Supply and Borrow APY in DeFi: Why They're Different
By Jorge Rodriguez — DeFi Protocols
The utilization rate mechanic that drives every lending protocol's interest rates
Why borrow APY is structurally always higher than supply APY, and by exactly how much
How to use supply and borrow rate dynamics to time lending entries and avoid yield compression
The Gap Between Supply and Borrow APY
You supply USDC on Kamino at 8% APY. You check the borrow rate on the same pool and it reads 14%. Same asset, same pool, same block. You are looking at a structural spread that exists on every DeFi lending protocol, from Aave on Ethereum to Kamino and Marginfi on Solana. That gap is not arbitrary. It is baked into how lending protocols work, and once you understand the mechanics behind it, the numbers stop being confusing and start being useful signals for timing your positions. Three forces create and maintain the gap: the **utilization rate**, the **reserve factor**, and the structural dilution of interest across a shared pool. This article breaks down each one, shows how they interact with real formulas and concrete examples, and translates that understanding into actionable positioning decisions. This is not a primer on DeFi lending from scratch. It assumes you have already supplied collateral or borrowed against it. If you track lending rates on the [Lince Yield Tracker](https://yields.lince.finance/tracker/solana/category/lending), you have seen these numbers move in real time. This guide explains what is actually moving them.
How DeFi Interest Rate Models Work
Every major lending protocol uses an algorithmic interest rate model. Rates are not set manually by any person or committee. They are computed automatically at every block based on a single input: the fraction of supplied capital that is currently borrowed. **The Two-Slope Model** The dominant design is the **kinked interest rate model**, sometimes called the two-slope model. It operates in two distinct phases: • Below a threshold called the **kink point**, borrow rates increase gradually as utilization rises. The protocol is operating normally, with adequate liquidity available for supplier withdrawals. • Above the kink point, borrow rates climb steeply. A gradual slope becomes a near-vertical wall. For most stablecoin pools, the kink point sits between 80% and 90% utilization. Below that threshold, the protocol encourages borrowing by keeping costs moderate. Above it, a steep rate increase makes additional borrowing prohibitively expensive and draws in new supply capital. This design exists for a critical purpose. Without the rate cliff above the kink, a pool could reach near-100% utilization, leaving suppliers unable to withdraw. The steep slope is a market-based circuit breaker: it raises borrow costs high enough to attract new suppliers and discourage fresh borrowing, restoring the liquidity buffer before it disappears entirely. **The Rate Formula** The standard kinked model computes borrow rates using these equations: ``` When U < Kink: Borrow Rate = Base Rate + (U / Kink) x Slope1 When U > Kink: Borrow Rate = Base Rate + Slope1 + ((U - Kink) / (1 - Kink)) x Slope2 ``` Below the kink, rates scale linearly with how full the pool is. Above the kink, a steeper second slope kicks in and can drive rates several times higher per percentage point of additional utilization. Slope2 is typically set at five to fifteen times the steepness of Slope1, which is what creates the dramatic rate spikes users encounter during high-demand periods. The specific parameters (Base Rate, Slope1, Slope2, and the kink threshold) are set by governance and differ across protocols and asset types. Understanding which parameters a protocol uses tells you a lot about how risk is managed. For a deeper look at protocol risk design, see our [DeFi risk framework guide](https://yields.lince.finance/blog/risk-management/defi-risk-framework).  **The Governance Layer** Rate parameters are not fixed permanently. Protocol governance can vote to adjust the kink point, modify slope steepness, or change reserve factors in response to market conditions. Aave regularly publishes risk parameter update proposals. Kamino and Marginfi have similar governance mechanisms. This matters for yield planning. A rate model that works well at current market conditions may behave differently after a governance update. Tracking parameter changes alongside raw APY figures is part of understanding what your position is actually exposed to.
