Cozy 🦥

@Ecminings we never left. thanks for pointing it out, discord has been cleaned up.

@cozyfinance Are you guy’s coming back because I noticed your discord server is full of scammers and spammers

Introducing protected borrowing on Cozy Earn. Borrow USDC against WBTC, cbBTC, and wstETH with defined criteria for qualifying infrastructure failures. Most DeFi borrowing makes one risk obvious: if your collateral falls and your LTV gets too high, you can be liquidated. What usually stays implicit is infrastructure risk. Oracles fail. Contracts break. Keys get compromised. Critical dependencies break. When that happens, users can end up with losses from a system failure they did not cause, and no clear framework for repayment. This is not hypothetical. In 2020, Compound liquidated 121 properly collateralized borrowers after a faulty Coinbase Pro price feed. In 2022, Rari Fuse lost $80M to a reentrancy exploit and left collateral depositors exposed. Protected borrowing makes that risk boundary explicit. If a qualifying infrastructure failure affects an eligible collateral vault, Disaster Reserve funds may be used toward repayment, subject to eligibility criteria, exclusions, terms, and reserve availability. A clearer answer to what happens if the infrastructure fails is now live on Cozy Earn. Review terms before borrowing. Not insurance or a guarantee. Repayment applies only in certain qualifying infrastructure-failure scenarios and may be partial or unavailable.

Browse borrowing opportunities here: earn.cozy.finance/vaults/borrow

Borrowing is now live on Cozy Earn Use WBTC, cbBTC, or wstETH to borrow USDC from the Cozy USDC vault

The real cost of Resolv was not just the ~$23M extracted. It was the race to exit once depositors stopped knowing where the loss would land. As of today, Resolv holds $95M in assets against $173M in liabilities. The depositors who left early were right to leave. We saw a similar pattern with Stream / xUSD last year. The initial loss was bad enough. But the bigger damage came from how uncertainty propagated through lending markets and curator vaults after the fact. That is the deeper problem with bad debt in DeFi. If losses are socialized immediately, every lender gets marked down at once. But if losses are not fully realized yet, early exiters can still get out while the hole is still sitting somewhere in the system, and the last holders are the ones who eat it. Affected vaults behaved differently after Resolv. Some were hit immediately. In others, automated curator liquidity services continued supplying capital to broken USR markets for hours after the exploit. Omer Goldberg of Chaos Labs found that most of the bad debt in Gauntlet's vaults was created after the initial exploit. Others were drained because users believed the bad debt had not been fully crystallized yet. But the depositor incentive was the same in both cases: leave before you find out which system you're in. And this panic does not stay contained. It spreads to vaults without obvious exposure. It spreads to protocols without direct exposure. It becomes a system-wide question of trust. Risk curator protocols lost roughly $2.5B in TVL within three days, including vaults with no Resolv exposure. Everyone deciding they no longer want to be around when accounting catches up. The second-order damage from incidents like Resolv is larger than the exploit itself. This is endemic to a deeper problem in DeFi lending: a lack of predictability and transparency around loss realization. The cost of exiting is usually low. The downside is highly uncertain. And in the worst case, that downside can be total loss. Last year, we proposed a solution to Euler: turn off bad debt socialization and repay bad debt atomically instead. The idea is simple: when bad debt is realized, a dedicated sidecar of assets is drawn down immediately to repay it — automatically, by burning staker shares, with no governance vote and no manual intervention. Users can see the size of that buffer onchain at any time. So instead of guessing where losses will land, they can evaluate the coverage directly. That changes the incentive entirely. If users know bad debt will be repaid automatically from a visible pool of assets, they do not have to worry about racing for the exits before accounting catches up. That mechanism is live in production on Cozy Earn. And we hope to see more curators adopt this model. As many have pointed out over the last 24 hours, the constant risk of large losses and the trust-breaking setbacks from these incidents paint a pessimistic picture of DeFi's future. But if lending markets move toward systems where risk is more transparent and recourse or balance protection is more predictable, we may get more orderly behavior in stress events instead of reflexive bank runs. If you want to lend with a clearer protection model, visit Cozy Earn. If you are a curator or protocol that wants to improve your offering on this dimension, reach out. We have spent the last several years building infrastructure for exactly these cases.

Original Euler pilot proposal to heal bad debt and prevent bank runs forum.euler.finance/t/cozy-safety-…

Lending with layers of protection live on Cozy Earn earn.cozy.finance

both @cozyfinance, @SteakhouseFi and @LidoFinance have first loss capital put up for their curated vaults and if you're a curator that wants to do the same, @cozyfinance has infrastructure you can use to easily set it up for your vaults

View borrowing opportunities here: earn.cozy.finance/vaults/borrow

earn.cozy.finance Lend with layers of protections And soon, borrow with protection too

A mispriced oracle led to a 350+ ETH loss realized by borrowers who were wrongfully liquidated Aave is going to reimburse these users out of pocket but the details are still unknown Cozy Earn's disaster reserve would handle this elegantly, repaying users predictably

Try it out at earn.cozy.finance

Cozy Earn has crossed 2M TVL in under a week since launch Yields are higher than market with the most lender protection for a vault of its kind

In Cozy Earn, assets staked by the curator to cover first losses are programmatically drawn from to repay bad debt It's the only lending system that aligns incentives between curators and lenders with deterministic repayments to cover user losses x.com/Defipeniel/sta…

Introducing Cozy Earn DeFi lending with actual recourse. First-loss capital backs every vault. Losses hit protection layers before they reach you. Live now.

Security isn’t only about preventing exploits. It’s also about how systems absorb damage when something goes wrong. Cozy’s risk infrastructure, first-loss capital and protection layers, complements the prevention stack. Congrats to the @cozyfinance team on the launch of Cozy Earn. Proud to support the protocol’s security. 🫡

first loss capital is essential but not enough for defi lending. that's why in cozy earn there's a vault that covers excess bad debt AND a disaster reserve that's held in uncorrelated assets without the same exposures as the lending vault x.com/defiprime/stat…