Proposal updated

Title: [ARFC] Aave V3 Deployment on BNB Chain

Author: @marczeller - Aave-Chan Initiative

Date: 2023-04-04

Summary:

This ARFC proposes the deployment of Aave V3 on the BNB Chain, expanding Aave’s presence and tapping into the growing DeFi ecosystem and user base of BNB Chain.

Abstract:

Upon approval by the Aave community, Aave V3 will be deployed on the BNB Chain with a selection of suitable assets for collateral, including BNB, WBTC, BETH, WETH, USDC, and USDT. This deployment aims to enhance the growth and collaboration between the Aave and BNB Chain communities, providing new opportunities for both ecosystems.

Motivation:

Deploying Aave V3 on the BNB Chain presents potential benefits for both Aave and BNB Chain ecosystems. Aave would gain access to BNB Chain’s user base, DeFi ecosystem, and potential additional revenue. In return, BNB Chain would benefit from having a flagship DeFi project like Aave on its platform, further diversifying DeFi options available to its users.

Specification:

The ARFC suggests deploying Aave V3 on the BNB Chain with the following assets for collateral: CAKE, BNB, BTCB, WETH, USDC, and USDT. The table below provides the suggested risk parameters for each asset.

Deployment Checklist:

• Risk service providers feedback
• BGD Labs Infrastructure/technical evaluation Report
• Oracle recommendation (Chainlink)
• The graph support
• Cross-chain messaging infrastructure for governance

Disclaimer

The Aave-Chan Initiative (ACI) is not compensated in any form by Binance & BNB for creating this ARFC proposal. This ARFC has been prepared solely to facilitate community discussion.
At the time of writing, the author owns ~70$ worth of assets on the BNB chain.

Next Steps

1. Integrate BGD Labs Infrastructure/technical evaluation report once available.
2. Refine the ARFC based on community feedback and risk service provider recommendations.
3. If consensus is reached and the proposal is refined, submit the ARFC for a snapshot vote for final approval.
4. If consensus is reached, submit a deployment AIP to activate Aave V3 BNB.

Copyright

Copyright and related rights waived via CC0.

Would you help clarify what BETH is here?
If it refers to Binance-PoS-ETH2, as far as I can tell, it has very bad liquidity on DEXs and even on Binance exchange, the trading volumn and depth of order books for BETH/ETH or BETH/BUSD or etc are not good.
Also holding BETH onchain (off Binance exchange) won’t receive any staking rewards from ETH2.
Recently Binance exchange announces that users can borrow other coins against BETH on Binance lending. So this “Binance only advantage” will discourage users to move BETH off Binance exchange to AAVE.

BETH is 0x250632378E573c6Be1AC2f97Fcdf00515d0Aa91B, we also completely agree that current liquidity is not great for this asset but:

• With the Shanghai upgrade in 8 days, BETH is expected, as every LST, to become more naturally peg-resilient due to the ability to redeem it at fair primary market value
• Due to the nature of BNB, it is expected that liquidity on Binance CEX is an essential consideration for bETH and other assets in this Aave V3 deployment, compared to more decentralized networks, the centralized exchange is an important component to consider
• it is expected that with Pancakeswap V3, (UNI V3 fork) on chain-secondary liquidity will become more efficient on the BNB chain.

as this ARFC will require some time to mature, it is likely that the risk profile of BETH will greatly improve in a couple of weeks, making it fitter for consideration.

Thanks for quick response. I just want to point out Binance quite centrally “hostages” the BETH that way from my point of view.

Chaos Labs Analysis and Recommendations

Overview

Chaos Labs supports listing CAKE, BNB, BTCB, BETH, WETH, USDC, and USDT as part of an overarching strategy to enable the expansion of Aave protocol swiftly to other chains.

Liquidity and Market Cap

When analyzing market cap and trading volumes of assets for listing, we look at the past 180 days data:

Liquidation Threshold

Analyzing assets’ price volatility over the past, we recommend the following LT settings:

For ETH and BTC we use the battle-tested settings used on other V3 deployments. These are initial LTs that will be optimized in the future utilizing our simulation platform.

Cake:

Untitled (39)1258×460 127 KB

BNB:

Untitled (40)1388×449 139 KB

BETH:

Untitled (41)1271×448 142 KB

We support listing all assets as borrowable under reasonable limits of supply cap, as we do not observe a significant risk to the protocol by allowing to borrow them, as long as it is bound by a well-defined cap.

CAKE

Of the above-listed assets, we recommend CAKE be listed in Isolation Mode, given its market cap and trading volume.

