Aave, Compound, Morpho, Spark Compared
The Decision You Are Making

You are choosing where to borrow stablecoins against crypto collateral. The rate you pay matters, but the rate is not fixed. It changes with pool utilization, and the mechanism that drives that change differs across protocols. A position that costs 6% today can cost 18% next week if utilization crosses a threshold you did not see coming.
Four lending protocols hold most DeFi borrow demand as of April 2026: Aave V3 at $19.4B TVL, Spark at $6.8B, Morpho Blue at $4.9B, and Compound V3 at $2.7B. Each uses a different interest rate model, collateral structure, and liquidation mechanism. Those differences determine what you pay to borrow and what happens when your position approaches liquidation.
This is not a choice between static rates. It is a choice between rate architectures. The utilization curve, the kink point, the liquidation threshold, and the liquidation penalty interact to produce both your ongoing cost and your downside risk. The protocol with the lowest rate today may have the highest spike risk tomorrow.
Utilization Curves: Why Rates Spike Suddenly

Most DeFi lending protocols set borrow rates algorithmically based on pool utilization: the percentage of deposited funds currently borrowed. Higher utilization produces higher rates. This is not a borrower-specific variable. It is a pool-level state that moves continuously as deposits enter, borrows open, and positions close.
Aave V3 and Compound V3 both use kinked utilization curves. Below an optimal utilization target (typically 80-90%), borrow rates increase gradually. Above that target, rates escalate sharply to discourage further borrowing and preserve withdrawal liquidity for lenders. This design is intentional. The kink protects liquidity, but it creates a discontinuity in your borrow cost.
Here is the mechanism in practice. Suppose Aave’s USDC pool sits at 75% utilization and you open a borrow at 7% APR. Demand increases. Utilization crosses 90%. The borrow rate jumps to 14% within hours. Your position did not change. The pool state changed. Your cost doubled.
Morpho Blue uses isolated markets rather than pooled liquidity. Each market is a single collateral-asset pair with its own utilization curve and parameters set by vault curators. This architecture concentrates borrow demand into specific markets, which typically produces higher utilization and higher rates than pooled models. Suppliers capture more of the spread because less capital sits idle, which is why Morpho vaults often pay more on USDC deposits than Aave or Compound.
Spark and Sky operate differently. Their rates are governance-set policy numbers, not market-clearing rates driven by utilization. Spark’s USDC rates track the DAI Savings Rate (DSR) because both are governed by the Sky ecosystem. As of September 16, 2026, the Sky Savings Rate was 3.60% APY. This is a floor set by DAO-managed reserves, not a function of borrower demand.
The durable takeaway: when evaluating borrow rates, check the current utilization level and the kink point for that asset. A rate quoted at 6% with 85% utilization and a 90% kink is more likely to spike than the same 6% rate at 70% utilization. On-chain data for utilization is live on DeFiLlama and each protocol’s dashboard.
Collateral Factors and Liquidation Thresholds

Borrow capacity is determined by the loan-to-value (LTV) ratio. Liquidation is triggered by the liquidation threshold (LT). These are not the same number, and the gap between them defines your risk buffer.
Aave V3 sets LTV at 80% for ETH collateral and liquidation threshold at 82.5%. That 2.5 percentage point spread is your buffer. If ETH drops 2.5% from your entry price, your position becomes eligible for liquidation. Over the past year, ETH’s daily price change exceeded 2.5% on 46% of days. This is not a comfortable margin.
Compound V3 uses two separate collateral factors: borrow collateral factors determine initial borrowing capacity, and liquidation collateral factors (set higher) determine the liquidation threshold. This creates a price buffer for all new positions. Suppose you deposit $1,000 of ETH. Compound’s 80% borrow factor lets you borrow $800 USDC. The liquidation factor is 90%, so liquidation triggers when your collateral value drops to $888 (a 11.2% decline from entry). That is a wider buffer than Aave’s 2.5%.
