Section III: Lending and Borrowing Fundamentals

With on-chain price formation and liquidity established through DEXs, these pricing mechanisms enable the next layer of DeFi infrastructure: lending and borrowing. These protocols form the foundation of the ecosystem, providing the liquidity and leverage that power more complex strategies.

The most common safety metric for lending protocols is the Health Factor (HF), which measures how close a position is to liquidation. While Chapter VI covered liquidation in the context of centralized exchanges and perpetual futures, DeFi protocols implement similar mechanics entirely on-chain through smart contracts. The Health Factor is calculated based on the ratio of collateral value to debt value, adjusted for liquidation thresholds. An HF above 1 means the position is healthy; below 1 means it can be liquidated.

Aave: Building the Automated Lending Infrastructure

Aave operates like an automated bank that never closes, using smart contracts to evaluate collateral and approve loans based on pre-defined rules rather than human underwriters. The protocol has evolved significantly since its inception, with each version addressing real limitations users faced in practice.

For lenders, the process remains straightforward across all versions. A participant deposits assets like ETH, USDC, or other supported tokens into shared liquidity pools and immediately starts earning interest. Deposits are represented by special tokens called aTokens, whose balance in your wallet automatically increases over time as interest accrues. Borrowers must maintain more collateral than they borrow, a design known as over-collateralization. For example, depositing $1,000 of ETH might allow borrowing only $800 of USDC, with the $200 buffer protecting lenders from price volatility. This collateral requirement is fundamental to trustless lending, since protocols can't sue defaulters or garnish wages. They need sufficient assets on hand to liquidate when positions become unhealthy.

Who Uses Collateralized Lending

Aave's lending model serves multiple use cases that explain its popularity, with the protocol having around $60B in total deposits and nearly $25B in active borrows in early 2026. Many users want liquidity without selling assets they believe will appreciate, an ETH holder may need stablecoins for expenses or new opportunities. Borrowing preserves upside potential while deferring capital gains taxes that selling would trigger immediately.

Leveraged trades represent another major use case. Users deposit ETH, borrow stablecoins, then buy more ETH through "looping" strategies that amplify exposure, for example, depositing $1,000 of ETH, borrowing $800 USDC, buying more ETH, and repeating until the Health Factor approaches the participant's risk tolerance (e.g., HF ≈ 1.2 for aggressive leverage). Alternatively, staked assets like stETH can serve as collateral to boost yield through measured leverage, combining staking rewards with borrowing strategies.

Beyond basic lending, these platforms enable shorting and hedging by allowing users to borrow assets they expect to decline and sell them immediately, creating on-chain prime brokerage functionality. Safe shorting requires the borrowed asset to have sufficient liquidity and reliable oracle pricing to prevent manipulation during liquidations. This helps hedge concentrated positions or farming rewards without unwinding entire strategies, maintaining core exposure while managing specific risks.

Professional traders use the platforms for arbitrage and carry trades, borrowing cheap stablecoins to earn higher yields elsewhere and capturing futures basis, funding rate premiums, or liquid staking token spreads. These strategies exploit rate differentials across DeFi protocols and traditional markets.

Risk Management Through Key Parameters

Aave manages lending risk through parameters that determine borrowing limits and liquidation triggers. Loan-to-Value (LTV) ratios set maximum borrowing power per asset, an 80% LTV means depositing $100 allows borrowing up to $80. Liquidation thresholds define when positions become undercollateralized and eligible for liquidation, always set higher than LTV ratios to create safety buffers. Liquidation bonuses provide incentives for third parties to maintain system solvency by repaying bad debt in exchange for discounted collateral.

Interest rates adjust automatically based on pool utilization through mathematical curves. High demand increases rates to attract lenders and discourage excessive borrowing. Low utilization decreases rates to encourage borrowing and provide competitive returns. Markets self-balance through these algorithmic adjustments.

Evolution Through Protocol Versions

Aave v1 introduced the basic concept of pooled lending with interest-bearing tokens and pioneered flash loans (see Section V: Infrastructure Dependencies), enabling users to borrow and repay large amounts of capital within a single transaction for arbitrage and liquidations.

