📖 How DeFi lending and borrowing protocols work
A DeFi lending & borrowing protocol is a set of smart contracts that accept deposits into a liquidity pool, issue collateralized loans, and trigger collateral liquidation according to predefined parameters: the interest rate model, LTV (loan-to-value — the ratio of debt to collateral value), liquidation threshold (the LTV level after which the protocol starts selling collateral), and liquidation bonus (the premium paid to a liquidator for closing a risky position).
This material explains how DeFi protocols issue collateralized loans, calculate LTV (loan-to-value — the ratio of debt to collateral value), use price oracles to liquidate positions, and what risks smart contracts, DAOs, and market volatility create. APR and APY are used to compare yield.
In a bank loan, part of the risk is covered by reserves, regulations, and a manual assessment of the borrower’s solvency by a bank employee; in DeFi, the position is closed programmatically when the collateral price falls and LTV crosses the liquidation threshold, because the smart contract sells part of the collateral to repay the debt.
When stablecoins are used, risk appears in de-peg scenarios: when the oracle records the stablecoin price below $1, the protocol’s calculation of collateral or deposit value changes, and some positions move into liquidation because LTV rises.
DeFi lending is a collateralized loan against a crypto asset, where debt size is limited by LTV, while loss risk is shaped by three parameters: the collateral price from the oracle, the liquidation threshold, and the speed of the market price decline.
A DeFi loan is issued against collateral, while liquidation risk is defined by LTV, the liquidation threshold, and the collateral price from the oracle.
ℹ️ What DeFi lending & borrowing protocols are and why they are used
DeFi lending & borrowing protocols are money market smart contracts: one contract holds lenders’ deposits as a shared liquidity pool, a second contract tracks borrowers’ debts, and separate protocol functions calculate the interest rate, check collateral, and trigger liquidations.
A loan opens without a manual decision: the borrower deposits collateral, the protocol values the collateral using the oracle price, and immediately limits the debt size through the LTV parameter. LTV is the upper limit of debt relative to collateral value: with $10,000 in collateral and 50% LTV, the protocol allows roughly $5,000 of debt. If the collateral price falls, the collateral value in the protocol’s calculation decreases, LTV rises without the debt increasing, and when the liquidation threshold is reached, the smart contract sells part of the collateral to repay the debt.
Core elements of DeFi lending protocols:
The lender earns yield because the borrower pays interest on the debt, and that interest is distributed to pool depositors. The borrower receives liquidity without selling collateral, but accepts liquidation risk when LTV rises.
Loss of funds can occur not because of the collateral price, but because of a smart contract vulnerability or a price oracle error. Basic interface and contract checks are described in the material on how to check a smart contract before connecting.
DeFi lending works as an automated money market: the rate depends on pool utilization, the debt size is limited by LTV, and liquidation is triggered by the collateral price that the protocol receives through an oracle.
Liquidation in DeFi occurs automatically if the collateral price falls and LTV crosses the liquidation threshold.
⚠️ DeFi lending risks: liquidations, smart contracts, and DAO decisions
Key DeFi lending risks:
Typical loss scenarios in lending protocols: a sharp 40–50% collateral price decline over a short period, an oracle failure that sends the protocol an asset price close to zero, or a market parameter change through a DAO vote after which the current position LTV moves closer to the liquidation threshold.
Factors that reduce the probability of risk materialization:
Stress scenario: calculating a 50% collateral price decline in advance shows what amount of collateral top-up or debt repayment returns LTV to a safe zone after a sharp market move.
Risk in DeFi lending is defined by liquidation parameters and price sources; reducing liquidation probability comes down to a collateral buffer (low LTV), health factor monitoring, and available liquidity to top up collateral or partially repay debt.
A comparison of ten protocols helps compare yield sources and risk classes (liquidations, oracle failures, network risks, token incentives) under one template.
📊 Overview of DeFi lending protocols: yield, risks, and use cases
The comparison captures use cases, risk parameters, fees, and network constraints for deposits and borrowing.
