A lending market with isolated per-asset risk and unusually high advertised LTVs — live on Monad since mainnet launch day and already carrying real scars from another protocol's exploit.
Curvance is a lending protocol with risk-isolated markets — each market stands alone rather than sharing one pool, so bad debt in one can be contained instead of spreading. Deposits earn a representative "pToken" that keeps accruing the underlying asset's own yield (staking, LST or vault yield) while also backing a loan, which Curvance calls "productive collateral" — for yield-bearing assets specifically, Curvance redirects the supplied token back into its underlying protocol so it keeps earning native yield while simultaneously serving as loan collateral, a design explicitly built for looping strategies. It advertises unusually high loan-to-value ratios — up to 97.5% on select assets — via a "Dynamic Liquidation Engine" that runs a short auction window for underwater positions instead of a flat liquidation penalty, aimed at cutting the cost of getting liquidated.
The engine scores every position on a continuous "lFactor" between a soft debt limit (healthy) and a hard debt limit (maximum risk) — below the soft threshold lFactor sits at zero, above the hard threshold it maxes out, and in between it interpolates linearly. Both the close factor and the liquidation incentive are functions of that same lFactor, so a partial, "soft" liquidation kicks in and scales up smoothly as a position deteriorates, reaching full liquidation only once the hard threshold is crossed. The auction buffer that determines how tight pricing is during liquidation differs by correlation: a much tighter buffer for correlated-asset pairs than for uncorrelated ones — the direct mechanical reason a 97.5% ceiling is offered on select correlated, top-tier pairs specifically, not across the board.
Curvance and Gearbox are the only two protocols in Monad's lending landscape that let one unit of capital do double duty simultaneously — external yield plus collateral or leverage utility — but they arrived at opposite risk designs. Gearbox bounds composability risk by whitelisting the execution path: a Credit Account can only ever touch a pre-approved list of external protocols, and borrowed funds never leave the account. Curvance's risk is less bounded, because the collateral asset itself is an external protocol's token, whose integrity Curvance's own contracts have no way to independently verify — which is exactly the failure mode the May 2026 incident below crystallized. Against Aave's standard 75-80% LTV, a Curvance position at 97.5% has a razor-thin roughly 2.5% buffer before liquidation risk begins, versus Aave's 20-25%+ cushion; the Dynamic Liquidation Engine's graduated curve exists specifically to manage that thinner margin in real time, genuinely more sophisticated than a flat liquidation penalty, but compensating for a more fragile starting position rather than eliminating the fragility. Against Morpho and Euler's vanilla-collateral isolated markets, where a curator picks oracle, asset and LTV explicitly market by market, Curvance's productive-collateral model adds a dependency neither of those takes on by default.
Curvance went live on Monad on November 24, 2025, mainnet launch day, with day-one markets spanning cbBTC, Monad's liquid-staking tokens (shMON, aprMON, sMON, gMON), stablecoins (AUSD, USDC, savUSD) and WMON/WBTC/WETH — despite originally being pitched as an omnichain money market, its only live production deployment today is Monad. It has grown considerably since launch, though its governance token, CVE, is not yet tradable on any tracked exchange even though it has defined utility (locking as veCVE for governance and a share of fee revenue).
In May 2026, an entirely separate protocol called Echo — a Bitcoin-focused DeFi platform primarily deployed on Aptos with a Monad expansion — had the admin key on its eBTC token contract compromised, with no multisig, timelock or minting cap in place to slow the attacker down. The attacker self-granted admin and minter roles, minted 1,000 fake eBTC (roughly $76-77M in paper value at BTC spot), then deposited 45 of those tokens (about $3.45M notional) into Curvance's isolated eBTC/WBTC market as collateral and borrowed real WBTC against it — with the actual loss reported in a range of roughly $816,000 to $870,000, later routed through Tornado Cash. The remaining 955 fake eBTC, worth roughly $73.2M on paper, went unlaundered because Monad's own DEX liquidity couldn't absorb dumping it, and were later burned once Echo regained control of its admin role. Curvance's own contracts were never breached — its lending logic behaved correctly given inputs it had no way to independently verify were fraudulent — and its isolated-market design contained the damage to that one paused market rather than letting it spread to Curvance's other listings. What isn't confirmed by any available source is whether the resulting bad debt was ultimately absorbed by a reserve, socialized to that market's lenders, or recovered from Echo — treat that specific outcome as an open question.
The Dynamic Liquidation Engine and 97.5% max LTV are unusual relative to standard Aave/Compound-style liquidation and have less real-world stress history than incumbent designs. And as the eBTC incident showed, isolating a market limits contagion but doesn't prevent losses in that specific market — collateral quality still matters, and accepting a third-party mintable or wrapped token as full-value collateral means inheriting that token issuer's own operational security (admin-key hygiene, timelocks, mint caps) as an unstated dependency of Curvance's own solvency.
Note: Listing a market for an asset you can't independently verify the backing of is a risk every lending protocol takes on — Curvance's eBTC market is a concrete example of what that looks like when it goes wrong.
There's no single binary trigger — the Dynamic Liquidation Engine scores your position on a continuous lFactor between a soft and hard debt limit. Cross the soft threshold and a partial liquidation kicks in, with the close factor and liquidator incentive both scaling up as your position worsens, reaching full liquidation only at the hard threshold.
Your posted asset keeps earning its native external yield while locked as Curvance collateral, since Curvance redirects the deposit back into the underlying protocol rather than letting it sit idle. The tradeoff: your collateral's integrity now depends on that external protocol behaving correctly, not just Curvance's own contracts — a dependency the May 2026 eBTC incident showed can be exploited when the external asset turns out fraudulent.
No — Curvance's own contracts were never breached. Echo Protocol's eBTC token had its admin key compromised on Monad, letting an attacker illegally mint 1,000 fake eBTC. The attacker deposited 45 of them into Curvance's isolated eBTC market and borrowed roughly $816,000-870,000 in real WBTC against it. Curvance's contracts executed exactly as designed — the failure was upstream, at Echo's own key security.
Not confirmed by any available source. Curvance paused the affected market and said isolation contained the loss to it, but no source specifies whether the shortfall was absorbed by a reserve, socialized to that market's lenders, or recovered from Echo — this remains an open question rather than a resolved one.
Sources: Curvance Docs, Curvance, Dynamic Liquidation Engine — Curvance Docs, Asset Types — Curvance Docs, Inside Echo's $76M mint exploit — coinmarketman.com, Echo Protocol suffers $76M exploit in eBTC minting attack on Monad — CoinDesk
Last reviewed 2026-07-08. More Monad research.
Monad DeFi board · Markets · Swap on NullTerminal