Following Chaos Labs’ Isolation Mode Methodology, we recommend an initial debt ceiling of $11M. Under the methodology for Isolation Mode, we consider two levels of probabilities for extreme price drops - Medium-High and High. We estimate the probability of an extreme price drop for CAKE as High. Given this debt ceiling, we do not identify a profitable attack vector under the current liquidity levels.

Borrowable in Isolation Mode

We recommend enabling USDC and USDT as borrowable in Isolation Mode

Supply Cap, Borrow Cap, and Liquidation Bonus

Following Chaos Labs’ approach to initial supply caps, as introduced with the Metis deployment recommendations, we propose setting the Supply Cap at 2x the liquidity available under the Liquidation Penalty price impact.

Given the observed liquidity of the suggested assets, we recommend the following:

Full Recommendations

We support the recommendations in the post for the Reserve Factor, Liquidation Protocol Fee, and Interest Rate curves. We will provide updated recommendations for IR Curves after publishing our methodology in the coming weeks.

Following the above analysis, we recommend listing the aforementioned assets with the following parameter settings:

Note on USDT and isolation mode:
Based on market risk and liquidity considerations, we support enabling USDT as collateral outside of isolation mode. However, it’s important to note that the primary concern, as mentioned in previous USDT listings, is the counterparty risk. Ultimately, the community must determine its stance on this issue and decide whether to enable USDT as collateral outside of Isolation mode. If the community is supportive of this, the discussion should be expanded to enabling USDT across the other Aave deployments too. Additionally, the community can choose to raise the reserve factor for USDT, which could serve as a form of compensation for the increased risk associated with USDT in comparison to USDC and DAI.

We once again would like to signal the inclusion of CAKE following on from the previous comments in the TEMP CHECK.

Hey guys, deploying Aave on BNB Chain seems like a great opportunity for growth and collaboration between both communities, while also providing opportunities not available on current Aave deployments. As pointed out by @PennBlockchain and @WintermuteGovernance Governance, introducing an Aave CAKE market would support the chain’s largest protocol while providing significant revenue potential.

Examining the CAKE market on Venus, the leading money market protocol on BNB Chain with a TVL of ~$1.7B, can provide valuable insights.

pp1920×1080 79.3 KB

Presently, on PancakeSwap, users have the opportunity to earn extra CAKE rewards by staking CAKE in a Syrup Pool for up to 41% APR in CAKE. Roughly 70% of the CAKE supply is staked in this pool with an average lock time of 43 weeks. Some investors adopt a strategy of borrowing CAKE on Venus and then staking it on PancakeSwap to capitalize on the rate differential.

[Cut-off from previous reply]

CRV and BAL are comparable assets to CAKE since both are DEX governance tokens with locking Mechanisms. Looking at Venus, a CAKE market on AAVE could potentially provide the treasury with 4x the margins (annualized revenues over average daily supply) earned on CRV and BAL. Anytime there seems to be outsized returns, we should check the risks (which fortunately @ChaosLabs is working on), but the opportunity is good.

Like Compound, Venus uses a jump interest rate model and recently reduced the kink on the CAKE market from .6 to .5, and increased the multiplier from .15 to .2. This had the effect of increasing the maximum borrow rate from 87% to 401% and the max supply rate from 20% to 235%. Since the parameter changes, utilization has only decreased by ~4% to 58%, revealing ~$6.7M in demand at a new average borrow rate of 36%.

However, demand at these levels is likely only sustainable with the current yield from staking CAKE. The APR for staking CAKE was 49% at the end of Q4, an 8% drop from current levels. In theory, a higher slope1 would capture elevated revenues while demand is there, with room for the strategy to still be profitable as staking yields fall.

Hi everyone, Uma from Succinct here. We’re very excited about this deployment and the community support for spreading Aave’s multichain reach to BNB chain! With this deployment, Aave governance (currently conducted on Ethereum mainnet) will need some way to communicate with the BNB deployment. Currently, the Aave governance cross-chain bridges repository (GitHub) does not have support for BNB-chain as there is no “native” message bridge for BNB-chain. With security and decentralization as two of Aave’s top priorities, we would love to support this deployment by adding our cross-chain messaging protocol, Telepathy, as a solution for bridging governance messages to BNB-chain.