Liquity
GlobalLiquity is a protocol where the borrower locks ETH as collateral and mints the LUSD stablecoin; no interest accrues on the debt, and borrowing cost is defined by a one-time fee when the position is opened.
- Best suited for: ETH holders who need a stablecoin loan against collateral without selling ETH.
- Appropriate when: borrowing against ETH, where the collateral buffer accounts for sharp ETH price moves and proximity to the liquidation threshold.
✅ Pros
- No interest accrues on the debt: borrowing cost is defined by a one-time fee when the position is opened.
- Protocol parameters do not change through DAO voting: the risk of a sudden LTV increase or liquidation threshold change is lower.
- LUSD is backed by ETH, and its return toward roughly $1 is supported by arbitrage between minting and redemption.
- Stability Pool: a LUSD deposit receives liquidated ETH at a discount in exchange for repaying the debt of liquidated positions.
- LQTY: token holders receive part of the protocol fees, which adds income to strategies using the Stability Pool.
❌ Cons
- High liquidation risk during sharp ETH moves: at the minimum collateralization level, a 10–15% ETH price drop can move the position close to liquidation.
- Limited asset set: only ETH can be used as collateral and only LUSD can be borrowed, so assets cannot be diversified inside the protocol.
- Risk of LUSD deviating from $1 during panic periods: closing a position may require buying LUSD above $1.
- The model is not suitable for a money market with dozens of assets: the protocol solves one task — borrowing against ETH.
Liquity is suitable for borrowing against ETH without accruing interest, but it requires strict collateral control because the liquidation threshold is close to the position’s operating range.
Aave
GlobalAave is a lending protocol with large liquidity pools for ETH and stablecoins, operating across several networks and providing money market functions: deposits, borrowing, liquidation, and variable rates.
- Best suited for: stablecoin deposit scenarios and borrowing against ETH/WBTC in a high-liquidity pool.
- Appropriate when: fees are critical: Aave is available on L2 (Layer 2 — networks built on top of Ethereum), including Arbitrum, Optimism, and Polygon.
✅ Pros
- Large liquidity: deposits and borrowing for major assets are available without a liquidity shortage.
- Multi-network access: network choice reduces fees and simplifies position rebalancing.
- Wide asset list: stablecoins, ETH, WBTC, and several major tokens are available in one interface.
- Protection module: deficit coverage mechanisms create an additional buffer for rare systemic events.
- Flash loans (a loan borrowed and repaid within one transaction) expand arbitrage and hedging strategies for advanced users.
❌ Cons
- Variable rates: deposit and borrowing rates change with pool utilization, so yield is not fixed inside the base product.
- A sharp increase in borrowing cost during liquidity shortages: rising utilization rate increases the rate and worsens borrowing economics.
- Different risk parameters across networks and assets: incorrect LTV and liquidation threshold selection shows up more often during market drawdowns.
- Risk surface is higher because of integrations: additional functions and external integrations increase the number of scenarios where configuration errors affect users.
Aave is suitable for deposits and borrowing in core assets if the variable rate and collateral buffer before the liquidation threshold are controlled.
Compound
GlobalCompound is an Ethereum money market protocol where deposit and borrowing rates are calculated algorithmically from pool utilization, while market parameters change through votes by COMP token holders.
- Best suited for: basic deposit and borrowing scenarios for ETH and stablecoins without additional modules.
- Appropriate when: a variable rate is acceptable and a simple money market model is important.
✅ Pros
- Long operating history: the protocol has passed through several market cycles and remains a benchmark for the basic money market model.
- Simple logic: deposits and borrowing without a large number of modes and add-ons.
- Algorithmic rates: the rate balances supply and demand through the utilization rate.
- COMP governance: market parameter changes are published through an on-chain voting process.
- Availability across several networks: migration to lower-fee networks improves the economics of smaller operations.
