[AIP-90] Deploy Alchemix on Arbitrum

ButlerAndTheThirdStringers

Proposal:

Deploy an alUSD Alchemist on Arbitrum with AAVE V3 strategies enabled for alUSD and alETH, and Jones DAO jUSDC vault (for alUSD).

Deploy an alETH Alchemist on Arbitrum with rETH, wstETH, and sfrxETH strategies enabled.

Collaborate with Ramses DEX to form liquidity, offering Alchemix substantial voting power in return for contributing TVL to Ramses DEX to help form liquid markets there.

The alUSD Alchemist on Arbitrum should utilize the following parameters:

Accepted Collaterals: aUSDC, and jUSDC
Max LTV: 50%
Deposit Caps: Each collateral asset will have a $1m cap (after min 1 week of 100k, at multisig discretion)
Repayment and Liquidation Caps: $500k per 60 minutes
Maximum Loss: 5bps (0.05%)
Cap Raises: All vaults are pre-authorized to have their caps increased to $5m

The alETH Alchemist on Arbitrum will utilize the following parameters:

Accepted Collaterals: wstETH, rETH, and sfrxETH
Max LTV: 50%
Deposit Caps: Each collateral asset will have a 200 WETH cap (after 1 week of 50, at multisig discretion)
Repayment and Liquidation Caps: alETH = infinite
Maximum Loss: alETH = 200 bps (2%) (Max loss based on st/rETH to ETH price)
Cap Raises: All ETH vaults are pre-authorized to have their caps increased to 500 WETH.

Context:

Ethereum Layer 2 networks continue to gain in adoption, TVL, and mind share. As part of the Alchemix multi-chain vision, the team has been evaluating various Layer-2 networks as potential candidates for the next Alchemix deployment. Arbitrum has emerged as a leading candidate in this regard, as it is the largest Layer-2 network by TVL, a hub for established DeFi protocols and innovation, and presents a large surface area for potential collaborations.

Any Alchemix deployment to a new chain requires the following:

Compatible yield strategies;
The ability to facilitate substantial liquidity on one or more DEXs to service loans;
New Alchemist configurations, and ability to port other existing contracts such as the Transmuters and keepers; and
The ability to bridge.

Initial Vault Strategies

With the integration of AAVE and ETH liquid staking vaults on Mainnet, Alchemix now has several existing yield strategies that can be ported to Arbitrum without much difficulty. In addition, the Alchemix core developers have begun development on an adapter for Jones DAO’s jUSDC vault to integrate an Arbitrum-specific option to earn yield. Jones DAO is a yield aggregator and vault protocol where users can enter various strategies to earn yield. The jUSDC vault is a passive strategy that allows users to lend USDC, serving as counterparties to another group of users that would like to lever up on GLP. For more information on jUSDC vaults see their documentation here: https://docs.jonesdao.io/jones-dao/features/understanding-jusdc-and-jglp

Partnership with Ramses DEX

On the liquidity side Alchemix will utilize Ramses to facilitate alAsset liquidity. Ramses is a (3, 3) DEX similar to Solidly and Velodrome. Ramses has offered to provide Alchemix with a veNFT, along with a substantial initial allocation of voting power to help get a liquidity flywheel started. For more information on Ramses see their documentation here: https://docs.ramses.exchange/introduction-to-ramses/what-is-ramses. As part of this collaboration, Alchemix will move some of its existing protocol-owned-liquidity to Arbitrum to seed Ramses pools for alUSD and alETH liquidity. (at the discretion of the Bizgov and operational teams) This collaboration will benefit both parties by increasing liquidity and fostering a mutually beneficial relationship. Alchemix will have a long-term goal of accumulating Ramses voting power.

Alchemists + Other contracts

The Alchemist parameters follow the precedents set by previous parameters for the mainnet LSD vaults, and the AAVE Mainnet launch. Deposit caps are significant but smaller than on Mainnet, with an initial “soft launch” period. Liquidation caps are also kept smaller, as the amount of arbitrage possible in the event of an underlying token depeg is a function of all liquidation caps combined. Other necessary contracts will be ported and deployed as is determined to be appropriate by the operational team. Bridging is undergoing an updated through Multifarm, but until this update is completed, the process will remain the same.

