DeFi — Lending, Bridges, Liquidity, and Collateral
How DeFi actually works in practice: lending markets, cross-chain bridges, liquidity programs, collateral risk, and the integrity of stablecoin design.
DeFi is easy to ship and hard to harden. Most of what is worth reading about decentralised finance is not the launch claim but the operational follow-through: what the protocol does in a drawdown, how it handles a bridge failure, and whether the parameters hold when correlations break.
Coverage at this desk runs across lending, bridges, liquidity programmes, and the collateral assumptions that quietly shape every yield figure.
Lending markets
Lending protocols look conservative on paper until correlated liquidations test the assumption that collateral types behave independently. Capital efficiency claims usually redistribute risk; they rarely reduce it.
When TrueFi extended decentralised loans to verified institutional borrowers, the working questions were underwriting framing and default expectations. Both apply more widely to any permissioned-credit DeFi product.
Bridges
Bridges have been the highest-leverage attack surface in DeFi for years. Even careful designs need to keep proving themselves, and a non-custodial label is not a guarantee of security.
Interlay's xCC bridging work and the trustless BTC-stablecoin bridge on Polkadot both surface the same theme: collateral assumptions and oracle risk decide whether trustless actually means trustless.
Liquidity programmes
Token-incentive driven liquidity compresses sharply once the incentive ends. Headline APR figures should be read against the durability of the underlying utility, not as forward indicators.
Coverage of Equilibrium's EQ Blast, EQIFi's payment integrations, and Chains.com's card onramps all treat the durability question as the central one.
Risk language
Yield figures should always be read alongside underlying collateral risk and smart contract risk. The denominator changes faster than the numerator. Read primary documents before allocating any capital.