How DeFi Lending Protocols Work

Introduction

DeFi lending protocols constitute one of the largest and most systemically important sectors of decentralized finance. As of 2024, lending protocols collectively hold tens of billions of dollars in total value locked (TVL), providing permissionless credit markets that operate 24/7 without intermediaries. These protocols enable users to earn yield on idle assets by supplying liquidity, while borrowers can access capital without credit checks, income verification, or lengthy approval processes.

This article provides a rigorous overview of how DeFi lending works at the protocol level, covering pool mechanics, interest rate models, tokenized positions, risk parameters, and how these systems compare to traditional lending infrastructure.

Supply and Borrow Pools

At the core of every DeFi lending protocol is the liquidity pool model. Unlike traditional lending, where a specific lender is matched to a specific borrower, DeFi protocols aggregate deposits into shared pools. Suppliers deposit assets into a smart contract, and borrowers draw from that same pool. This pooled architecture provides several advantages:

• Instant liquidity — Suppliers can withdraw at any time (subject to pool utilization), and borrowers can access funds immediately without negotiation.
• Continuous interest accrual — Interest compounds every block (approximately every 12 seconds on Ethereum), not on monthly or annual cycles.
• Permissionless access — Any address can supply or borrow without identity verification, provided collateral requirements are met.
• Transparent risk parameters — All pool states, interest rates, and collateral ratios are publicly visible on-chain.

Each asset typically has its own pool. For example, a protocol may have separate pools for ETH, USDC, WBTC, and DAI. The supply side earns interest from borrower payments, while the protocol may retain a small percentage as a reserve factor to build a safety buffer.

Interest Rate Models

DeFi lending protocols use algorithmic, utilization-based interest rate models to dynamically price the cost of borrowing. The key variable is the utilization rate, defined as:

Utilization Rate = Total Borrows / Total Deposits

When utilization is low, interest rates remain low to incentivize borrowing. As utilization increases, rates rise to attract more deposits and discourage excessive borrowing. Most protocols implement a kinked (piecewise linear) rate curve with two distinct slopes:

1. Below the optimal utilization threshold (typically 70–90%): Rates increase gradually along a gentle slope.
2. Above the optimal utilization threshold: Rates spike sharply along a steep slope, creating strong economic pressure to rebalance the pool.

This design ensures that pools never become fully utilized, preserving withdrawal liquidity for suppliers. The "kink" in the curve acts as a critical safety mechanism. For example, Compound's interest rate model for USDC may specify a base rate of 0%, a slope of 4% up to 80% utilization, and a jump slope of 75% beyond that threshold.

Interest-Bearing Tokens: cTokens and aTokens

When users deposit assets into a lending protocol, they receive interest-bearing tokens that represent their claim on the underlying deposit plus accrued interest. The two dominant implementations are:

cTokens (Compound)

Compound issues cTokens (e.g., cUSDC, cETH) at an exchange rate that increases over time. When a user deposits 1,000 USDC, they might receive 50,000 cUSDC at the current exchange rate. As interest accrues, the exchange rate increases, meaning those 50,000 cUSDC can be redeemed for progressively more USDC. The user's cToken balance remains constant; the value per cToken grows.

aTokens (Aave)

Aave uses a rebasing model with aTokens (e.g., aUSDC, aETH). When a user deposits 1,000 USDC, they receive 1,000 aUSDC. As interest accrues, the aToken balance increases automatically. The user's wallet shows a growing number of aTokens, maintaining a 1:1 peg with the underlying asset. This approach is more intuitive for users but introduces complexity for smart contract integrations.

Both models are ERC-20 compliant, meaning interest-bearing tokens can be transferred, traded, or used as collateral in other protocols, enabling composability across the DeFi ecosystem.

Health Factor and Risk Parameters

Because DeFi lending is overcollateralized (borrowers must deposit more value than they borrow), protocols use a health factor to measure the safety of each borrowing position:

Health Factor = (Collateral Value × Liquidation Threshold) / Total Borrows

A health factor above 1.0 means the position is solvent. A health factor at or below 1.0 triggers liquidation, where external liquidators repay part of the debt and claim collateral at a discount. Key risk parameters that govern lending positions include:

• Loan-to-Value (LTV) — The maximum percentage of collateral value that can be borrowed. For example, an LTV of 80% means a user can borrow up to $800 against $1,000 of collateral.
• Liquidation Threshold — The collateral ratio at which liquidation becomes eligible. This is always higher than the LTV to create a buffer zone.
• Liquidation Penalty — The bonus paid to liquidators, typically 5–10%, which incentivizes rapid liquidation of undercollateralized positions.
• Reserve Factor — The percentage of interest payments retained by the protocol as a safety reserve.

DeFi Lending vs. Traditional Lending

Major Lending Protocols

Aave

Aave is the largest DeFi lending protocol by TVL, deployed across Ethereum, Polygon, Arbitrum, Optimism, Avalanche, and other chains. It supports a wide range of collateral types, offers both variable and stable interest rates, and introduced flash loans as a core primitive. Aave V3 introduced features like efficiency mode (eMode) for correlated assets and isolation mode for newly listed, higher-risk assets.

Compound

Compound pioneered the pooled lending model and the cToken standard. Compound V3 (Comet) simplified its architecture by moving to a single-asset borrowing model, where each market has one borrowable asset (e.g., USDC) and multiple collateral assets. This design reduces risk surface and simplifies governance.

MakerDAO (Sky)

MakerDAO operates differently from Aave and Compound. Rather than lending out deposited assets, users deposit collateral into Vaults (formerly CDPs) and mint DAI, a decentralized stablecoin. The system functions as a decentralized central bank, with governance controlling stability fees (interest), liquidation ratios, and collateral types. MakerDAO's Peg Stability Module (PSM) maintains DAI's dollar peg through direct USDC-DAI swaps.

Key Risks

Despite their robust design, DeFi lending protocols carry meaningful risks:

• Smart contract risk — Bugs or exploits in protocol code can result in loss of funds.
• Oracle risk — Lending protocols depend on price oracles (Chainlink, Uniswap TWAP) for accurate collateral valuation. Oracle manipulation or failure can trigger improper liquidations or allow undercollateralized borrowing.
• Governance risk — Malicious or poorly considered governance proposals can alter risk parameters in dangerous ways.
• Liquidity risk — During high-utilization periods, suppliers may be unable to withdraw their deposits until utilization decreases.
• Cascade liquidation risk — Rapid price declines can trigger mass liquidations, further depressing prices in a reflexive feedback loop.