Morpho Blue’s isolated markets allow curators to set collateral factors per vault. Parameters vary. Some markets offer 85% LTV with tight liquidation thresholds. Others use 70% LTV with wider buffers. There is no protocol-wide standard. You must verify the specific vault’s parameters before depositing collateral.
Here is the liquidation interaction with rate spikes. You are borrowing at 85% LTV with a health factor near 1.2. Two events compress that health factor simultaneously: (1) collateral price drops 3%, and (2) utilization crosses the kink, spiking your borrow rate from 7% to 15%. The rate increase does not reduce health factor immediately, but it increases the accrued debt faster, narrowing the distance to liquidation threshold. If collateral continues declining, you hit liquidation sooner than you would have at the lower rate.
Liquidation Penalties: Protocol-Specific Structures
When your health factor falls below 1.0 on Aave, or your LTV exceeds the liquidation threshold on Compound, external liquidators can close part or all of your position. The protocol penalizes the borrower to compensate the liquidator and protect the pool. The penalty structure differs meaningfully across protocols.
Aave V3 uses a health-factor-based liquidation model. When health factor drops below 1.0, liquidators repay part of the debt in exchange for collateral. Aave’s documentation does not specify a fixed liquidation penalty percentage. The penalty is embedded in the liquidation bonus paid to liquidators, which varies by asset.
Aave V4 introduces dynamic liquidation bonuses that scale with the position’s health factor. Lower health factors offer higher bonuses, creating a Dutch-auction-style incentive. Liquidators prioritize the riskiest positions first. Instead of a fixed close factor, liquidators repay only enough debt to restore the position to a target health factor set at the protocol level. This prevents over-liquidation and leaves the borrower with remaining collateral.
Compound V3 applies a 5% penalty on the collateral when a position is liquidated. This penalty is protocol-absorbed, not paid directly to liquidators. Example: you deposit $1,000 of ETH and borrow $800 USDC. ETH drops to $880. Your LTV hits 90.9%, crossing the 90% liquidation threshold. A liquidator calls the absorb function. The protocol absorbs your $880 of ETH collateral. You lose 5% of the collateral value ($44) as the penalty, and the debt is cleared.
Morpho Blue’s liquidation mechanics are set per vault. Some vaults use Aave-style health factors, others use Compound-style LTV thresholds. The liquidation penalty is specified in the vault parameters. This is not discoverable at the protocol level. You must check the individual vault’s configuration on Morpho’s dashboard before opening a position.
The practical implication: Compound’s fixed 5% penalty is more predictable than Aave’s dynamic structure, but Aave V4’s partial liquidation mechanism may preserve more collateral in a volatile drawdown. Morpho’s vault-specific model requires the most diligence, because penalty structures are not standardized.
Reserve Factors and the Supply-Borrow Spread
Borrow rates set the cost you pay. Supply rates determine what lenders earn. The difference is the protocol’s spread, captured as the reserve factor. This spread funds protocol operations, DAO treasuries, and insurance reserves. It also explains why supply rates lag borrow rates even when utilization is high.
The formula: Supply Rate = Borrow Rate × Utilization × (1 – Reserve Factor).
Example: borrow rate is 10%, utilization is 50%, reserve factor is 20%. Supply rate = 10% × 50% × (1 – 20%) = 5% × 80% = 4%. Lenders earn 4%. Borrowers pay 10%. The 6 percentage point gap does not disappear. Half goes to reserve factor, half reflects idle capital in the pool.
Aave V3’s reserve factors vary by asset but typically run 10-20%. Compound V3’s reserve factors are similar. Morpho Blue’s isolated markets have curator-set reserve factors, which tend to be lower (5-15%) because less capital sits idle. Spark’s model is different: rates are subsidized by the DAO, so the reserve factor is effectively zero. The spread between deposit and borrow rates on Spark reflects governance policy, not pool mechanics.
