The Alchemix Comeback: Vaults, Fixed Yields, 10x Looping, and Why They've Been Building V3 For Years

Alchemix began as a way to deposit DAI, receive a token representing future yield, and eventually reclaim the DAI as yield bought the token back. The idea evolved into a stablecoin and what Scoopy calls self-repaying loans: borrowers could deposit collateral and borrow while yield automatically repaid debt. Early vaults earned roughly 5–6% in 2020; later Yearn strategies reached about 70% on DAI and 50% on USDC. Launched in early 2021 expecting perhaps $10 million in TVL, Alchemix later approached $1 billion at its peak.

V1 and V2 let users borrow against yield-bearing collateral, but their design constrained growth and could leave LPs waiting a long time to withdraw when pools were unfavorable. Backing alETH required roughly $1 of TVL for every $1 borrowed, a constraint Scoopy says kept the protocol around $30 million to $60 million. The team recognized the need to change the system in late 2022 and early 2023, amid the damage to confidence and liquidity from the Terra/Luna and FTX collapses. V3 is a complete redesign intended to balance depositors, borrowers, and LPs while requiring less reliance on token incentives.

V3’s mixed-yield token is meant to serve both as a passive-yield product and as collateral. It would combine third-party yield tokens with custom strategies in a customized Morpho V2 vault, with up to 15 strategies running at once. The DAO would set weights as risk curator, with proposed allocation caps of 10% for high-risk strategies and 30% for medium-risk strategies, and no cap for low-risk ones. Q4 2025 modeling estimated returns of about 5–7% for USDC and 4–5% for ETH, but these are estimates based on the mix of yields, not guaranteed rates. The vault page is expected to show strategies, allocations, yield explanations, and external information.

Borrowers would be able to take up to 90% LTV, with collateral valued using the mixed-yield token’s internal oracle and cost basis rather than mark-to-market pricing. Repeatedly borrowing and redepositing could turn an initial 100 into nearly 1,000 in principal—approaching 10x leverage—but would also create substantial debt. Liquidation is designed to trigger at 95% LTV if strategy losses push a position that far, then reduce it to 89% LTV. Scoopy says the goal is to avoid liquidations through risk curation and correct integrations, while keeping them as a safeguard.

Fixed-term transmitter redemptions are intended to provide a fixed-yield mechanism and help balance borrowing and selling that might otherwise put pressure on pegs. In Scoopy’s example, depositing 100 alUSD takes two months to convert into 100 USDC; an early exit returns the amount already converted and the remainder in L assets. The implied yield depends on the market discount rather than a posted rate: if ALUSD is at $0.98 and can be redeemed for $1 after two months, the implied return is 2%, or 12% annualized. The rate is locked in when users execute, and redemption periods are expected to be no longer than three months. AlUSD and alETH borrowing carries no interest, according to Scoopy, though fees apply to mixed-yield-token yield and collateral when redemption occurs.

Looping can magnify both returns and peg pressure. Scoopy says a $10 million deposit looped 10 times could create close to $90 million in alUSD, which, if sold into Curve, Balancer, and other markets, could substantially depress the peg depending on liquidity. His simulations suggest looping at $0.99 or higher is likely to outperform passive use; at $0.98 it may still do so with high redemption activity, while at $0.97 it may cost more than remaining passive. A looper tool and UI simulator are in development, but Scoopy did not expect them to be ready at launch. The V2-to-V3 migration is intended to preserve existing assets, and the team expects redemption and leverage activity to increase LP churn and volume and support more natural APY.

Scoopy identifies external yield providers as V3’s main risk. A catastrophic 20% loss in a yield token while the protocol is around 85% LTV could leave liquidations unable to cover the shortfall. In insolvency mode, the transmitter rate would be adjusted to match solvency, and deposits and borrowing would be blocked while the protocol attempts recovery. He calls this a real black-swan risk. Six human audits have been completed, and additional automated and AI-assisted testing found bugs that the team patched; a recent human audit found no critical issues, though other findings remain under review.