Subject: Alchemix Bridging and Fantom Update
Proposal
The Alchemix team proposes to change the current bridge, which is Multichain, to the Connext bridge. The new bridge will have similar functionality as the current Multichain bridge, with UI/UX and security improvements as outlined in the Specifications section. One notable change is that with the upcoming veALCX launch, Alchemix will retire Layer 2 gALCX in favor of ALCX being the incentivized liquidity token on other chains.
Fantom is not supported by Connext, as Fantom is not a Mainnet rollup. As part of this proposal, Alchemix on Fantom will begin to be wound down in a way that encourages existing depositors and liquidity providers to unwind their positions and move to Arbitrum, Optimism, or Mainnet.
The primary milestones to complete this proposal are:
- Launch ALCX, alETH, and alUSD bridging to Arbitrum and Optimism through Connext
- Alchemix Mainnet, Arbitrum, and Optimism Multisig to carry out operations to drain multichain contracts of existing alAsset liquidity
- Alchemix Fantom Multisig to re-enable swapping to bridge tokens (but not from), in order to make it possible to bridge gALCX and alUSD to mainnet again.
- Alchemix Fantom Multisig to lower deposit caps to zero on Fantom and disable new alUSD loan mints, and kill all incentives for alUSD and gALCX on Fantom.
- Alchemix to loan $50k alUSD, $50k worth of alETH, and $50k of ALCX to a routing service, which will create liquidity for faster and cheaper bridging between approved L2s (Arbitrum and Optimism to start) as well as faster bridging from L2 to L1 than the typical 7 day period.
Context and Motivation
Multichain bridging services suffered outages in the last few weeks, and there have been rumors circulating about the Multichain team and their multisig. The Multichain team has been slow to respond to the outages and rumours, and vaguely cited “force majeure” as reasoning. With the current Alchemix bridging system, Multichain has the effective ability to mint unlimited canonical alAssets and gALCX on any chain on which those assets are deployed. Given the uncertainty surrounding Multichain and the power that their protocol has over L2 Alchemix, it is proposed to migrate away from Multichain to a new bridging protocol, while simultaneously improving the decentralizability of Alchemix on L2s.
technical note: technically Multichain is requesting the alAsset be minted by the alAsset contract itself, but this is effectively the same as minting for the purpose of this proposal.
For Alchemix to become sufficiently decentralized, it must eliminate dependencies that rely on operators and third parties. The easiest way to eliminate dependencies would be to silo every chain on which Alchemix is deployed - alUSD on Optimism would be an entirely different token than alUSD on Arbitrum. This means that each chain would need to fully support liquidity for the TVL on that chain. The “frictionless” way to eliminate dependencies would be to allow infinite transfer of alAssets between chains (referred to as an OFT, Omnichain Fungible Token, or more generally, burn/mint functionality). An OFT means that any chain can support liquidity for any other chain, which makes for very efficient bizdev/liquidity incentives - however, this means Alchemix is only as secure as its weakest chain or the oracles/relayers that underpin the system.. Lastly, there is a way to “soft-silo” the chains that allows limited bridging for efficiency gains without sacrificing decentralization. The soft silo has a few features:
- alAssets can be infinitely bridged from Mainnet to any Layer 2.
- Bridging alAssets from any Layer 2 to Mainnet is capped by the amount of alAssets that have been bridge from Mainnet.
- A side effect of #1 and #2 is that any alAssets bridged to a Layer 2 chain may be moved to any other supported Layer 2 chain.
- Any bridge (including Connext) can create liquidity pools to allow users to loan funds to anyone wishing to bridge more quickly between chains (mostly, L2<>L2 or L2->Mainnet). Alchemix can incentivize or loan assets to these liquidity pools, to earn revenue and allow near instant bridging.
This approach is ultimately quite similar to what was built through Multichain. However, Multichain relies on off-chain entities and multisigs to secure the bridge, which introduces third-party dependencies. This is necessary for communication between many chains, as not every chain has a way to communicate with Mainnet.
However, Layer 2 chains do have a way to communicate with Mainnet - through the canonical bridges. Communication through the Arbitrum and Optimism bridge requires no intermediaries, meaning cross-chain messaging can be secured entirely by the Ethereum validator set and the L2 Fraud Proofs. This is ultimately a more secure and decentralized approach to L2 bridging that Connext has built around, and is the approach to which the Alchemix BizGov subDAO and devs wish to migrate. Additionally, the use of the Connext bridge means bridge time is 30 minutes back to mainnet, instead of 7 days like with canonical bridges. Generally speaking, this is due to the specific use case of bridging requiring less time for security guarantees than the general canonical bridge, which must account for any and all forms of cross chain settlement.