Aave v2 added debt tokenization (non-transferable tokens that represent the borrower's debt), plus credit delegation, collateral swaps, and repay-with-collateral, all of which improved composability and UX. The version also reduced gas costs and improved user experience. Credit delegation allowed trusted parties to borrow against others' collateral without direct access to the underlying assets.

Aave v3 brought targeted improvements for risk management and liquidity optimization. Isolation modes allowed the protocol to safely list long-tail assets without endangering the broader system, while efficiency modes offered better rates for closely correlated asset pairs like stablecoins. The protocol added variable liquidation close factors, allowing liquidators to close up to 100% of very unhealthy positions to remove bad debt efficiently.

The forthcoming Aave v4 represents a fundamental architectural shift. Instead of separate pools for each market, the protocol is moving to a Unified Liquidity Layer with a central Liquidity Hub and asset-specific Spokes. This design dramatically improves how markets share liquidity while maintaining safety through compartmentalized risk management per asset type.

This evolution illustrates DeFi's constant push toward better liquidity utilization while managing risk. Each version solved real problems users faced, from capital fragmentation to gas costs to risk isolation.

Aave's ecosystem extends beyond lending through GHO, its own over-collateralized stablecoin, transforming the platform from a simple lender into a broader monetary system. When users mint GHO by supplying collateral to Aave, the interest payments flow directly to the Aave DAO treasury, creating a revenue stream for the protocol itself. This makes GHO both a stablecoin and an integral part of Aave's ecosystem, governed entirely by Aave governance.

Euler and Morpho: Isolated Permissionless Markets

The pooled, blue-chip-focused design that Aave popularized is not the only way to build a lending protocol. Euler and Morpho push further toward isolated, permissionless markets with explicit separation between infrastructure and risk decisions.

Euler's original design already stood apart through permissionless listing and a tiered risk system that isolated riskier assets. Euler v2 expands this modular approach through the Euler Vault Kit (EVK), a framework for deploying credit vaults. Anyone can launch an isolated lending vault for an ERC-20 asset and configure custom parameters: accepted collateral types, LTVs and caps, interest rate models, and oracle sources. Each vault functions as its own market with its own risk parameters, so issues in one vault don't contaminate others. Tools like the Ethereum Vault Connector and EulerEarn connect vaults, enable cross-collateralization, and aggregate yields. Euler becomes a meta-lending layer that supports everything from conservative blue-chip markets to experimental long-tail configurations while preserving risk isolation.

Morpho evolved in a parallel direction. The project began as a P2P optimizer on top of Aave and Compound, but Morpho Blue re-architected it as a minimal trustless lending primitive. A Morpho Blue market is extremely simple: one loan asset, one collateral asset, a liquidation LTV, an oracle, and an interest rate model. Markets are permissionlessly created and fully isolated with parameters fixed at creation from governance-approved menus. Above this base layer sits MetaMorpho, a protocol for lending vaults built on Morpho Blue. Anyone can create a vault that allocates deposits across multiple Morpho Blue markets according to a strategy. This is where risk curators come in.

Risk Curators and Vault-Based Lending

A risk curator is an entity (often a specialized risk firm, DAO, or fund) that designs and deploys vaults, chooses which markets the vault supplies liquidity to and in what proportions, sets risk parameters at creation time within protocol constraints, and earns a fee for providing this risk management service. On Morpho, curators use MetaMorpho vaults to route depositor funds into selected markets. They decide which markets a vault can lend into, adjust allocation weights over time, and impose additional vault-level rules like caps and fee structures. Curators include specialist risk firms and DeFi-native asset managers: Gauntlet, Steakhouse, MEV Capital, RE7 Labs, and Moonwell have all launched or managed curated vaults.

There's an important distinction between risk service providers like Chaos Labs on Aave and risk curators on Morpho and Euler. On Aave, risk firms advise the DAO and publish parameter recommendations, but governance executes changes for all users. Users don't opt into specific risk managers; they use Aave's globally-set parameters. On Morpho and Euler, risk curators own the strategy for a given vault. Users choose a particular vault and thereby opt into that curator's allocation and risk decisions.

By early 2026, risk-curator-style vaults had grown to nearly $11 billion in deposits, about 10% of all DeFi lending TVL (Total Value Locked, the standard measure of assets deposited in a protocol), down from a $13 billion peak after de-risking. Several aggressive vaults that chased yield by accepting riskier stablecoins or thin-liquidity collaterals suffered losses or severe liquidity constraints, with some lenders temporarily stuck or taking haircuts when underlying assets de-pegged or markets froze.