❌ Cons
- The asset list is limited: rare tokens often do not meet listing risk and liquidity requirements.
- The rate is not fixed: when demand falls, deposit yield declines along with pool utilization.
- Feature development is slower: several competitors add new modes and markets more quickly.
- COMP rewards may be small: total yield may match the ordinary market rate without a premium.
Compound is suitable for basic deposits and borrowing on Ethereum if the priority is a simple money market model and a transparent link between rates and utilization rate.
Spark Lend (Sky)
GlobalSpark Lend is a lending protocol from the Sky Protocol ecosystem, formerly MakerDAO, focused on borrowing and deposits in stablecoins and liquid assets.
- Best suited for: DAI/USDS scenarios that use stablecoin deposits or borrowing against liquid assets.
- Appropriate when: the risk of parameter changes through Sky governance is accepted.
✅ Pros
- USDS/DAI rates are formed inside the ecosystem: yield relies on internal cash flows from the Sky model.
- Stablecoin borrowing rates are often below the average level: the economic model is focused on liquidity in base stablecoins.
- ETH and LSD (liquid staking derivative — a liquid staking token) support: ETH, wstETH/rETH, and WBTC can be used as collateral.
- Architecture based on Aave v3: core money market mechanics inherit a tested model.
- Growth in stablecoin liquidity improves deposit and borrowing conditions.
❌ Cons
- Parameters depend on Sky Protocol governance: changes to limits and risk settings alter borrowing cost and liquidation risk.
- Focus on a limited set of markets: the asset set is usually smaller than in universal money markets.
- Shorter operating history: the protocol has passed through fewer stress periods where liquidations and price oracle correctness are tested.
- Token incentives may change: incentive changes affect final deposit yield.
Spark Lend can be used for DAI/USDS strategies if market risk parameter changes through Sky governance (limits, LTV, and liquidation thresholds) are monitored and a collateral buffer before the liquidation threshold is maintained.
Morpho
GlobalMorpho is an optimization layer on top of Aave and Compound that matches lenders and borrowers directly, and if no match is available, sends liquidity to the underlying pools.
- Best suited for: scenarios that seek a higher rate compared with Aave/Compound without changing the base protocol.
- Appropriate when: it is acceptable that part of the position may operate as a regular pool deposit when P2P matching (direct matching of a deposit and loan between users) is unavailable.
✅ Pros
- P2P matching increases depositor income and reduces the borrower’s rate when the protocol finds a direct match.
- Assets sit in base protocols: when no match exists, Morpho uses Aave/Compound as a liquidity backend.
- Optimization fees are low: most of the gained rate remains in the position.
- Access to base protocol markets: stablecoins, ETH, WBTC, and major tokens are available through integrations.
- Audits and bug bounties reduce the risk of a missed vulnerability in the optimization code.
❌ Cons
- A layer on top of the base protocol adds complexity: rate assessment depends on the share of the position in P2P and the share in the pool.
- Base protocol risks carry over to the position: Morpho places liquidity inside Aave and Compound pools, so a failure, liquidation, or incident in those pools directly affects the Morpho position.
- LTV and liquidation parameters are inherited: Morpho uses the risk parameters of Aave or Compound markets, so risk configuration is limited by values set in those protocols.
- Incentives and rewards depend on markets: yield changes when reward programs are enabled or disabled.
Morpho is used as a rate optimization layer while preserving dependence on the risk parameters and incidents of the underlying protocols.
JustLend DAO
GlobalJustLend is a lending protocol on the Tron network where deposits and loans operate under a money market model with low network fees.
- Best suited for: deposit and borrowing scenarios on Tron using TRX and USDT-TRC20.
- Appropriate when: frequent operations are expected and the effect of network fees on deposits and withdrawals is critical.
✅ Pros
- Liquidity in core Tron assets is often sufficient for deposits and borrowing.
- Low network fees simplify frequent collateral top-ups and partial debt repayments.
- Tron-native scenarios connect with each other: lending can be combined with other Tron applications.