Why this matters for borrowers: a lower reserve factor does not directly reduce your borrow cost, but it indicates a more efficient market. Morpho’s isolated markets and lower reserve factors mean utilization rises faster for the same level of borrow demand, which pushes borrow rates higher sooner. Aave’s pooled model spreads demand across more capital, which keeps rates more stable but also leaves more capital idle.
Rate Architecture By Protocol
Aave V3 uses a two-slope interest rate model with an optimal utilization point. Below the optimal point (typically 90% for stablecoins), borrow rates rise with a shallow slope. Above it, rates rise steeply. This kink is visible in the protocol’s governance documentation, which specifies slope parameters per asset. As of April 2026, Aave V3 holds $19.4B TVL and offers USDC borrow rates in the 5.5-8.5% range depending on utilization.
Compound V3 maintains $2.7B TVL and runs borrow rates slightly lower than Aave (3.5-5.8% USDC supply APY as of mid-2026). The protocol uses a similar kinked model but with different slope parameters. Compound’s liquidation structure (separate borrow and liquidation collateral factors) creates a wider safety buffer than Aave’s, which may justify the slightly lower rates as a risk-adjusted outcome.
Morpho Blue splits into two layers: the 650-line immutable primitive that handles isolated markets, and the vault layer where curators allocate deposits across those markets. Morpho’s rate discovery in version 2 moves away from formula-driven curves toward market-negotiated rates. This is a structural shift. For most of DeFi’s history, rates were driven by embedded formulas. Morpho V2 allows markets to discover rates rather than enforcing them algorithmically. This will produce more volatile rates but potentially more efficient capital allocation.
Spark’s rates are set by Sky governance, not utilization. The Sky Savings Rate was 3.60% APY as of September 16, 2026. Spark’s USDC rates track DSR closely. This is a policy floor, not a market rate. If you are willing to accept governance risk (the DAO could lower the rate at any time), Spark offers rate stability that utilization-based protocols cannot match.
Who Each Protocol Is Right For
Aave V3 is for borrowers who want deep liquidity, wide collateral options, and predictable liquidation mechanics. The health-factor model is well-documented. The kinked utilization curve is standard across assets. You pay a moderate spread for that stability. If you are borrowing large size or using less-common collateral, Aave’s $19.4B TVL provides the liquidity depth to handle your position without moving the market.
Compound V3 is for borrowers who want a wider liquidation buffer. The separate borrow and liquidation collateral factors create more downside protection than Aave’s tight LTV-to-threshold spread. If you are borrowing near your max LTV and expect volatility, Compound’s structure gives you more room before liquidation triggers. The tradeoff: slightly lower liquidity and fewer collateral options than Aave.
Morpho Blue is for borrowers who understand isolated market mechanics and are willing to verify vault parameters manually. Curators set collateral factors, reserve factors, and liquidation penalties per vault. This produces higher rates on the supply side, which often translates to higher borrow costs, but it also allows more efficient capital allocation. If you are comfortable reading contract parameters and monitoring vault-specific utilization, Morpho offers rate structures that pooled protocols cannot match.
Spark is for borrowers who want rate stability and are comfortable with governance risk. The Sky DAO sets rates by vote, not by utilization. This eliminates spike risk but introduces policy risk. If the DAO votes to lower the rate, your cost drops. If the DAO votes to raise it, your cost rises, regardless of market conditions. The 3.60% APY floor as of September 2026 is attractive, but it is not guaranteed.
My Recommendation
If you are borrowing stablecoins against ETH or BTC collateral and you plan to hold the position for more than 30 days, use Compound V3. The wider liquidation buffer (10+ percentage points between borrow LTV and liquidation threshold) matters more than the 50-100 basis point rate difference versus Aave. Rate spikes are temporary. Liquidation is permanent.
If you are borrowing against liquid staking tokens or less-common collateral, use Aave V3. The collateral support is broader, the liquidity is deeper, and the health-factor model is well-tested. The narrower liquidation buffer is a real risk, but the protocol’s maturity and liquidity depth reduce execution risk during volatile periods.
