By Eco research. Updated May 2026.
​

Aave, Morpho, Spark, and Fluid are the four largest non-custodial lending protocols on Ethereum and major L2s in May 2026. They share a common job (match suppliers to borrowers against onchain collateral) but use different architectures: pool-based, isolated-market, governance-aligned, and DEX-integrated. This is a side-by-side comparison of TVL, stablecoin supply rates, audit history, governance tokens, and risk parameters (LLTV, oracle stack, liquidation thresholds). Figures cite DeFiLlama on May 24, 2026.
​

The headline differences:

Aave V3 runs a shared liquidity pool per market: every supplier of USDC into the Ethereum mainnet market shares one risk surface with every other USDC supplier. Isolation mode and e-Mode are guardrails on this shared pool, not separate markets. Morpho Blue inverts the model: each market is a tuple of (loan asset, collateral asset, oracle, interest-rate model, LLTV) and lives in its own contract. There is no cross-contamination between Morpho markets, but suppliers must read every market's parameters or delegate that work to a MetaMorpho vault curator.
​

Spark is a fork of Aave V3 deployed by Sky governance. Spark inherits Aave V3's pool model but routes idle USDS into the Sky Savings Rate, so the supply APY tracks SSR plus borrow-demand premium. Fluid takes a different angle: a Fluid USDC position can act as both lending collateral and DEX liquidity. The "smart collateral" and "smart debt" primitives let a single deposit earn lending yield, trading fees, and reduce borrow cost simultaneously. Architecturally, Fluid is closer to a unified money market than a pool lender or an isolated primitive.
​

The matrix below pulls protocol-level TVL, stablecoin supply rates (Ethereum mainnet markets, 30-day window), audit firms, governance tokens, and the load-bearing feature for each protocol. Rates are variable.
​

The spread between the cheapest and the richest USDC supply APY is roughly 3 percentage points on any given day in May 2026, and the spread tracks borrow demand. When ETH funding turns negative or basis-trade demand thins, the Morpho-over-Aave premium compresses to 50bps. When demand for sUSDe collateral surges, the Morpho vaults that accept sUSDe push USDC APY well above the pool lenders.
​

Aave V3 holds $14.6B TVL across Ethereum mainnet, Arbitrum, Base, Optimism, Polygon, Avalanche, and several other deployments. The Ethereum mainnet USDC market alone holds roughly $3.1B in supply. Supply APY ranges 3.8-5.2% on the trailing 30-day window. The protocol uses a shared liquidity pool per asset with three risk modes: standard (full collateral set), isolation mode (one collateral, capped debt ceiling), and e-Mode (correlated assets, higher LTV).
​

Risk parameters on the Ethereum mainnet USDC market in May 2026: 77% LTV, 80% liquidation threshold, 5% liquidation bonus for the liquidator. The oracle stack is Chainlink for the canonical price feeds with a custom AaveOracle wrapper that adds a fallback path. The Safety Module is staked AAVE plus ABPT (Aave Balancer Pool Token) that backstops shortfall events; size moves with AAVE price and is not a fixed dollar figure. Suppliers receive aUSDC, a rebasing receipt token that accrues interest each block. Audits include OpenZeppelin, Trail of Bits, SigmaPrime, Certora, and ABDK across more than ten formal reviews since the V3 launch in 2022. AAVE governance sets parameters via offchain Snapshot vote followed by onchain Aave Governance V3 execution. See the Aave V4 GHO yield breakdown for the borrow-mint mechanics coming with V4.
​

Morpho Blue ships a minimal, immutable lending primitive. Each market is defined by five parameters: loan token, collateral token, oracle, interest-rate model, and LLTV (the Morpho term for the liquidation loan-to-value, the threshold above which a position becomes liquidatable). Markets are permissionless, so anyone can deploy a new (loan, collateral, oracle, IRM, LLTV) tuple. Risk does not bleed between markets; a bad oracle in one Morpho market never affects another market's solvency.
​

