How DeFi Interest Rates Are Set: The Full Mechanics
By Jorge Rodriguez — DeFi Protocols
How the utilization rate model determines both borrow and supply APY in real time
Why rates spike sharply after the kink point and what that signals about pool liquidity
A framework for reading utilization as a forward indicator of DeFi yield changes
Introduction
A single lending pool can move from 4% to 45% APY in under an hour without any human making a decision. If you have ever watched a DeFi yield change mid-session and wondered why, the answer lives in one formula. Understanding how DeFi interest rates are set is not optional for anyone deploying capital in lending protocols. This article breaks down the full mechanics: the **utilization rate** formula that drives everything, the kinked curve that creates sudden rate spikes, the reserve factor math behind supply APY, how Aave and Compound implement these models differently, and what all of it means when you are deciding where to put capital. The same model runs on Ethereum, Base, and Solana-native protocols like Kamino and Marginfi. The parameters differ; the core mechanics do not. If you want to track these mechanics across 30+ protocols without watching dashboards manually, the [Lince Yield Tracker](https://yields.lince.finance/tracker) surfaces utilization rates, borrow APYs, and supply APYs in one place.
What the Utilization Rate Actually Measures
**The Core Formula** The **utilization rate** (U) is the single input that drives all interest rate calculations in algorithmic lending. The formula is: ``` Utilization Rate (U) = Total Borrows ÷ Total Deposits ``` If $80 million is borrowed from a $100 million pool, utilization is 80%. If a pool holds $200 million in deposits but only $60 million has been borrowed, utilization is 30%. Every other rate number on a protocol dashboard is downstream of this one figure. At 0% utilization, the pool is full of idle capital earning nothing for lenders. At 100%, every deposited dollar has been lent out and no withdrawals are possible. Both extremes are problematic, which is why protocols are designed to keep U in a target range. **Why Protocols Care About Utilization** At very high utilization, lenders face a withdrawal problem. The **liquidity buffer** (the unborrowed portion of the pool) shrinks, and lenders who want to exit may have to wait for borrowers to repay first. In extreme cases, withdrawal can be temporarily blocked. The protocol's job is to balance incentives so utilization stays in a range that keeps capital working efficiently while preserving enough liquidity for normal operations. For most assets, a healthy range sits between 70% and 90%, varying by asset risk profile and borrow demand patterns. The utilization rate is not just a dashboard number. It is the upstream input to every other rate calculation in the protocol. Borrow rates, supply APYs, and the incentive for new capital to enter all flow from this one figure.
The Kinked Interest Rate Curve
 **Two Slopes, One Kink** Rather than a single straight line, the interest rate curve in most lending protocols is piecewise: two different slopes joined at a threshold called the **kink point**, also referred to as **optimal utilization**. Below the kink, the rate curve rises gently. Borrowing is relatively cheap, which encourages demand, and the pool has enough liquidity to absorb withdrawals. Above the kink, the curve steepens sharply. Borrow rates can jump from 8% to 30% or higher in the range between 85% and 95% utilization. This is deliberate design, not an anomaly. Once utilization climbs past the kink point, the protocol needs to shock the system back toward balance: • Discouraging new borrowers from entering by making rates expensive • Incentivizing existing borrowers to repay sooner • Attracting new deposits by making supply APY temporarily very attractive **What the Reserve Factor Does** There is always a spread between what borrowers pay and what lenders earn. That spread is captured by the **reserve factor**, a percentage of borrow interest routed to the **protocol treasury** rather than distributed to lenders. A worked example makes this concrete: ``` Borrow Rate = 10% Utilization Rate = 80% Reserve Factor = 20% Supply APY = 10% × 0.80 × (1 − 0.20) = 6.4% ``` Lenders always earn less than borrowers pay. The reserve factor is the protocol's operating margin, and it varies by asset and protocol. For a deeper breakdown of how borrow and supply rates diverge, see [Supply vs Borrow APY: What the Difference Means](/blog/defi-protocols/supply-borrow-apy-defi-explained).