More about Telepathy:
Telepathy is a trustless cross-chain messaging protocol that is secured by zkSNARK verification of consensus as opposed to multisigs. With Telepathy, protocols can send messages across chains without needing to trust centralized actors to deliver their message. Telepathy’s core innovation is a zkSNARK-powered light client that can directly verify the Ethereum light-client protocol on BNB chain, allowing for messages to be sent from Ethereum to BNB chain without requiring a sign-off from a multisig. Because of this, our protocol is significantly more decentralized and censorship resistant compared to existing bridging solutions. You can read more here: docs and website.

Telepathy as a “native” bridge:
Because of Telepathy’s significantly reduced trust assumptions, it can be treated as a “native” bridge from Ethereum to BNB chain. Across Protocol, a liquidity layer with over $1.6B in total volume, requires native message-passing bridges for their protocol. They view Telepathy as a native bridge to send messages from Ethereum to BNB chain, and have already integrated Telepathy into their protocol.

Our Support:
If this proposal passes, we would love to support the deployment by adding a PR to the current Aave cross-chain governance repo and make cross-chain governance between Ethereum and BNB chain as seamless and secure for AAVE holders as possible.

We would love to connect with someone on the Aave team to discuss current bridging plans and any questions you have, as well as look forward to engaging with the Aave community more broadly!

Gauntlet Risk Analysis

For our recommendations for the initial deployment on BSC, we consider a few main metrics for BNB, BTCB, WETH, USDC, USDT, CAKE, and BETH following our asset listing methodology.

Five of the six proposed assets have significant liquidity on BSC. BETH is the exception, with a 25% depth of less than $1M. We recommend that Aave V3 launch on BSC with BNB, BTCB, WETH, USDC, USDT, and CAKE as the initial assets, and wait to list BETH pending both the Shanghai upgrade and greater on-chain liquidity for the token.

A note on bridged assets

Some assets have multiple bridged versions on the BNB chain. For instance, there’s Binance-Peg USDC and also anyUSDC, both of which have significant on-chain market caps ($779M and $4.9M). We recommend the Binance-Peg version of all bridged tokens, as these are the versions with the highest circulating liquidity in all cases.

For the sake of clarity, here are all the token addresses we use in our analysis:

Recommended Initial Parameters

Borrow and Supply Caps

For initial supply and borrow caps, we use the lower of the conservative and aggressive recommendations generated under Gauntlet’s Borrow and Supply Cap Methodology.

*For USDT and USDC, our standard methodology would actually support higher caps than those in the proposal, but we cap them at 20M borrow cap and 50M supply cap.

Enable Borrowing

We recommend enabling borrowing for BTCB, ETH, USDC, and USDT. Conservative borrow caps are sufficient to mitigate the risk from borrowing.

Liquidation Threshold, Loan to Value, Isolation Mode

For BTCB, ETH, USDC, and USDT, we use the most conservative (lowest) LB and LTV of those assets on other chains.

BNB and CAKE are new assets. BNB, the native token of the BNB chain, is the 4th largest token by market cap, after BTC, ETH, and USDT, with a market cap of $49B. It has very significant on-chain liquidity, with a 25% depth of $65M. We recommend initially setting LT and LTV to 70% and 65% respectively, as a conservative starting point.

CAKE is the token for PancakeSwap, the leading DEX on the BNB chain. It is used for governance and staking. The CAKE token has substantial on-chain liquidity, with a 25% depth of $33M. CRV is a relevant comparable to CAKE, as both are DEX tokens with similar market caps ($686M for CAKE, $793M for CRV). CRV is listed on Ethereum Mainnet in isolation mode with LTV of 55% and LT of 61%. We recommend the same parameters for CAKE: isolation mode on, with LTV = 55% and LT = 1%. We recommend an initial debt ceiling of $10M for CAKE. Please see here for Gauntlet’s Isolation Mode methodology.

Liquidation Bonus

We recommend setting LB to 5% for stablecoin assets and 10% for non-stablecoin assets.

Reserve Factor

We recommend using the same Reserve Factors used on other chains for BTCB, ETH, USDC, and USDT. For the new tokens, BNB and CAKE, we recommend starting at 20%.

Interest Rate Curves

We recommend setting the interest rate curves for existing assets to the same values they have on other chains. For the new assets, BNB and CAKE, we recommend starting with the IR curves used for BTCB, as these are relatively conservative.

Borrowable in Isolation

We recommend allowing users to borrow the stablecoins, USDT and USDC, in isolation mode.

Supporting Data

Liquidity

25% depth figures reflect liquidity on Binance chain and are sourced from 1inch; we use USDC as the target currency for USDT, and USDT as the target currency for the other assets. 2% depth figures reflect liquidity across CEX and DEX sources and are sourced from Coingecko.