Most retail USDC supply lives one level higher in MetaMorpho vaults curated by Gauntlet, Steakhouse Financial, Block Analitica, and Re7 Labs. Curators allocate vault deposits across markets they deem acceptable risk and rebalance as parameters change. The base protocol is immutable; the vault layer is governed by each curator. LLTV ratios on Morpho commonly run 86-94% for blue-chip stablecoin-against-ETH markets (compared to Aave's 80% for the same collateral), which is why supply APYs run 50-150bps richer than Aave. Liquidations on Morpho run through MEV-resistant auctions with no fixed liquidation bonus; the bonus is set per market. Oracle choices are per market and commonly Chainlink for blue-chip assets, with composite oracles (Chainlink × Redstone × Pyth) for newer collateral. Audits include Spearbit, Cantina, ChainSecurity, and OpenZeppelin. MORPHO token holders vote on protocol-level fee parameters, not market-level risk.
​

Spark is the lending protocol deployed by Sky (formerly MakerDAO). It started as a fork of Aave V3 and retains the pool model, but its economic engine is different. Idle USDS on Spark routes into the Sky Savings Rate via the DSR Manager, and that yield passes through to USDS suppliers. USDC supply APY on Spark tracks the SSR (3.75% in May 2026) plus a borrow-demand premium that compresses or expands with utilization.
​

Spark holds $3.5B TVL across USDC, USDS, ETH, and wstETH markets in May 2026. Spark's USDC market on Ethereum runs at 75% LTV and 78% liquidation threshold with a 5% liquidation bonus, slightly tighter than Aave to reflect the protocol's younger operating history. The oracle stack mirrors Aave's: Chainlink primary, fallback path through a custom oracle adapter. Audits include ChainSecurity, Cantina, and Spearbit. Spark does not have its own governance token in the canonical sense; SPK has launched but parameters flow from Sky executive votes routed through the Spark proxy. The trade-off versus Aave: simpler governance, tighter spreads versus a pool that is fully Sky-controlled. For a Sky-supportive depositor, Spark is the natural USDC venue; for an Aave-or-nothing depositor, Spark is a secondary position.
​

Fluid is the newer entrant, built by the Instadapp team. It combines lending and DEX liquidity into a unified position framework. The two load-bearing primitives are "smart collateral" (collateral that simultaneously acts as DEX liquidity and earns trading fees while securing a loan) and "smart debt" (debt that is paired with a DEX position to reduce effective borrow cost). A Fluid USDC-USDT smart collateral position can earn lending APY, swap fees, and rebalance gains in one slot.
​

Fluid holds $1B TVL across markets in May 2026, with USDC supply APY in the 4.3-5.5% range and USDT supply APY in the 4.2-5.3% range. LTV ratios on Fluid commonly run 80-86% for blue-chip pairs, supported by tighter liquidation parameters: 1-3% liquidation penalties and continuous liquidation curves rather than fixed-bonus auctions. Fluid's oracle stack is Chainlink-primary with TWAP fallback for DEX-paired positions. Audits include Statemind and OpenZeppelin. FLUID token holders govern protocol parameters. The main risk-versus-yield trade-off: Fluid's smart collateral and smart debt are newer code paths, so audit count is lower than Aave or Morpho. Depositors comfortable with newer protocols capture 50-100bps over Aave equivalents.
​

The four protocols differ most sharply in how they price and enforce liquidation risk. Aave V3 uses a single liquidation threshold per asset (e.g., 80% for USDC backed by ETH) with a fixed 5% liquidation bonus for the liquidator. The threshold is governance-set per Aave Risk Working Group recommendation, supported by Gauntlet and Chaos Labs risk models. Oracles are Chainlink-primary across all V3 markets.
​

Morpho Blue's LLTV is per-market and ranges 70-94.5% depending on the (loan, collateral) pair. Common LLTVs in May 2026: 86% for USDC against wstETH, 91.5% for USDC against sUSDe, 94.5% for USDC against weETH. Higher LLTVs reflect tighter collateral correlation. Spark uses pool-model thresholds inherited from the Aave V3 codebase but with slightly more conservative parameters (75-78% LTV on USDC markets versus Aave's 77-80%). Fluid runs continuous liquidation curves (positions liquidate proportionally rather than in fixed-size chunks at a single threshold), which reduces liquidation cascade risk but adds smart-contract surface area. Oracle stacks vary: Aave and Spark default to Chainlink primary with fallback adapters; Morpho lets each market pick its oracle (commonly Chainlink, sometimes composite Chainlink × Redstone × Pyth); Fluid uses Chainlink with TWAP fallback for DEX-paired positions.
​