- The protocol is widely used inside Tron, which increases the probability of available liquidity.
- DAO governance: token holders participate in changes to market parameters.
❌ Cons
- Governance and network validator risk is higher: protocol operation depends on Tron’s resilience.
- The strategy is tied to one network: moving capital to other ecosystems requires bridges and adds bridge infrastructure risk.
- Fewer independent analysis tools: parameter and risk monitoring sometimes has to be done manually.
- JustLend pool utilization on Tron changes with borrower and depositor activity; when pool utilization rises, the interest rate is recalculated upward by the protocol’s interest rate model.
JustLend can be used for low-cost operations on Tron, but network risk and governance risk come along with the fee savings.
Venus Protocol
GlobalVenus is a lending protocol on BNB Chain that combines a money market (deposits/borrowing) with incentives through the XVS token.
- Best suited for: deposit and borrowing scenarios on BNB Chain using ecosystem assets.
- Appropriate when: BNB Chain network risk and dependence of final yield on XVS incentives are accepted.
✅ Pros
- Large BNB Chain market: liquidity for loans and deposits is more often available in popular assets.
- Low network fees improve the economics of small operations and frequent rebalancing.
- Wide ecosystem asset set: BNB, stablecoins, and several tokenized assets are available in one protocol.
- XVS incentives can noticeably add yield to deposits in specific markets.
- Reserves and audits reduce the risk of missed vulnerabilities, although they do not eliminate it completely.
❌ Cons
- History of stress events in specific assets: risk parameters may change sharply after incidents.
- BNB Chain network risk affects position management: a network failure or fee increase changes the ability to rebalance.
- Vote concentration in the DAO: large XVS holders can influence market parameters.
- Inflationary pressure from the incentive token: XVS rewards may lose value and reduce final yield.
Venus can be used for deposits and borrowing on BNB Chain if network risk and dependence of yield on XVS rewards are taken into account.
Euler Finance
GlobalEuler is a lending protocol on Ethereum that creates isolated markets: the risk of one asset has less impact on other protocol markets.
- Best suited for: isolated market scenarios for specific tokens with separate risk parameters.
- Appropriate when: complex configuration risk and limited liquidity in niche assets are accepted.
✅ Pros
- Isolated markets reduce system contagion: a problem with one token is less likely to create a deficit across all markets.
- Support for less liquid tokens expands the list of strategies unavailable in conservative protocols.
- Partial liquidations reduce the one-time collateral sale volume and lower the price impact on the market.
- Flexible price sources: TWAP (time-weighted average price — the average price over a period) reduces the impact of short-term price spikes.
- Audits and bug bounties reduce the risk of missed code errors, although they do not eliminate it completely.
❌ Cons
- Complex market configuration: an error in assessing liquidity and oracles leads to unexpected liquidations and slippage.
- Past incidents show the risk of complex attacks: an audit does not guarantee the absence of vulnerabilities.
- Low liquidity in niche assets limits borrowing size and complicates position exit.
- Ethereum L1 fees (Layer 1 — Ethereum’s base network) worsen the economics of frequent operations if the strategy requires regular transactions.
Euler is used for isolated markets and niche assets if liquidity and price source correctness are controlled.
Maple Finance
GlobalMaple is a platform that issues on-chain loans to institutional borrowers with partial collateral or without full collateral; depositor income comes from interest payments on such loans.
- Best suited for: scenarios where borrower default credit risk is accepted in exchange for a higher interest rate.
- Appropriate when: term capital lockup and limited exit liquidity are acceptable.
✅ Pros
- Borrowers are institutional organizations; the loan is tied to a specific borrower and loan terms, not an anonymous address.
- Pool rates are often higher than rates in classic overcollateralized protocols.
- Loan terms are visible on-chain: maturity, rate, and parameters are fixed in contract logic.
- Risk delegate (a participant responsible for borrower selection and loan parameters) manages the pool: loan parameters and borrower selection are concentrated in one governance module.
