markets comes from the difference between borrowing and lending interest rates. Users can earn interest by depositing assets into the protocol, while borrowers must pay interest to access liquidity. The protocol profits from the difference between the borrowing rates paid by borrowers and the deposit rates received by depositors. This spread is typically small, but as more users participate in the protocol, earnings accumulate over time. For example, in Aave v3's Ethereum market, the deposit rate for $ETH is 1.99%, while the borrowing rate is 2.67%, resulting in a spread of 0.68%.
2. Liquidation fees: Lending markets also generate additional income through liquidation fees. When a borrower's collateral falls below the required threshold due to market fluctuations, the protocol initiates liquidation procedures to maintain system solvency. Liquidators pay off part of the borrower's debt in exchange for discounted collateral. Typically, the protocol retains a portion of this reward, and in some cases, the protocol itself operates liquidation bots to ensure timely liquidations and generate additional income.
3. CDP-related fees: Some protocols charge specific fees for their CDP (Collateralized Debt Position) products, which may come from interest on borrowed CDP assets, charged either over time or as a one-time fee (or both).
4. Flash loan fees: Most protocols allow users to take out flash loans and charge a small but very profitable fee. A flash loan is essentially a loan that must be repaid within the same transaction, allowing users to instantly access the capital needed to perform specific actions (such as liquidation).
5. Treasury revenues: Protocols sometimes also utilize their treasury to generate income, usually opting for the safest return methods.
It is these mechanisms that have made lending markets one of the most profitable protocols.
These fees are sometimes shared with governance tokens, redistributed through incentive mechanisms, or used to cover operational expenses.
Risk <> Lending Markets
As previously mentioned, operating a crypto lending market can be one of the most profitable businesses, but it is also one of the riskiest.
One of the first challenges faced by emerging lending markets is the "cold start" problem.
The cold start problem refers to the difficulty of launching liquidity in a new protocol or market. Due to concerns about insufficient liquidity, limited borrowing opportunities, and potential security vulnerabilities, early users are often reluctant to commit funds to a liquidity pool that has not yet reached sufficient scale. Without enough initial deposits, interest rates may be too low to attract borrowers; borrowers may find it difficult to obtain the loans they need or face excessively volatile rates due to liquidity changes.
Protocols typically address the cold start problem through liquidity mining incentives, where users earn native tokens as rewards for providing or borrowing liquidity (the incentives for one party indirectly affect the other, especially in cases where cyclical borrowing is possible). However, if these incentives are not managed effectively, they may lead to unsustainable token issuance, which is also a trade-off that protocols need to consider when designing launch strategies.
Timely liquidations are another key element in maintaining the protocol's solvency. When a borrower's collateral value falls below a certain threshold, the protocol must liquidate it to prevent further losses. This faces two main issues:
First, the success of this process largely depends on the liquidator—whether operated by the protocol party or managed by a third party—they need to monitor the protocol in real-time and execute liquidations quickly.
Source: Chaos Labs Benqi Risk Dashboard
To ensure that liquidations can proceed smoothly, liquidators need to receive sufficient incentives through liquidation rewards, which must be balanced with the protocol's revenues.
Secondly, the liquidation process must be triggered when it is economically safe to liquidate: if the value of the seized collateral is close to or nearly equal to the outstanding debt, the risk of that position slipping into bad debt increases. In this process, defining safe and up-to-date risk parameters (such as loan-to-value ratio LTV, collateral ratio CR, and setting a liquidation buffer between these parameters and the liquidation threshold) is crucial. At the same time, the whitelisting of assets on the platform needs to go through a rigorous selection process.
Additionally, to ensure the smooth operation of the protocol, ensure timely liquidations, and prevent user abuse of functionalities, lending markets largely rely on functional oracles that provide real-time collateral valuations and indirectly reflect the health and liquidability of loan positions.
Oracle manipulation is a significant risk, especially in protocols that rely on low-liquidity assets or single-source oracles, where attackers can trigger liquidations by distorting prices or borrow at incorrect collateral levels. There have been multiple similar incidents in the past, the most famous being Eisenberg's exploitation of a vulnerability in Mango Markets.
Delays and latency are also key factors; during market volatility or network congestion, delays in price updates can lead to inaccurate collateral valuations, resulting in lagging or mispriced liquidations, ultimately causing bad debts. To address this issue, protocols often adopt a multi-oracle strategy, aggregating information from multiple data sources to improve accuracy, or setting up backup oracles in case the primary data source fails, while also using time-weighted price feeds to filter out sudden changes in asset value caused by manipulation or outliers.
Finally, we must also consider security risks: among projects that suffer attacks, money markets are typically the second most affected after cross-chain bridges.
Managing a lending market's code is extremely complex, and only a few protocols can proudly say they have an impeccable background in this regard. At the same time, we see many protocols, especially forks of some complex lending products, encountering multiple security vulnerabilities when modifying or handling the original code. To mitigate these risks, protocols usually take measures such as bug bounties, regular code audits, and strict processes for approving protocol modifications. However, no security measure can be foolproof; the possibility of exploits is always a continuous risk factor that a team needs to be careful about.
How to handle losses?
When protocols incur losses, whether due to bad debts from failed liquidations or sudden events like hacker attacks, there is usually a standard mechanism in place to share the losses. Aave's approach can serve as a typical example.
Aave's Safety Module serves as a reserve mechanism to cover potential funding gaps in the protocol. Users can stake AAVE tokens in the Safety Module and earn rewards, but if necessary, these staked tokens may be reduced by up to 30% to cover losses. This acts as an insurance mechanism, and recently, this mechanism has been further enhanced by the introduction of stkGHO.
These mechanisms essentially provide users with "high risk, high reward" opportunities and align user interests with the overall interests of the protocol.