Supply

Thank you for the proposal and analysis on assets. BNB chain was haulted in the past via the a centralised entity, if such a situation were to arise again, what risks does this post to AAVE DAO and users of the protocol?

The ACI is presenting a consensus solution, taking into account the feedback from all parties involved in the discussion thus far.

Although BETH is treated similarly to other LSTs (and we already strongly stated our opinion on them), we acknowledge the concerns raised and will propose its integration at a later stage through a separate AIP. This approach aligns with the feedback received.

In addition, we recognize that incorporating CAKE as a “day one” asset may not be ideal. However, the suggestion to introduce it as an isolated asset serves as a suitable compromise. Therefore, the ACI supports its inclusion under these conditions in the current ARFC.

Consensus Solution

Onboarded Assets: BNB, BTCB, WETH, USDC, USDT, and CAKE.

BETH Onboarding Deferred until after the Shanghai upgrade and when improved on-chain liquidity is achieved, as recommended by Chaos Labs and Gauntlet.

Risk Parameters (LTV, LT, LB, Reserve Factors, etc.) & Interest Rate Curves as suggested by Gauntlet are adopted.

Ongoing Risk Parameter Monitoring: Continuously assess and update risk parameters as necessary, considering market conditions and input from all involved parties.

Updated Recap table:

If risk teams @ChaosLabs & @wfu (Gauntlet) agree on this consensus proposition, this ARFC will consider risk recommendation mature and no further discussion on this topic will be needed in the context of this ARFC.

Thanks @MarcZeller for putting this forward.

We suggest a couple of modifications to the consolidated suggestions provided above:

1. We recommend adopting the Chaos Labs recommendation for LT (80%) and LTV (77.5%) for USDC. The Gauntlet recommendation is based on the lowest configurations on other V3 networks, in this case, Ethereum. We set the parameters for USDC on Ethereum conservatively for the launch of V3, and we plan to recommend an increase soon. Therefore, there is an opportunity to launch here with higher figures.
2. Generally, when the risk providers recommend differing supply and borrow caps, we converge on the lower figures as they can continuously be raised. In this case, Gauntlet proposed lower caps for BNB and CAKE, while Chaos proposed lower caps for USDC, DAI, BTCB, and WETH. We think it would be prudent to adopt the lower figures for BTCB and WETH and are comfortable with the higher caps for USDC and DAI.

To summarize, we recommend the following consolidated parameters:

Thanks for your reply,

The proposed amendment to ARFC is accepted, the ARFC proposal has been updated to reflect this, and the checklist item is resolved.

This ARFC will now move forward when @bgdlabs will provide feedback with their evaluation report & cross-chain messaging solution.

Is @bgdlabs currently working on this? I’m looking foward to their evaluation report!

Although I have my personal reservations when it comes to BSC and to which extent it is decentralized, I do see it as a strategic place for Aave to deploy to and grab a decent market share.

Actively working on the report, expect it really soon

the report has been posted. Checklist is now complete.

The ACI will sync with @bgdlabs and escalate this proposal to next governance stage.

Upon reevaluation of the proposal and market data, we confirm our initial recommendations, with one amendment - we propose a 5% decrease in BNB LT and LTV to 70% and 65% respectively.

Furthermore, we’d like to highlight that while our supply and borrow cap methodology allows for caps higher than those outlined in the original proposal, we recommend launching with the more prudent recommendations. These caps are sufficiently high for the initial growth phase of the deployment and can be easily increased utilizing the Risk Steward process.

A fair point you raise here.

After re-analyzing the risks associated with deploying Aave V3 on the BNB chain in context of current market conditions, Gauntlet affirms its original parameter recommendations. Liquidity, market depth and token concentration have not substantially changed for the relevant tokens since the original analysis. We agree that placing CAKE in isolation mode is a good balance between risk management and participating in the BNB ecosystem. As we mentioned in our collateral recommendation, Gauntlet has not quantified the centralization risk associated with USDT so this recommendation abstracts from any risks around a USDT de-peg. Further, given how market conditions have improved around USDC, we recommend increasing USDC LT to 80% in line with our recent proposal. We believe these settings lay a good foundation for Aave to expand onto BNB with the understanding that they are conservative and meant to mitigate risks at inception. Once V3 is live on BNB, we will reevaluate the parameter recommendations in line with our Borrow and Supply Cap methodology to maximize risk-adjusted value accretion to the Aave community.