Aave vs Compound: Two Implementations of the Same Model
**Aave V3: Per-Asset Optimal Utilization** Aave V3 applies the kinked rate model with per-asset parameter sets. Each asset gets its own **optimal utilization** target, its own **slope1** (the gentle gradient below the kink), and its own **slope2** (the steep gradient above it). ETH, USDC, and a volatile altcoin will each have different kink thresholds because their borrow demand profiles differ significantly. Aave also separates variable and stable **borrow rates**. Only the variable rate responds to utilization in real time. Stable borrow rates are set at origination and can be rebalanced by the protocol, but they do not fluctuate block by block. The full parameter set for each Aave V3 market is publicly available in the [Aave risk documentation](https://docs.aave.com/risk/liquidity-risk/borrow-interest-rate), covering how slope1, slope2, and the base variable borrow rate are defined. **Compound III: The Jump Rate Model** Compound uses what its documentation calls the **jump rate model**. The structure is similar: a low slope up to a **utilization kink**, then a steep jump above it. Compound's kink parameter is explicitly documented in its contracts, making the exact threshold visible to anyone reading the [Compound interest rate docs](https://docs.compound.finance/interest-rates/). **Governance parameters** (slope values, kink thresholds, reserve factors) are voted on by token holders for both protocols. Once set, the smart contract executes them autonomously without any human override in real time. **What Both Share** Both protocols recalculate rates on every block. Every transaction that changes the borrow or deposit balance triggers a fresh calculation. This is fundamentally different from scheduled policy decisions: • No calendar, no committee, no discretion • The formula runs on every block • Any participant can verify the calculation on-chain • Governance sets parameters but cannot override the formula in the moment The result is a system where rates respond to supply and demand at the speed of the blockchain, not the speed of a meeting room.
How Borrow Demand Drives Supply APY
 **The Transmission Chain** The direction of causation matters. **Borrow demand** comes first; **supply APY** follows. More borrowers means higher utilization, which means a higher borrow rate, which means more interest flowing to lenders after the reserve factor is applied. This **rate transmission** chain has direct implications for lenders. Yield is not controlled by the lender; it is earned based on whatever the current utilization and borrow rate produce. When borrow demand is strong, lenders benefit. When borrowers repay en masse, supply APY can drop sharply and immediately. **A Worked Example** Start with a pool at 60% utilization. Supply APY is moderate: useful for yield-seekers but not exceptional. A new leverage cycle arrives. Traders want to borrow to fund positions. Utilization climbs from 60% to 88%, crossing the kink point. The borrow rate jumps from 12% to 32%. ``` New Supply APY = 32% × 0.88 × 0.80 ≈ 22.5% ``` Lenders notice the elevated yield. New capital enters the pool, deposits rise, utilization falls back toward 75%, and rates normalize. This self-correction mechanism is how algorithmic lending balances itself. But the lag matters: the adjustment can take hours or days, depending on how quickly new capital responds. **The Flip Side: When Rates Collapse** A single large borrower repaying a substantial position can drop utilization from 85% to 40% in one transaction. The borrow rate falls immediately. Supply APY follows. Lenders who assumed a stable rate will see their yield cut significantly with no warning. There is no lock-up protection and no rate guarantee in variable-rate pools. The rate visible at deposit time reflects that moment's utilization, not the next block's.
Algorithmic Rate-Setting vs Central Bank Policy
 **How Central Banks Set Rates** Traditional monetary **policy rate** decisions follow a fixed calendar. The Federal Reserve's FOMC meets roughly eight times per year. Rate changes happen in discrete increments, typically 25 or 50 basis points. Between meetings, the policy rate is fixed regardless of intraday market conditions. Transmission lag from a policy decision to a retail loan rate can span weeks or months. Political constraints, inflation targets, and employment mandates all factor into the decision. Transparency exists in the form of published votes and forward guidance, but so does substantial opacity in committee deliberations. **How DeFi Protocols Set Rates** DeFi rate-setting is continuous, algorithmic, and permissionless. On Solana, rates update approximately every 400 milliseconds as new blocks confirm. On Ethereum, the cadence is roughly every 12 seconds. In either case, the contrast with scheduled policy meetings is stark: • No meeting calendar; rates update on every block • No basis-point increment constraint; the formula produces any value • No discretion; the smart contract executes the formula exactly • No transmission lag; the rate change is immediate and global Governance communities set the governance parameters through token votes, but once parameters are on-chain, no vote can override the formula in real time. **Where They Connect: The Carry Trade Floor** When traditional risk-free yields are elevated, rational capital migrates out of lower-APY DeFi stablecoin pools. As deposits exit, utilization rises, and DeFi rates climb in response. This **carry trade** arbitrage creates a soft floor: DeFi stablecoin lending rates rarely fall far below comparable traditional yields for extended periods when capital is free to move between markets. Academic research, including work from the Banque de France on interest rate transmission between DeFi and traditional finance, documents this connection. DeFi rates are not isolated from global monetary conditions; they are linked through the behavior of yield-seeking capital moving across markets.
