Curve's founder, Michael Egorov, has put forth a market-driven approach to resolve approximately $700,000 in bad debt associated with LlamaLend, Curve's lending platform. In a governance post, Egorov outlined a free-market based recovery method with an option-like payoff, open to all who wish to participate.
This approach invites, but does not require, the involvement of Curve DAO. The bad debt is currently situated in LlamaLend's CRV-long market, where users can borrow Curve's crvUSD stablecoin against CRV, the protocol's governance token. This trade functions as a wager that CRV will maintain or increase its value. However, if CRV's value drops too rapidly, the collateral may not be sold quickly enough to fully repay lenders.
This scenario occurred following the October 10 crash, triggered by President Donald Trump's announcement of tariffs on all Chinese goods via a post on Truth Social. Rather than relying on Curve's DAO to cover the shortfall, Egorov proposes packaging the affected lender positions into a tokenized vault, allowing traders to buy and sell them through a dedicated Curve pool. The primary objective is to provide trapped lenders with an exit strategy while enabling outside buyers to determine the value of the distressed claims.
The bad debt resulted from the October 10 crash, which saw over $19 billion in leveraged liquidations within hours, marking the largest single-day deleveraging on record. Although Curve's crvUSD minting markets withstood the sell-off, LlamaLend was not entirely unaffected. Prices plummeted, and gas costs rose, resulting in a scenario where some liquidations could not occur in time. Lenders in the CRV-long market were left with deposits backed by approximately 70% of their stated value.
The market is designed to mitigate this risk through an automated market maker, LLAMMA, integrated into the lending system. Instead of selling a borrower's collateral all at once when prices fall, LLAMMA converts the collateral in stages as the market moves. According to Egorov, the providers of borrowable liquidity in this market were exposed to losses during liquidation protection and cannot withdraw their positions, which are currently around 70% backed. However, during the October 10 crash, the market moved too rapidly.
Arbitrage traders, who help maintain the system's balance by buying and selling across price gaps, were unable to keep up. Some lender positions ended up in a vault token that cannot be redeemed at full value today.
Egorov argued that the token still holds value because the loss is not open-ended. The distressed positions already hold crvUSD converted from CRV, so further CRV declines should not deepen the shortfall. If CRV rises above approximately $0.96, the conversion starts to reverse, and the positions begin taking in CRV collateral again.
Full recovery would occur around $1.24. Egorov stated that if CRV's price grows, positions with bad debt will deliquidate, meaning the system would start converting crvUSD back into CRV collateral. However, if CRV's price drops, the collateral is already converted to crvUSD, so the vault deposits will not be less backed. At the time of writing, CRV is trading near $0.23, well below both levels.
The proposed pool would utilize Curve's Stableswap design, with a 1% swap fee and liquidity centered around 71% solvency rather than full value. This means the pool would not treat the distressed token as if it were worth one dollar on the dollar; instead, it would price the token closer to the amount currently backing it. For trapped depositors, the pool offers a choice: they can either wait for a CRV recovery or sell their vault tokens at a discount and move on. For buyers, the trade resembles a long-term bet on CRV, as they purchase a claim that is partly backed today and could become more valuable if CRV recovers.
This gives the token an interesting option-like property, as its fair price and price floor increase if CRV's price rises, without decreasing if CRV's price falls. Liquidity providers in the new pool would earn swap fees and any CRV incentives allocated by Curve's DAO. Admin fees would partly accrue in the distressed vault token itself. Egorov has requested that the DAO retain these tokens rather than convert them, which would gradually transfer some of the bad debt onto Curve's balance sheet through trading activity.
The proposal's timing carries added weight, as it follows an attacker's exploitation of Kelp DAO's LayerZero bridge, resulting in the release of 116,500 unbacked rsETH worth approximately $292 million. The attacker then deposited the unbacked rsETH into Aave as collateral and borrowed real WETH against it. Aave now faces up to $230 million in bad debt, prompting an industry-wide response in the form of a coordinated bailout through DeFi United, a recovery effort led by Aave service providers. Egorov presents Curve's pool as a distinct model, one that builds a market for distressed claims and allows buyers to determine the price, rather than relying on a bailout.
If this proves to be a successful pilot study, it could be applied to similar challenging situations at Curve or other protocols.