This proposal outlines the new bridge solution, the migration plan, and how to treat the Fantom deployment given it is not a rollup (and therefore not compatible with the proposed bridging system).
Connext Specifications (Arbitrum and Optimism)
Specifications will be as follows (note, the next- prefix represents the Connext-specific bridge token on the specified chain).
ALCX
- Alchemix will route ALCX through the canonical Arbitrum and Optimism bridges. This will create the canonical ALCX on these chains. This is the ALCX that will have liquidity incentivized.
- In order to bypass the 30 min settlement time of Connext bridge and avoid the new for all bridge transactions to route through mainnet, Alchemix will loan $50k of ALCX from its treasury to Connext to provide atomic swap bridge liquidity.
alETH and alUSD
- Alchemix will deploy a Connext specific representation of alAssets on Arbitrum and Optimism: next-alETH and next-alUSD.
- Connext will map L1 alETH and alUSD to the corresponding representations on L2s.
- Alchemix will give next-alAssets the ability to mint deployed alAssets on each L2.
- Alchemix will integrate Connext to wrap next-alAssets into alAssets on L2 (and vice versa) in a single transaction so that next-alAssets are invisible to the user.
- Alchemix will loan XX alETH and YY alUSD from its treasury to a Connext router to provide “fast” liquidity to enable bridging between Arbitrum and Optimism in 45-120 seconds.
- In the future, Alchemix can considering creating a similar system for the canonical bridge tokens for each chain, or upgrading the connext implementation as they work towards their xERC20 standard (see below paragraphs).
This model gives fungibility of alUSD and alETH minted by both Connext and the canonical bridge on each L2 from the perspective of the user. It also caps the risk (defined as the total amount of funds that can flow from any L2->L2 or L2 to L1) in the system to the amount bridged through Connext from L1.
Note, Connext is working on a bridging solution built into Ethereum that Alchemix would have the option to upgrade to in the future. The current solution is similar to multichain, however there are two key improvements. The first is that the bridge-specific assets are now removed from the experience, which ensures that bridge liquidity is not fractured. The second is that the mint/burn functionality is now secured by rollup fraud proofs and Ethereum mainnet validators, instead of off-chain relayers/oracles. The future solution would build mint/burn functionality into Ethereum and canonical bridges themselves, allowing bridging to be bridge-protocol agnostic and not require bridge-token intermediaries.
More information can be found here: https://connext.notion.site/xERC20-Explainer-0ac9ac7dfcaa4e30b0b29c1012e72ac1
Migration Plan
Step 1 - Disable Minting (Already Executed by Multisig)
The first step is to disable the ability for multi-assets to mint Alchemix assets on any layer 2 chain or side chain. This will secure the assets on that chain. This will also mean that for a period of time, users will not be able to bridge alAssets between L2s or from Mainnet to L2s. This step has already been carried out as a result of precautions taken by multisig signers in response to aforementioned events around Multichain.
Technical note: Each alAsset and gALCX deployment on each L2 has the ability to grant and revoke minting ability to different contracts. Currently, the Alchemists and the multichain assets are the only tokens with this ability.
Step 2 - Launch New Bridge
See above section for the specifications of the new Connext bridge.
Step 3 - Migrate Mainnet Bridge Liquidity
The Multichain contracts currently hold alETH, alUSD, and gALCX on Mainnet. Step 1 secures L2 alAssets, however this does not solve the issue of the Mainnet funds, which are backing L2 assets, being at risk by being held in the multichain contracts.
To migrate the alETH and alUSD backing to the new bridge, the Alchemix dev multisig can first bridge funds to any L2 through the new bridge. Next, the multisig can bridge those funds back to Mainnet through Multichain to drain the multichain contracts. At this point, the ability of Multichain assets to mint alAssets on Arbitrum and Optimism can be disabled again.
Incentivized gALCX only exists on Fantom, and the Alchemix BizGov SubDAO and Devs do not wish to grow Alchemix’s presence on Fantom presently. Additionally, Fantom is not a rollup, so it falls outside the scope of what is possible to secure with Ethereum Validators. The net risk is about 5800 gALCX on Mainnet, and the risk is borne by anyone holding gALCX on Fantom. In this manner (because gALCX cannot be minted by anyone else on Fantom), the risk is opt-in. Alchemix will disable the ability for gALCX to bridge to Fantom, and will allow anyone to migrate gALCX back to Mainnet. In the future (near the launch of veALCX), Alchemix can potentially offer a swap contract to convert gALCX to ALCX on Fantom at the then-current conversion rate. The approach above can also be used if anyone happened to bridge gALCX to other chains through multichain.
