The question that separates a durable yield from a fragile one is not how large the number is. It is what has to keep happening, week after week, for that number to keep getting paid. A yield backed by fees someone actually pays keeps working as long as that activity continues. A yield backed by a token subsidy stops the moment the subsidy is cut. Below is how to sort a given APY into the category it actually belongs to, using specific products as worked examples.
Start by asking what pays the yield, not what the rate is
Every yield has to be funded by something. Block3 Finance’s guide to evaluating crypto yield puts this plainly: yield either comes from real economic activity – fees paid by users, interest paid by borrowers, spreads captured on trading volume – or it comes from incentives such as token emissions and liquidity mining, which Block3 describes as transfers rather than revenue. Stripe’s own explainer for businesses and investors draws the same line, describing crypto yield as return generated by staking, lending, or providing liquidity – each a different mechanism with a different source of payment. Neither outlet disputes the other on this basic split; both use it as the first filter.
Four places yield actually originates
Interest paid by borrowers. Stripe describes lending yield as a matter of depositing crypto – often stablecoins – onto a platform that lends it to traders or institutions willing to pay for access to capital; the lender’s yield is simply the borrower’s interest payment, mediated either by a centralized platform vetting borrowers or a decentralized protocol doing so algorithmically through pooled smart contracts. Jurgis Pocius, writing in The Internet Economy on 6 February 2026, extends this logic to fintechs and stablecoin issuers more broadly: banks borrow from depositors at close to 0% and lend at 8%+, with interest income accounting for roughly 70% of bank revenue against 30% from non-interest sources such as fees, per The Internet Economy. Fintechs such as Revolut, which Pocius values at roughly $75B, can pass through more of that spread because they run leaner operations with no branch network – and stablecoin issuers push that same model further still, since they carry even lower overhead than a fintech.
Trading and liquidity-provision fees. A second real-activity source is fees generated directly by trading volume. Pocius points to Jupiter, where 75% of liquidity-provider fees come from position opening and closing, price impact, borrowing costs and trading fees, according to The Internet Economy’s reading of Jupiter’s own data. This yield rises when trading activity and leverage increase and compresses when volumes fall – it tracks market activity, not a fixed schedule.
Delta-neutral funding-rate strategies. A third source sits between lending and active trading. Ethena’s sUSDe, as described by The Internet Economy, is a synthetic stablecoin backed by a delta-neutral structure: the protocol holds collateral such as ETH or stETH while running an offsetting short position in derivatives markets, and the yield comes from funding rates paid on those perpetual futures rather than from a redemption mechanism against fiat reserves. Resolv Labs’ RLP token works on a related principle – it functions as a risk buffer for Resolv’s USR stablecoin, absorbing losses that arise from the delta-neutral hedge, and in exchange RLP holders capture higher returns sourced from staking rewards and funding fees, per The Internet Economy. This is yield paid for managing basis risk and hedge execution, and the outlet notes it carries the added sensitivity that funding rates themselves are volatile.
Borrowing tied to tax avoidance rather than trading. A fourth source is less visible on the surface. Gauntlet’s USDC Prime vault on Base allocates $326.87M of its $327.68M in assets to a single cbBTC/USDC pool, lending USDC against Coinbase’s wrapped Bitcoin product at roughly 4%, according to The Internet Economy. Pocius argues that because cbBTC is specifically Coinbase’s wrapped Bitcoin, a meaningful share of the borrowing behind that pool likely reflects holders borrowing against Bitcoin rather than selling it, to avoid triggering a capital-gains tax event – a strategy The Internet Economy compares to wealthy individuals borrowing against equities or property in traditional finance. If that reading is right, the yield paid to lenders in that pool is funded less by speculative leverage demand and more by holders’ reluctance to realize a taxable gain, a demand pattern that can persist through calm markets precisely because it isn’t driven by trading conviction.
A checklist for classifying a specific product
Once a yield is placed into one of the categories above, Block3 Finance’s framework and Stripe’s own risk section converge on the same set of follow-up questions, each testing whether the yield survives stress rather than just describing it in calm conditions:
- Liquidity risk – Block3 notes this risk rarely appears in marketing and instead shows up as withdrawal queues or slippage during volatility; the real test is what happens on exit, not entry.
- Smart-contract risk – Block3 argues an audit reduces risk without eliminating it, since composability across protocols creates dependencies no single audit can model.
- Counterparty risk – Block3 argues that DeFi does not remove counterparties, it layers them: protocols depend on oracles, oracles on data providers, bridges on validators, and a failure anywhere in that chain can unwind a yield strategy faster than expected.
- Volatility and liquidation risk – Block3 notes that yield is usually quoted in annualized terms while risk materializes intraday; a strategy earning steady returns can still lose value if the underlying asset drops sharply, and for leveraged or collateralized positions, liquidation risk is mechanical rather than theoretical.
- Incentive drift – Block3 warns that emissions schedules and governance parameters change quietly over the life of a position, so a yield that looked stable when first assessed can carry a materially different risk profile later without any single disclosed event marking the shift.
What this page does not tell you
This page cannot verify The Internet Economy’s on-chain figures – the DeFi TVL estimate, the Gauntlet vault balances, or the Jupiter fee breakdown – against a live block explorer or protocol dashboard; they are reported as of that newsletter’s 6 February 2026 publication with no independent confirmation available here, and the piece gives no separate on-chain snapshot timestamp beyond that publication date. Stripe’s figure that yield-generating assets made up roughly 8%-11% of global crypto market capitalization is dated only to “2025” with no stated methodology in the material reviewed, so a reader cannot reproduce or check it directly. Block3 Finance’s risk framework is published on a page that also markets the firm’s own advisory services, and it cites no data, audit or named case to support its categories – it should be read as one firm’s framework, not as verified fact. Finally, this page describes categories of yield source in general terms; it cannot tell a reader which category the specific product they are holding falls into. That requires reading the protocol’s own documentation and, where one exists, its audit.
Sources
Every fact above is attributed to one of these reports. Where they disagree, the article says so.