The Utilization Rate: The Core Driver
If there is one number that drives everything in lending protocol rate dynamics, it is the **utilization rate**. The formula is: ``` U = Total Borrowed / Total Supplied ``` If $80 million is borrowed from a pool with $100 million supplied, utilization is 80%. If borrowers repay $10 million, utilization drops to 70% and rates fall. If new borrowers push total borrowed capital to $90 million, utilization rises to 90% and rates spike. **Three Utilization States** How a pool behaves across the utilization range is predictable and consistent across protocols: • Low utilization (below 40%): Liquidity is abundant. Borrow costs are cheap and supply APY is low. This zone is attractive for borrowers but unrewarding for suppliers, who see most of their deposited capital sitting idle. • Moderate utilization (60% to 80%): Rates are climbing and supply APY becomes genuinely interesting. This is the productive zone for suppliers, offering elevated rates without the withdrawal risk that comes with a near-full pool. • High utilization (above 85%): Borrow rates spike past the kink. Supply APY rises as well, but not proportionally. Available withdrawal liquidity thins out significantly. This zone carries meaningful risk for both sides of the market. **Utilization Changes Every Block** Utilization is not a static number. It shifts with every transaction: new borrows push it up, repayments push it down, supply deposits push it down, and withdrawals push it up. On Solana, where blocks finalize in under a second, the rate you see when you enter a position is not guaranteed to be the rate five minutes later. This is why DeFi lending rates are variable by design. There is no lock-in mechanism in most lending protocols. Every block recalculates the rate, and all open positions pay the current rate going forward. A borrower who opened a position at 12% APY can find themselves paying 30% or more if a large capital event shifts utilization across the kink threshold. When utilization climbs to extreme levels and liquidity gets thin, the dynamics can become self-reinforcing. Borrowers who are underwater delay repayment, suppliers who want to exit find the pool tapped out, and a liquidity crunch forms. For a detailed analysis of how this dynamic feeds into bad debt formation across lending protocols, see our guide on [bad debt in DeFi lending](https://yields.lince.finance/blog/risk-management/bad-debt-defi-lending).
Why Borrow APY Is Always Higher Than Supply APY
This is the core question. There is a precise, mathematical answer that has nothing to do with protocol manipulation or hidden spreads. Three structural mechanics guarantee the gap, and together they make it permanent. **The Reserve Factor** Every lending protocol takes a cut of borrow interest before distributing it to suppliers. This percentage is called the **reserve factor**, and it flows to the protocol treasury or insurance fund. A reserve factor of 20% means that for every $1.00 of interest paid by borrowers, $0.20 goes to the protocol and only $0.80 reaches suppliers. Reserve factors typically range from 10% to 30%, depending on asset risk and protocol design. Conservative stablecoin markets run lower reserve factors. Assets with higher volatility, less liquidity, or more complex liquidation risk carry higher ones. This single mechanism guarantees supply APY is always less than borrow APY. Even at 100% utilization with every dollar in the pool deployed, suppliers receive at most 70% to 90% of what borrowers pay. **The Utilization Discount** Beyond the reserve factor, supply APY is further discounted because only the borrowed fraction of the pool generates interest. Capital sitting idle in an underutilized pool earns nothing. The complete relationship is: ``` Supply APY = Borrow APY x Utilization Rate x (1 - Reserve Factor) ``` A concrete example: borrow APY is 15%, utilization is 80%, and reserve factor is 20%. ``` Supply APY = 15% x 0.80 x 0.80 = 9.6% ``` The borrower pays 15%. The supplier earns 9.6%. The gap is 5.4 percentage points, explained entirely by the reserve factor and the utilization discount. No hidden spread, no intermediary capturing value. Just the math of how interest is distributed across a shared pool. **Structural Dilution** There is a third factor worth naming. By design, every borrower must supply more collateral than they borrow. Total supplied capital always exceeds total borrowed capital. This means borrow interest is always spread across a larger pool of suppliers than the pool of borrowers, diluting per-unit yield further even after the reserve factor is applied. **The Core Answer** Two reasons keep borrow APY structurally above supply APY: the reserve factor and dilution. Protocols take a percentage of borrow interest (typically 10% to 30%) for their treasury. The rest is split across all suppliers, who collectively hold more capital than borrowers do. Supply APY is always a fraction of borrow APY, regardless of how high rates climb. Token incentive emissions can temporarily push the reported total supply APY above borrow APY, but this is separate from interest income. Those rewards are not generated by the lending mechanics and are not sustainable once emissions end. To understand how reserve factors function as a hidden cost when comparing protocols, see our guide on [hidden fees in DeFi](https://yields.lince.finance/blog/yield-strategies/hidden-fees-defi). For a broader look at the risks that come with being on the borrow side of a lending market, see our [DeFi yield risks guide](https://yields.lince.finance/blog/risk-management/defi-yield-risks-explained).
