A return with no clear source is not a bargain — it's an unmapped risk. The single most useful habit in evaluating any DeFi yield opportunity is refusing to move forward until you can state, in one sentence, exactly which economic activity is generating the payment you'd receive. This article walks through the main sources, in roughly increasing order of complexity and risk.
Lending
You deposit an asset into a lending protocol; borrowers post collateral and pay interest to borrow it; you receive a share of that interest. This is the most legible yield source in DeFi — it's structurally similar to how a bank pays interest on deposits, except the "bank" is a smart contract, collateral requirements are enforced automatically, and interest rates float in real time based on supply and demand within the protocol rather than being set by a committee.
Primary risks: smart contract failure, a collapse in collateral value outrunning the liquidation engine's ability to keep up, and — during periods of high withdrawal demand — temporary illiquidity if too much of the deposited asset is currently lent out.
Liquidity provision
You deposit a pair of assets into a decentralized exchange's liquidity pool; traders swap against that pool and pay a fee; you earn a share of fees proportional to your share of the pool. This is closer to running a small market-making operation than to lending — your return depends on trading volume, not on anyone paying you interest directly.
Primary risk — impermanent loss: when the price ratio between your two deposited assets shifts, the pool automatically rebalances your holdings toward the asset that fell in relative value and away from the one that rose. Compared to simply holding both assets separately, this can leave you with less value than you'd have had without providing liquidity at all — a cost that's easy to underweight when only looking at the advertised fee yield.
Staking
On proof-of-stake networks, staking means locking up the network's own asset to help secure it, in exchange for a share of newly issued tokens and transaction fees — essentially getting paid for taking on the validator role, or delegating your stake to someone who does. Compared to lending or liquidity provision, staking's yield is tied directly to the base network's own security model rather than to a separate application built on top of it.
Primary risks: "slashing" — an automatic penalty, including partial loss of staked funds, if a validator misbehaves or goes offline — and lock-up periods that can prevent withdrawal during exactly the moments you might most want liquidity.
Layered and leveraged strategies
Many of the highest advertised yields in DeFi combine the above mechanisms — for example, depositing an asset as collateral, borrowing against it, and redeploying the borrowed funds into another yield source, repeated across multiple protocols. Each layer typically adds a separate smart contract risk, a separate liquidation risk, and dependency on every underlying protocol continuing to function correctly at the same time.
A stated yield that stacks several protocols together is not one risk multiplied by a nice number — it's several independent risks added together, any one of which failing can unwind the entire position. Advertised APY rarely communicates that stacking clearly.
A framework for evaluating any yield opportunity
-
01
Identify the source.
Interest from real borrowers, fees from real traders, network security rewards, or new token emissions the protocol is printing to attract deposits — these are not equally durable.
-
02
Check whether the yield is subsidized.
A protocol paying rewards in its own newly minted token is effectively diluting existing holders to pay new depositors — a rate that's often unsustainable once the incentive program ends or the token's price falls.
-
03
Count the smart contracts involved.
Every additional protocol in a strategy is an additional point of failure, whether from a bug, an exploit, or a governance decision you didn't anticipate.
-
04
Understand your exit.
Lock-up periods, withdrawal queues, and impermanent loss can all mean the value you can actually withdraw differs meaningfully from the value shown on a dashboard.
Why "too good to be true" is a genuinely useful filter here
DeFi yields significantly above what comparable lending or fee-generating activity elsewhere in the market would suggest are almost always compensating for a specific, identifiable risk — token emission dilution, unaudited or experimental code, or unsustainable incentive spending meant to bootstrap early activity. That doesn't make every high yield a trap, but it does mean the size of the number alone tells you nothing useful; only the source does.