How lending protocols work
Two sides of every market:
- Suppliers deposit assets (ETH, USDC, WBTC) and earn a variable interest rate based on how much of the pool is being borrowed.
- Borrowers post eligible collateral, borrow other assets up to a protocol-defined loan-to-value ratio, and pay interest on the outstanding balance.
Interest rates adjust algorithmically: as utilization climbs, rates rise to attract more supply and deter borrows; as utilization falls, rates drop. The supply side earns a fraction of the borrow APR (typically 70-90%), with the protocol keeping a reserve factor.
Collateralization ratios vary by asset. On Aave, ETH might allow 82% LTV borrows; a more volatile altcoin might cap at 50%. Cross positions where you post multiple collaterals are common; isolated mode (one collateral per loan) reduces contagion risk.
What lending is used for
- Leveraged longs — deposit ETH, borrow USDC, buy more ETH, deposit again. Each loop amplifies exposure.
- Tax-efficient liquidity — borrow stables against appreciated crypto without selling (avoiding capital gains).
- Yield generation — lend idle stablecoins for 3-10% APY; lend ETH for 2-4% on top of validator yield via restaking integrations.
- Short selling — borrow an asset you expect to decline, sell for stables, buy back cheaper later.
Risks and considerations
The specific failure modes:
- Liquidation — collateral price drops or borrow balance grows past LTV; a keeper liquidates for a bonus, and the borrower loses a chunk of collateral (often 5-10%) beyond the repaid loan.
- Oracle manipulation — if the protocol’s price feed can be pushed off-market briefly, an attacker can borrow against inflated collateral or trigger mass liquidations. The Mango Markets 2022 exploit is a canonical case.
- Smart-contract bugs — Euler lost $200M in 2023 to a flawed accounting check; recovered most but not all.
- Bad debt — during rapid crashes, liquidations can’t always execute fast enough; the protocol ends up with under-collateralized positions and a hole in the supply side.
Modern protocols (Morpho Blue, Euler V2, Aave V3) use isolated markets and more conservative parameters specifically to contain these risks. For users, the practical advice: use established protocols with long audit histories, keep LTV well below the liquidation threshold (50-60% of max is a common rule), and diversify across protocols rather than stacking a whole portfolio in one.