Multichain, Fantom, and Future Chains
Fantom is not supported by Connext, as Fantom is not a rollup. Some chains that are not L2s do have canonical bridges to Mainnet (such as Matic) but are not rollups, which means a new validator set is being trusted in addition to the Ethereum validator set. At present, the only chains that would be approved for future bridging and minting through connext are rollups/L2s secured by Ethereum. Zk Rollup chains would also be future candidates, including ZkSync and Matic’s ZkEVM.
For this reason, the Alchemix instance on Fantom needs to be wound down. Alchemix V2 does not currently have a graceful way to wind down, in that depositors have no obligation to repurchase the alUSD they originally minted (especially since alUSD is over $1 on Fantom). In absence of alUSD liquidity, anyone holding alUSD on Fantom would need to wait for it to slowly convert to stablecoins using the transmuter.
In order to allow the wind-down process to begin without depegging alUSD on Fantom, Alchemix will continue to supply protocol owned liquidity to BeethovenX on Fantom to the extent that is necessary to maintain alUSD price close to the price of Mainnet alUSD.
Alchemix will re-enable Multichain bridging, and will allow anyone to bridge alUSD to Mainnet. Note that this bridging is capped by the amount of liquidity that has been bridged to Fantom. As of June 12th, there is 153k of anyalUSD available to be bridged (https://ftmscan.com/address/0xB67FA6deFCe4042070Eb1ae1511Dcd6dcc6a532E). There is 186k alUSD in the BeethovenX pool. (https://beets.fi/pool/0xff2753aaba51c9f84689b9bd0a21b3cf380a1cff00000000000000000000072e). The DAO owns 35k of the alUSD in the pool (https://debank.com/profile/0x6b291cf19370a14bbb4491b01091e1e29335e605). Therefore there is an excess of 2k alUSD even if all existing LPers were to bridge. This would still cause some level of alUSD depeg on BeethovenX - however, incentives are set to continue until mid-July and it is expected some depositors will unwind their positions before then, which helps shrink supply further.
Additionally, anyone is able to sell alUSD for another asset on Fantom and bridge that asset to Mainnet or another network (and this is possible at all times, not dependent on Multichain functionality). On the depositor side, Alchemix will lower all Fantom Opera Alchemist Fantom deposit caps to 0 and will disable additional mints of debt from those Alchemist Fantom positions. Note that the TVL of the Alchemists on Fantom is under $100k. If the above strategies are unsuccesful in winding down Fantom, the DAO could vote to force-retire Fantom and make everyone whole. This may become preferable if the cost of maintaining Fantom alUSD liquidity were to be too high, but that is not the current proposal.
In the medium-term future, Alchemix will likely only consider deploying on rollups. If another L1 wishes to host an Alchemix deployment, it would need to be siloed. In this manner, Alchemix would be open to friendly forks, parent/child relationships, or SubDAO-managed deployments of Alchemix on other chains.
In summary, the following changes are being made to the Fantom alAssets incentive program and Alchemists:
- Alchemists:
a. Lower deposit caps for all assets to zero
b. Disable ability for new alUSD loans to be minted (controlled by the alUSD FTM deployment)
- Liquidity:
a. Leave Alchemix alUSD POL staked in fantom (Approx $60k of alUSD liquidity, 35k of which is alUSD)
b. End all incentives for alUSD on Beethoven and gALCX on Spookyswap
- Bridging
a. Re-enable swapping of alUSD to multi-alUSD, which allows users to bridge alUSD to Mainnet again.
b. gALCX will continue to exist on FTM but will not have liquidity incentivized, and users may migrate back to mainnet at their leisure through multichain. The entirety of the risk is borne by users choosing to continue holding gALCX on FTM.
A note on risk
Below are a few scenarios on risks related to bridging alAssets:
There can never be more alAssets or ALCX on Mainnet than was minted on Mainnet. For this reason, if you bridge an alAsset to an L2, you do run the risk that you will not be able to bridge back (if someone else minted debt on that L2 and used the liquidity you created to bridge to Mainnet). This does not apply to ALCX, as ALCX is not minted on L2s.
Under normal operating conditions (no active exploits requiring limitations), you will always be able to bridge alAssets from Mainnet to a supported L2. It would be possible for an alAsset on an L2 to depeg relative to the price of that same alAsset on Mainnet if significant alAssets were minted on an L2, but then in the future there are not enough liquidity incentives to support the demand. alAssets on Mainnet would not depeg relative to those same assets on supported L2s, as users are always able to bridge from Mainnet to a supported L2.
If you have a debt position on a supported L2, then you are not exposed to bridging bottleneck risk, as you are able to simply repay your debt on L2 if you are unable to bridge and do not wish to sell or hold the alAsset on that chain.
Voting Options
- For: Vote in favor of the proposal.
- Against: Vote against the proposal.
- Abstain: Do not vote on the proposal.