DeFi Lending Rates on Solana
**Same Model, Different Parameters** Kamino, Marginfi, and Save all implement the same piecewise utilization-based model used by Aave and Compound. The core mechanic is identical: utilization drives the borrow rate, the borrow rate minus the reserve factor spread produces supply APY. What differs are the parameters. Solana lending protocols often calibrate tighter kink thresholds for volatile assets and broader ones for stablecoins. Solana's low transaction fees also mean utilization arbitrage resolves faster. When a rate imbalance exists between pools, traders can act on it quickly, compressing the window where rates remain mispriced. Kamino's isolated lending markets separate rate calculations by asset. A spike in utilization for one market does not affect the rate curve of another asset in the same protocol. This design prevents contagion between asset classes but also means each market can reach extreme utilization levels independently. High utilization environments on Solana increase pressure on leveraged borrowers. For a closer look at the risks that come with this territory, see [DeFi Yield Risks Explained](/blog/risk-management/defi-yield-risks-explained). **Point Incentives and Emissions** A pool advertising a high APY on a Solana protocol may not be generating that yield entirely from lending interest. **Incentive emissions** in the form of protocol token rewards are frequently layered on top of the **base lending rate**. A pool showing 20% APY might have a 5% base lending rate and 15% in token rewards. Those rewards are subject to: • Token price volatility • Vesting and unlock schedules • Governance decisions about program continuation • End dates when the incentive program expires When an incentive program ends or the token price falls, the visible APY can drop sharply while the underlying lending rate remains unchanged. Always verify what percentage of visible APY comes from the actual lending rate before committing capital.
Reading Rates as a Yield Investor
**What Utilization Tells You Before APY Changes** Utilization is a leading indicator. Supply APY is the lagging result. Tracking utilization over time reveals the trend before the rate move happens. Rising utilization signals that borrow demand is increasing and rate increases are coming. If utilization is trending from 65% toward 85% over several days, the borrow rate and supply APY will follow. Watching this trend gives yield investors a head start on repositioning before the rate spikes. Utilization above 90% is a liquidity warning. The pool is nearly fully deployed. Withdrawals may face delays or be blocked entirely until borrowers repay. High APY at extreme utilization is real, but so is the exit constraint. **The Sweet Spot Zone** The most favorable position for yield investors is typically the 70% to 85% utilization range. In this zone: • The borrow rate is elevated enough to produce a strong supply APY • The liquidity buffer is large enough to allow normal withdrawals • The pool is not yet in liquidity-stress territory Below 60% utilization, supply APY is typically low. The pool has excess capital relative to borrow demand, and yield-seekers compete for a thin slice of modest borrow interest. Above 90%, the APY may look exceptional, but liquidity risk is real. The elevated rate reflects stress in the pool, not a durable opportunity. **Sustainable Yield vs Incentive Inflation** The **base lending rate** is sustainable as long as genuine borrowers exist. It is tied to real economic demand: traders leveraging positions, protocols borrowing for liquidity operations, yield strategists running arbitrage. Boosted APY from emissions depends on protocol token price, reward schedule duration, and governance decisions about continuation. It can disappear quickly and without warning. The rule of thumb: always check what percentage of visible APY comes from the actual lending rate. [Lince Yield Tracker](https://yields.lince.finance/tracker/solana/category/lending) surfaces base vs. incentive APY breakdowns across Solana lending protocols, making this comparison straightforward without manually reading each protocol's dashboard.