This highlights both sides of the model. Risk curators can specialize, build sophisticated portfolios across many isolated markets, and offer higher risk-adjusted yields than generic pools. Long-tail assets can be supported without forcing every depositor to bear their risk. However, depositors are exposed not only to protocol-level smart contract and oracle risk, but also to curator behavior: their asset selection, concentration, and reaction speed in stressed conditions. For protocols like Morpho and Euler, the decision of which vault or curator to trust can be just as important as the choice of underlying protocol.

Sky: The Decentralized Central Bank

While Aave revolutionized peer-to-pool lending, another approach emerged that treats stablecoin issuance fundamentally differently. Sky (formerly MakerDAO) operates like a decentralized central bank that issues USDS stablecoins backed by crypto collateral and real-world assets. (For a broader overview of stablecoin types and mechanisms, see Chapter IX.)

The Vault system operates through protocol allocators ("Stars") who mint USDS via Vaults and deploy liquidity. Most end users typically upgrade DAI to USDS 1:1 or acquire USDS on markets, then opt into sUSDS to earn the Sky Savings Rate (SSR). Like Aave, the system requires collateral buffers, but the protocol creates newly minted stablecoins rather than lending from existing pools. This distinction matters because it means Sky can create new money supply based on collateral deposits.

Maintaining the peg requires multiple mechanisms working together. The LitePSM acts like an exchange window, enabling fixed-rate swaps between USDS/DAI and other stablecoins (like USDC) to help maintain the $1 peg. This provides immediate arbitrage opportunities when USDS trades away from $1. The Sky Savings Rate works like a demand lever, governance can adjust the rate to influence demand for holding and saving USDS, which supports the peg by making the stablecoin more attractive to hold.

Sky represents evolution from its original DAI system to the new USDS framework, with DAI and USDS currently coexisting during the Sky rebrand and voluntary upgrade migration. The protocol increasingly backs stablecoins with real-world assets like Treasury bills alongside crypto collateral, blending DeFi innovation with traditional finance stability.

Wildcat: Institutional Credit On-Chain

Both Aave and Sky require substantial collateral buffers, but Wildcat brings traditional credit relationships on-chain instead. The protocol connects institutional borrowers like market makers, hedge funds and even protocols with crypto lenders seeking potentially higher yields than fully-collateralized protocols can provide.

This alternative approach stems from a fundamental difference in collateralization philosophy. Unlike Aave and Sky's asset-backed collateral, Wildcat is intentionally under-collateralized and relies on a reserve-ratio liquidity buffer rather than full asset backing. This fundamental difference explains why Wildcat can offer higher yields while introducing explicit counterparty credit risk.

Wildcat operates as a marketplace where borrowers set all key parameters including fixed APR rates, lockup periods, and withdrawal windows without any protocol-level underwriting. They can also implement access control through allowlists or enable self-onboarding with OFAC screening via Chainalysis oracle. Additionally, borrowers may require lenders to sign legal agreements off-chain to establish formal credit relationships.

Risk management mechanics become especially critical when things go wrong. If reserves fall below the required level, the market becomes delinquent and withdrawals are restricted while penalty fees accrue until the borrower replenishes reserves. Actual losses only materialize if the borrower ultimately defaults, which is why Wildcat requires participants to actively manage counterparty risk through due diligence on borrower reputation.

These risks aren't merely theoretical, they materialized in mid 2025 when Kinto, a DeFi platform that had borrowed through Wildcat's facility following a major hack, announced its shutdown and became Wildcat's first official default. There were more than ten lenders in Kinto's facility and they faced a 24% haircut, recovering 76% of their principal from the borrower's remaining assets. This default demonstrated both the isolation of losses to specific facilities, with no contagion to Wildcat's other $150+ million in outstanding loans, and the real-world implications of Wildcat's undercollateralized lending model.

The Kinto default illustrates a broader principle about DeFi's evolution: while programmability doesn't eliminate credit risk, it can make it more transparent and controllable through fully on-chain, transparent credit markets with customizable terms. Wildcat represents this philosophy in practice, bringing traditional credit relationships into the programmable, transparent world of DeFi.