Voting

A vote “FOR” authorizes Alchemix to launch alUSD and alETH Alchemists on Arbitrum with AAVE V3 strategies and Jones DAO jUSDC vault integrated for the alUSD Alchemist, and AAVE v3 WETH, wstETH, rETH, and sfrxETH vaults integrated for the alETH Alchemist, utilizing the listed parameters AND authorizes Alchemix to collaborate with Ramses DEX to increase liquidity and voting power.

A vote “AGAINST” does not authorize Alchemix to launch on Arbitrum with the above parameters and partnership with Ramses DEX.

ov3rkoalafied

PUMPED FOR ARBITRUM

two notes:

POL can be viewed in alchemix-stats: https://alchemix-stats.com/. There is $2.21m of alETH and 500k of alUSD liquidity on velodrome that could be moved. Other dex's on arbitrum have also reached out, meaning this liquidity could eventually be split between a few Opti/Arb dex's. Given OP rewards are about to begin (because yearn just finally began their program), I would not want to move ALL of the liquidity, but that will depend partially on what APRs we are seeing on RAMSES.
only aave USDC is included for aave strats. It has the most TVL of any strategy on arbitrum. Low yield are bad for users and bad for LPers/the transmuter, so in my opinion aWETH can be left out as there are 3 better options, and aUSDT and aDAI can be left out due to low demand. aUSDC is present just to give people a more familiar option if they are not comfortable with jUSDC.

KW710

ov3rkoalafied I am also pumped about Arbitrum, but I do worry about the protocol being overextended given how little has happened on Optimism and Fantom after it was deployed to each.

I think the Optimism ETH vault has seen maybe one harvest and repayment? So the only thing making it worth it for depositors there is the secondary yield farming possibilities through other protocols.

Meanwhile, Fantom still only has one yield source available even though you have chain-native options like Geist which could be very interesting to explore.

It sounds like the protocol would be deploying to Arbitrum with significantly more partners and integrations at the onset, so that is promising and encouraging. But I worry about whether the DAO and contributor team will have enough bandwidth to service it appropriately given the other two instances above.

ov3rkoalafied

KW710

Optimism has been awkward - we were well positioned to launch AAVE and Yearn there. We did launch AAVE, then their OP grant ended and they didn't renew, which took the yields to zero. Therefore, the DAO voted to incentivize yearn - however, we were waiting for their OP grant to start up so we could double-incentivize (didn't make sense to incentivize a yearn vault with low yield). Additionally, mainnet needed more strategies and it didn't make sense to launch a new strat on OP first. We've actually had the solidity code ready for yearn vaults for months at this point, it hasn't been an issue of dev availability. We also couldn't expand LSD vaults until recently, due to waiting for unlocks to be possible.

Now - a lot has changed. Yearn just started their OP program this week, so we can use that vault. The merge happened, so LSD vaults are now feasible on optimism. veALCX audit comments have gone a bit faster than expected which has freed the dev up that had gotten pulled into help. L2 has a lot of new vaults now so we can expand elsewhere.

velodrome has been a great partner on OP, we've done exceptionally well there, I think the DAO has made over $2m in treasury value on that based on the intial NFT + initial purchase of 50k or so worth.

The debatable one here is FTM -> note that of all the vaults launching on OP/ARB, only jonesDAO is new. Frankly we just haven't seend the TVL, yield, or demand on FTM that has made us interested in putting specialty dev effort into it. The yield on alchemix assets tends to be high too indicating people view it as risky, so we would have to pay more for liquidity. Ie, spooky and beethoven have been alright but not nearly the benefit to liquidity that we got with velodrome (and hopefully with ramses too, and perhaps some other arb dexes)

tl;dr it hasn't necessarily been an issue of resources, moreso timing. And the timing feels very right at the moment to expand OP and ARB for the reasons listed above.

Note that building a new yield adapter for a yield strat is quite simple, about a week of work then time to deploy, integrate to the UI, test, run comms, etc. Most of the work is on the risk evaluation side of things (verify the system can handle the yield strat) and operations to bribe for liquidity.