- The portfolio format distributes risk across several loans instead of one position.
❌ Cons
- Default risk is direct: if the borrower becomes insolvent, the depositor bears losses.
- Exit liquidity is lower: some pools lock capital until the end of the loan term.
- The entry threshold is higher: the model is more convenient for larger amounts and longer horizons.
- Dependence on selection quality: a risk delegate’s mistake affects all pool participants.
- Regulatory uncertainty has a stronger effect on institutional lending models.
Maple is used for exposure to corporate credit risk in an on-chain format while accepting default risk and lower exit liquidity.
Pendle
GlobalPendle is an adjacent DeFi yield instrument, not a classic lending protocol: it splits a position into the asset principal and future interest — PT (Principal Token) and YT (Yield Token) — after which these parts can be traded separately.
- Best suited for: fixed-term yield locking scenarios and trading expectations for rates.
- Appropriate when: maturity risk, PT/YT market liquidity risk, and price risk of yield tokens are considered.
✅ Pros
- Yield locking: selling YT turns future interest into a fixed payout for the selected term.
- Rate trading: the YT price reacts to yield expectations, giving tools for speculation and hedging.
- Integrations with yield sources: positions can rely on lending, staking, and LP yield (yield from a liquidity provider position).
- Markets in core assets are often more liquid: entering and exiting is simpler in ETH and stablecoins.
- PENDLE incentives support liquidity in specific yield markets.
❌ Cons
- Product complexity: mistakes in understanding PT/YT lead to confusion between yield and token price risk.
- Rate risk: when the yield of the underlying source falls, the YT price usually falls with it, and the strategy becomes unprofitable.
- Liquidity is unevenly distributed: spreads and slippage are higher in niche markets.
- Maturity management is mandatory: before the redemption date, the PT/YT price can fluctuate noticeably even if the underlying protocol operates normally.
Pendle is used to lock yield or trade rate expectations while controlling position maturity and PT/YT market liquidity.
A comparison by network, assets, yield type, and risk profile helps compare protocols without mixing different lending models.
📋 Comparison of DeFi lending protocols: parameters for evaluating terms
Four parameter groups are used for comparison: network, core assets, yield format, and risk profile.
| 🏦 Protocol | 🌐 Networks | 💰 Key assets | 📈 Yield type | ⚠️ Risk profile |
|---|---|---|---|---|
| Aave | Ethereum, L2 (Arbitrum, Optimism, Polygon) | ETH, WBTC, stablecoins | Often 2–6% on stablecoins, higher on altcoins | 🟢 Low/medium, high liquidity in core assets |
| Compound | Ethereum, L2 | ETH, stablecoins | Often 2–4% in conservative markets | 🟢 Low, simple money market model |
| Spark Lend | Ethereum | DAI, USDS, ETH, LSD | Often 4–5% on stablecoins | 🟢🟡 Low/medium, parameters depend on Sky governance |
| Morpho | Ethereum | Aave/Compound assets | Rate optimization (P2P + pools) | 🟡 Medium, inherits risks of base pools |
| JustLend | Tron | TRX, USDT-TRC20, USDJ | Market yield with low fees | 🟡 Medium, network governance risk |
| Venus | BNB Chain | BNB, BTCB, ETH, stablecoins | Market yield + XVS rewards | 🟡🔴 Medium/elevated, BNB Chain network risk and XVS incentive risk |
| Euler | Ethereum | Wide range of ERC-20 (Ethereum token standard) | Higher rates in riskier markets | 🟡🔴 Medium/elevated, complex configuration |
| Maple Finance | Ethereum, EVM networks (networks compatible with Ethereum Virtual Machine) | USDC, stablecoins | Institutional loans (credit risk) | 🔴 High, default risk |
| Pendle | Ethereum, L2 | stETH, lending/LP positions | Yield locking and yield trading | 🟡🔴 Medium/high, maturity and liquidity risk |
| Liquity | Ethereum | ETH / LUSD | 0% on debt, income through liquidations and LQTY | 🔴 High, low collateralization threshold |
When comparing rates, it is important to clarify the metric format (APR or APY) and the accrual method; when assessing risk, check LTV, liquidation threshold, the oracle price source, and the protocol’s incident history.