How Utilization Rate Affects Your Yield
The mechanics become tangible when you watch them play out in a live position. Here is what rate dynamics look like in practice for each side of the market. **The Supply Side: Rate Compression** Elevated supply APYs attract capital. This is a consistent pattern across every lending protocol. When utilization climbs and rates spike, the opportunity becomes visible. New suppliers enter the pool. As total supplied capital grows, utilization drops and borrow interest is spread across more capital. APY compresses. The window closes. The cycle repeats. If you enter as a supplier when utilization is already near its peak (88% to 92%), you are likely entering near the compression phase rather than the opportunity phase. The optimal supply entry window is when utilization sits between 70% and 85%: rates are elevated, but the inflow of competing capital has not yet begun. After a utilization spike drives visible APY highs, watch for compression signals: rising total deposits, declining utilization readings, or a steady fall in reported supply APY over consecutive hours. These indicate the rate cycle has peaked and the window is closing. **The Borrow Side: The Rate Cliff** Borrowers face a different risk. A position opened at a comfortable borrow rate can become significantly more expensive if utilization crosses the kink. Consider a collateral loop opened at 12% borrow APY with pool utilization at 75%, and a kink set at 80%. A surge in borrow demand pushes utilization to 88%. The steeper Slope2 kicks in and borrow APY jumps to 28%. A carry strategy that worked at 12% is now deeply loss-making at 28% if supply APY has not risen proportionally. This rate spike risk is not rare or hypothetical. It occurs during market volatility events when traders rush to borrow stablecoins for leverage, or during protocol incentive periods that draw in large borrowers quickly. Positions opened during quiet periods carry the hidden risk of a rate cliff during active market conditions. For a full breakdown of how rapid rate increases feed into liquidation mechanics on Solana lending protocols, see our [DeFi liquidations guide](https://yields.lince.finance/blog/risk-management/defi-liquidations-solana). **The Rate Cycle Pattern** High rates attract supply. New supply compresses rates. Some suppliers exit. Utilization tightens. Rates rise again. Knowing where you are in this cycle at any moment is as important as knowing the current APY number. The rate itself is a lagging indicator. Utilization is the leading one.