If you are borrowing small size (under $50,000) and you actively monitor positions, consider Morpho Blue vaults with curator-managed parameters. You will pay higher rates, but the isolated market structure allows more precise risk management. Verify the specific vault’s collateral factor, liquidation threshold, and reserve factor before depositing. Do not assume Morpho’s documentation covers vault-level mechanics.
Avoid Spark unless you are borrowing very short-term or you have strong conviction that Sky governance will maintain the current rate floor. The governance risk is asymmetric. The DAO can lower your rate, but it can also raise it or shut down the program. Rate stability is valuable, but only if the governance structure that provides it is durable.
The Takeaway
Borrow rates are not the decision. Borrow rate mechanisms are the decision. Aave and Compound both use kinked utilization curves, but Compound’s liquidation buffer is wider. Morpho uses isolated markets with curator-set parameters, which produces higher rates and more variability. Spark uses governance-set rates, which eliminates utilization risk but introduces policy risk. The durable check: verify current utilization, locate the kink point, calculate the distance between your target LTV and the liquidation threshold, and confirm the liquidation penalty structure before opening a position. The protocol with the lowest rate today may have the highest liquidation risk tomorrow. For more on DeFi protocol comparisons across lending, DEX, and derivatives categories, see the full breakdown. Rates move weekly, but the mechanism that drives them is stable. Read the mechanism.
Frequently Asked Questions
Why do borrow rates spike suddenly on Aave and Compound?
Both protocols use kinked utilization curves. Below an optimal utilization threshold (typically 80-90%), rates rise gradually. Above that threshold, rates escalate sharply to protect liquidity for lenders. When pool utilization crosses the kink due to increased borrow demand, your rate can double within hours even if your individual position has not changed. This is a pool-level state change, not a borrower-specific event.
What is the difference between LTV and liquidation threshold?
Loan-to-value (LTV) determines how much you can borrow against your collateral. Liquidation threshold (LT) is the point at which your position becomes eligible for liquidation. Aave sets LTV at 80% for ETH and LT at 82.5%, creating a 2.5 percentage point buffer. Compound uses separate borrow and liquidation collateral factors, typically creating a 10+ point buffer. The gap between LTV and LT is your downside protection before liquidation triggers.
How does Morpho Blue differ from Aave and Compound?
Morpho Blue uses isolated markets rather than pooled liquidity. Each market is a single collateral-asset pair with curator-set parameters including collateral factors, reserve factors, and liquidation penalties. This concentrates borrow demand and typically produces higher utilization and higher rates than pooled models. Morpho V2 also introduces market-discovered rates rather than formula-driven curves, which is a structural departure from how most DeFi lending has operated historically.
What happens during liquidation on Compound V3?
When your LTV exceeds the liquidation threshold (typically 90% for ETH collateral), a liquidator can call the absorb function. The protocol absorbs your collateral and applies a 5% penalty. Example: you deposit $1,000 ETH and borrow $800 USDC. If ETH drops to $880, your LTV hits 90.9%. The protocol absorbs the $880 collateral, you lose $44 as penalty, and the debt is cleared. The 5% penalty is protocol-absorbed, not paid to liquidators.
Why are Spark’s rates lower and more stable than other protocols?
Spark’s rates are governance-set by the Sky DAO, not determined by pool utilization. The Sky Savings Rate was 3.60% APY as of September 16, 2026. This is a policy floor funded by DAO-managed reserves rather than a market-clearing rate. The stability comes from governance control, which also introduces policy risk. The DAO can vote to change rates at any time regardless of market conditions, so rate stability is conditional on governance continuity.
The Weekly Yield Report
You just compared four lending protocols with borrow rates between 3.6% and 15% and liquidation buffers ranging from 2.5 to 10+ percentage points. Those numbers will shift by next week.
Every Thursday: where crypto yield actually is – stablecoins, liquid staking and DeFi lending, with the risk named next to the rate and what changed since last week.
Free. No trade calls, no allocations, no hype. Unsubscribe in one
click.