Common Misconceptions
**"Higher APY always means better opportunity"** APY above the kink point is often temporary and comes with liquidity risk. A 35% lending APY at 95% utilization means the pool is nearly fully deployed. Withdrawal may require waiting for borrowers to repay, which can take hours or longer in volatile market conditions. High APY at extreme utilization signals stress, not safety. The opportunity is real but so is the exit constraint. **"DeFi rates are random or arbitrary"** Every rate follows a deterministic formula that is fully visible on-chain. The perception of randomness comes from not tracking utilization or understanding the kink curve. There is no hidden decision-maker, no discretion, and no opaque committee. The formula runs on every block and anyone can verify it. **"A 20% lending APY equals a 20% savings account"** Savings accounts offer principal guarantees, fixed rates, and deposit insurance. DeFi lending pools carry smart contract risk, floating rates that update every block, no deposit insurance, and collateral-dependent solvency. The yield number may look similar; the risk profile is structurally different. **"Rates are stable once you deposit"** The rate shown at deposit time is not locked. A large borrower repaying in a single transaction can cut your supply APY by half before the next block confirms. Variable-rate pools do not offer rate stability; they offer participation in whatever rate the current utilization produces. Planning around a fixed yield number in a variable-rate pool is a common and costly mistake.
FAQ
### What is the utilization rate in DeFi lending? The utilization rate is the ratio of borrowed assets to total deposited assets in a lending pool. If $70 million is borrowed from a $100 million pool, utilization is 70%. This number is the single input that drives all interest rate calculations in algorithmic lending protocols. ### Why do DeFi interest rates spike so suddenly? The kinked rate curve is designed to spike above a threshold called the kink point, usually around 80% to 90% utilization. This jump discourages further borrowing and attracts new lenders by making supply APY temporarily very high. It is a deliberate self-correction mechanism built into the protocol's interest rate formula. ### How does borrow demand affect my lending APY? Higher borrow demand raises the utilization rate, which raises the borrow rate, which raises your supply APY through the formula: Supply APY equals Borrow Rate multiplied by Utilization multiplied by (1 minus Reserve Factor). Borrow demand is the upstream driver of lender yield. When borrowers repay, demand falls and so does your APY. ### What is the kink point in an interest rate curve? The kink point (also called optimal utilization) is the utilization threshold where the interest rate slope steepens sharply. Below it, rates rise gradually. Above it, rates jump steeply to deter more borrowing and attract new capital into the pool. This threshold is set by governance and varies by asset and protocol. ### How do Aave and Compound set their interest rates? Both use piecewise-linear rate curves with a kink point. Aave V3 assigns per-asset optimal utilization levels with two distinct slopes calibrated to each asset's risk profile. Compound uses a similar jump rate model with an explicit kink parameter in its contract documentation. Governance communities set the parameters; the smart contract executes them automatically on every block. ### Can DeFi lending yields drop to zero? Theoretically yes, if utilization approaches 0% and all deposited assets sit idle with no borrowers. In practice this is rare for liquid markets, but it can happen when a protocol loses borrow demand after a market cycle ends, an incentive program expires, or a competing protocol draws borrowers away with better terms. ### Are DeFi interest rates correlated with central bank rates? Yes, indirectly. When traditional risk-free yields are elevated, rational capital migrates out of lower-APY DeFi stablecoin pools, reducing supply, raising utilization, and pushing DeFi rates up. This carry-trade arbitrage creates a soft floor where DeFi stablecoin lending rates rarely stay far below comparable traditional yields for extended periods when capital is free to move. ### What is a reserve factor and how does it affect my yield? The reserve factor is the percentage of borrow interest routed to the protocol treasury rather than distributed to lenders. If the borrow rate is 10%, utilization is 80%, and the reserve factor is 20%, supply APY works out to 10% multiplied by 0.80 multiplied by 0.80, which equals 6.4%. The reserve factor is the protocol's operating margin and it reduces what lenders earn relative to what borrowers pay.
Conclusion
The utilization rate is the foundation of every DeFi lending rate. Understand that one formula and you can trace every other number on a protocol dashboard back to its source: borrow rates, supply APYs, liquidity risk at high utilization, and the brief rate spikes that appear when the kink point is crossed. The model is the same whether you are looking at Aave, Compound, Kamino, or Marginfi. Parameters differ; the mechanics are universal. Base lending rates are sustainable as long as real borrow demand exists. Emissions-inflated APYs are not. The practical skill is learning to read utilization as a forward indicator rather than chasing the APY number that is already visible. Rising utilization signals incoming rate increases. Utilization above 90% signals liquidity stress. The 70% to 85% range is where the yield-to-liquidity tradeoff tends to be most favorable. If you would rather have this analysis done automatically across vetted Solana protocols, [Lince Smart Vaults](https://yields.lince.finance/vaults) routes capital based on utilization dynamics, risk parameters, and base yield data, without requiring you to monitor each protocol manually.