Supply vs Borrow APY Across Major Protocols
The kinked rate model is universal, but each protocol configures it with different parameters. Those parameters drive meaningfully different rate behavior at the same utilization level.  | Protocol | Asset | Typical Supply APY | Typical Borrow APY | Kink Point | Reserve Factor | |---|---|---|---|---|---| | Kamino | USDC | 5-12% | 10-22% | 85-90% | 10-20% | | Marginfi | USDC | 5-11% | 10-20% | 80-90% | 10-20% | | Drift | USDC | 4-10% | 9-18% | 80-85% | 10-20% | | Aave (ETH) | USDC | 4-9% | 8-16% | 90% | ~10% | All figures are illustrative ranges based on typical conditions. Actual rates fluctuate with utilization and governance changes. **Kamino (Solana)** Kamino is the largest lending protocol on Solana by TVL. Its rate model uses kink points typically set at 85% to 90% for stablecoin pools. For volatile assets like SOL or liquid staking tokens, kinks are usually set lower to protect liquidity during volatility events. One critical nuance: the supply APY displayed on Kamino often bundles base interest with token incentive rewards. The base APY from borrow interest alone may be considerably lower than the headline figure. When comparing Kamino against other protocols, always isolate the base interest component to make a fair comparison. Reserve factors on Kamino vary by asset, typically ranging from 10% to 20% for stablecoins. **Marginfi (Solana)** Marginfi runs a competitive lending market with similar structural mechanics. One practical advantage: Marginfi's interface displays current utilization alongside rates, which makes it easier to assess where a pool sits relative to the kink before entering a position. This transparency is useful for applying the rate dynamics covered in this article to live decision-making. **Drift (Solana)** Drift is primarily a perpetual futures exchange, but it includes a spot borrowing and lending market for collateral management. Because the primary borrower base on Drift consists of leveraged traders managing perp collateral rather than yield seekers, utilization patterns can differ from pure lending protocols. Rates on Drift's lending markets follow the same kinked model, but spike behavior during high-volatility trading sessions can be more sudden and sharp. Positions opened on Drift's spot lending market carry rate risk that correlates with perp trading activity, not just lending market conditions. **Aave (Ethereum and Multichain)** Aave is the reference standard for rate model transparency. Reserve factors, kink thresholds, and risk parameters are published in Aave's governance documentation, making it the most accessible protocol for studying rate mechanics in detail. On Aave's USDC pools, reserve factors have historically been around 10%, with kink points near 90%. Aave V3 introduced efficiency mode (eMode), which allows correlated asset borrowing at higher LTV ratios. eMode affects rate dynamics because higher capital efficiency changes the effective utilization math for positions using correlated collateral and collateral pairs. Aave serves as a useful benchmark. When evaluating Solana protocol parameters, comparing kink thresholds and reserve factors against Aave's published governance data provides a solid reference point for whether a protocol's risk configuration is conservative, standard, or aggressive.
Strategic Implications
The mechanics of supply and borrow APY are not just academic. They translate directly into better positioning decisions.  **For Suppliers** The optimal entry window is when utilization sits between 70% and 85%. Rates are elevated, but the capital inflow triggered by those elevated rates has not yet begun compressing them. You are positioned at the front half of the rate cycle rather than the back. After a utilization spike drives APY to visible highs, watch for compression signals: rising total deposits, a steady decline in utilization from peak levels, or a falling APY trend over consecutive hours. These indicate the compression phase has started and the marginal benefit of entering is declining. When comparing protocols, do not sort by gross APY alone. A protocol with a 10% reserve factor passes more borrow interest to you than a protocol with a 25% reserve factor, even if the displayed supply APY numbers look similar. Understand what fraction of borrow revenue is actually reaching you before committing capital. **For Borrowers** Identify the kink point of every pool you borrow from before entering. If current utilization is within 10 percentage points of that threshold, you are exposed to a potential rate cliff. Model your position at kink-level borrow rates, not current rates. If the position breaks at kink utilization, you are carrying more rate risk than your initial entry analysis shows. For large or long-term borrow positions, check whether fixed-rate alternatives exist. Protocols built around PT/YT mechanics or fixed-term lending can offer rate certainty that variable markets cannot provide. Our guide on [leveraged yield farming risks](https://yields.lince.finance/blog/risk-management/leveraged-yield-farming-risks) covers the full spectrum of market risk factors that interact with rate spikes in leveraged borrow positions. **For Loopers** Looping strategies depend on positive carry: the supply APY you earn must exceed the borrow APY you pay, net of your leverage ratio. ``` Net Carry = Supply APY - (Borrow APY x Leverage Ratio) ``` At normal utilization, carry is positive. Near the kink, borrow APY spikes faster than supply APY rises. Carry can turn negative before you have time to unwind, especially on Solana where liquidations execute in seconds. The full mechanics of carry risk and how to model it are covered in our guide on [leveraged yield looping in DeFi](https://yields.lince.finance/blog/yield-strategies/leveraged-yield-looping-defi-explained). To stay ahead of utilization shifts across Solana lending markets, the [Lince Yield Tracker](https://yields.lince.finance/tracker/solana/category/lending) lets you monitor live lending rates across Kamino, Marginfi, and other protocols in one place.