Strategies show how deposits, borrowing, and leverage change the yield source and liquidation risk through LTV growth when the collateral price falls.
🧩 Basic DeFi lending strategies: using protocols in typical scenarios
- Stablecoin deposit (often 2–5% annually across the market)
- Placing USDT, USDC, or DAI in a liquidity pool without opening a loan.
- Income comes from interest paid by borrowers in that pool and possible protocol rewards in tokens.
- Liquidation risk is absent because no debt is opened; the main risks are smart contract vulnerability and stablecoin de-peg.
- The holding period affects the share of network fees in the final result when funds are deposited and withdrawn.
- Long ETH through borrowing stablecoins (moderate leverage)
- ETH is deposited as collateral, against which stablecoins are borrowed.
- Part of the borrowed stablecoins can be swapped for ETH and added to collateral, which increases exposure to ETH.
- 30–40% LTV leaves a buffer before the liquidation threshold; a 20–30% ETH price correction raises LTV and moves the position closer to the liquidation zone.
- LTV (loan-to-value) is the ratio of debt to collateral value; LTV rises when the collateral price falls or debt grows because of interest.
- LSD + lending: staked ETH as collateral
- ETH is staked, and the received LSD (liquid staking derivative — a liquid staking token, such as wstETH) is used as collateral in a lending protocol.
- Income consists of staking rewards (part of network rewards) and lending yield/incentives if the position is additionally placed in a pool.
- Correlation increases risk: when ETH falls, both the collateral and the staking token usually get cheaper, so LTV rises faster.
- 35–40% LTV and a stress scenario with a 40–50% ETH decline reduce the probability of liquidation during a sharp market move.
- Institutional credit risk (often 8–15% annually depending on the loan structure)
- Placing stablecoins in pools that issue loans without full collateral or with partial collateral.
- Income comes from borrowers’ interest payments, but in the event of default, part of the capital is lost because collateral is insufficient for full repayment.
- Capital is often locked for a fixed term, such as 30–90 days, and exiting earlier may be impossible.
- Default risk is not tied to collateral volatility and can materialize without a market move.
Moving from a deposit without borrowing to a debt-based strategy adds liquidation risk because a collateral price decline raises LTV and lowers the health factor.
A stablecoin deposit earns interest without liquidation risk, while any strategy with borrowing adds liquidation risk through LTV growth when the collateral price falls.
The FAQ captures the basic definitions and key mechanics of DeFi lending: TVL, liquidation, LTV, and why yield changes.
❓ FAQ on DeFi lending and cryptocurrency loans: key questions
What is TVL and why is it important for DeFi lending?
What happens when a loan position is liquidated?
What LTV level is considered safe for a long-term loan?
Which metrics determine a position’s liquidation risk?
Why does deposit yield (APY) in DeFi constantly change?
What are the main risks of DeFi lending?
Does using DeFi lending for small amounts make sense?
Which criteria are used to compare DeFi protocols for borrowing and deposits?
🧾 What investors should remember about DeFi protocols
DeFi lending & borrowing protocols turn a crypto asset into collateral or a deposit: the depositor receives interest from borrower payments, while the borrower receives stablecoins or other assets without selling collateral. In basic money market protocols, the rate depends on pool utilization, while position risk is defined by LTV and the liquidation threshold, which are recalculated using the collateral price from the oracle.
Losses occur in three ways: liquidation when the collateral price falls and LTV rises, a smart contract vulnerability or oracle failure that changes collateral value calculation, and market parameter changes through DAO voting. Risk control comes down to a collateral buffer (low LTV), health factor monitoring, and available liquidity to top up collateral or partially repay debt.