No single protocol dominates on every dimension. Aave wins on audit depth, TVL, and battle-tested code; the trade-off is the lowest supply APY of the four. Morpho wins on rate when curators are aggressive; the trade-off is that depositors must read curator parameters or delegate that work. Spark wins on simplicity if a Sky-aligned exposure is acceptable. Fluid wins on integrated lending plus DEX yield; the trade-off is shorter operating history.
​

Practical heuristic: a depositor prioritizing audit depth and battle-tested code defaults to Aave V3 mainnet. A depositor hunting an extra 50-150bps with willingness to read curator dashboards picks Morpho Blue via Steakhouse, Gauntlet, or Re7 Labs vaults. A depositor who wants a governance-set rate without exposure to borrow-demand volatility uses Spark or Sky directly. A depositor who runs LP positions already and wants the same capital to earn lending yield uses Fluid. Across all four, splitting USDC supply across two protocols is a cheap hedge against single-protocol smart-contract risk. See the best DeFi lending protocols 2026 breakdown for a wider field that includes Compound, Euler, and Sky lending.
​

Eco Routes settles USDC across 15+ supported chains, so a depositor on Base who wants USDC into a Morpho vault on Ethereum mainnet can route the cross-chain leg through Eco intents without holding gas tokens on the destination. Eco is one of several USDC-bridging paths; Circle's native CCTP, Across, and LI.FI are alternatives. The supply APY rates above are protocol-level and do not depend on the bridge. Bridge choice affects gas cost and settlement latency, not the underlying yield.
​

Aave V3 has the longest audit history (10+ formal reviews since 2022) and the deepest TVL ($14.6B in May 2026), which together make it the conservative default. Morpho Blue's immutable base protocol is also a strong audit story, but most retail capital sits in MetaMorpho vaults, which add a curator layer of risk on top of the base.
​

LLTV (liquidation loan-to-value) is the Morpho Blue per-market parameter that defines the threshold above which a position becomes liquidatable. Morpho markets in May 2026 run LLTVs between 70% and 94.5% depending on collateral correlation. Higher LLTV = more capital efficient = thinner liquidation cushion.
​

Spark inherits Aave V3's pool model but routes idle USDS into the Sky Savings Rate and passes that yield to suppliers. Governance flows from Sky executive votes (not an independent Spark DAO). Risk parameters on Spark are slightly more conservative than Aave's. Token governance is via SPK and Sky-aligned proposals.
​

No. Morpho vaults run 50-150bps richer than Aave on most days in May 2026, but the spread compresses to roughly 50bps when borrow demand falls, and individual conservative vaults (Steakhouse USDC) can briefly match Aave during slack-demand periods. The Morpho premium reflects looser LLTVs and concentrated collateral types, both of which carry risk that does not exist on Aave.
​

Yes. Fluid uses continuous liquidation curves rather than fixed-size auction liquidations. A position approaching its liquidation threshold is liquidated proportionally rather than in a single chunk at a fixed bonus, which reduces cascade risk during volatility events. The trade-off is more complex smart-contract logic and shorter audit history than Aave's fixed-threshold model.
​

Sources and methodology. Protocol TVL pulled from DeFiLlama on May 24, 2026. USDC and USDT supply APYs from DeFiLlama yields, 30-day window on Ethereum mainnet markets. Audit lists verified against each protocol's GitHub audit registry. Risk parameters (LTV, liquidation threshold, LLTV) pulled from each protocol's public risk dashboard: Aave, Morpho, Spark, Fluid. Sky Savings Rate from Sky governance documentation. Figures refresh quarterly; supply APYs and risk parameters are variable.
​