FAQ
### What is the difference between supply APY and borrow APY in DeFi? Supply APY is the annualized yield that lenders earn for depositing capital into a lending pool. Borrow APY is the annualized cost that borrowers pay to access that capital. The gap between them exists because protocols take a percentage of borrow interest as a reserve factor for their treasury, and because that interest is spread across all suppliers rather than just those whose capital is currently deployed as loans. ### Why does my supply APY keep changing? Supply APY is variable and recalculates at every block based on current utilization. When new suppliers enter the pool, utilization drops and supply APY falls. When borrowers enter or suppliers withdraw, utilization rises and supply APY increases. On Solana, where blocks finalize in under a second, rates can shift noticeably within minutes of a large capital event. ### What is a good utilization rate for a DeFi lending pool? For most stablecoin pools, utilization between 60% and 85% represents a healthy balance. Supply APY is meaningful, borrow rates are not prohibitive, and enough liquidity buffer exists for withdrawals. Above 85% to 90%, most protocols cross the kink point and borrow rates spike aggressively. Below 50%, supply APY tends to be low and the pool is largely underutilized. ### Can supply APY ever exceed borrow APY in DeFi? No, not from base interest income alone. The reserve factor and dilution mechanics make it mathematically impossible for supply APY to match borrow APY through interest mechanics. However, token incentive emissions layered on top of base supply APY can push the reported total supply APY temporarily above the borrow rate. These incentive yields are separate from interest income and are not sustainable once emissions end. Our guide on [yield sustainability in DeFi](https://yields.lince.finance/blog/yield-strategies/yield-sustainability-defi) covers how to evaluate whether a displayed APY is organic or incentive-driven. ### What happens when utilization exceeds the kink point? When utilization crosses the kink point (typically 80% to 90% for stablecoin pools), the second, steeper rate slope activates. Borrow rates can jump from 15% to 40% or more within hours if demand persists. Supply APY rises as well, but more slowly. For borrowers, this creates a rate cliff risk. For suppliers, it creates a short-term yield opportunity paired with a withdrawal liquidity risk if utilization approaches 100%. ### How do reserve factors differ across protocols? Reserve factors vary by protocol and by asset. Conservative stablecoin markets typically run reserve factors of 10% to 20%. Riskier or more volatile assets often carry reserve factors of 20% to 40%. Aave publishes its reserve factors in governance documentation. Kamino and Marginfi display per-market parameters in their risk documentation. When comparing protocols, check the reserve factor for the specific asset and market you plan to use, not just the headline APY figure.
Conclusion
The gap between supply and borrow APY in DeFi is not a black box. It is the output of three interacting mechanics: the utilization rate, the reserve factor, and structural dilution. Each one is visible, measurable, and useful for making better positioning decisions. Suppliers who understand utilization cycles enter at better moments and avoid the compression trap. Borrowers who know the kink point of their pool model realistic worst-case costs rather than assuming current rates hold. Loopers who track the supply-borrow spread can see when carry is turning negative before it shows up as a realized loss. These mechanics apply across every protocol covered here, and across every lending protocol in DeFi. The parameters differ, but the structure does not. Use the [Lince Yield Tracker](https://yields.lince.finance/tracker/solana/category/lending) to monitor live utilization and rate data across Solana lending markets and put these mechanics to work in your next position.