[ARC] Launch Aave v3 on Celo

Hi all,

Guangye here from @Michigan_Blockchain.

Summary

Deploy Aave V3 on Celo to expand Aave real-world lending and green asset collateral to users around the world, with the provision of rewards to bootstrap liquidity.

Proposal

@Michigan_Blockchain and @PennBlockchain, in partnership with the Celo Foundation, propose the deployment of the Aave v3 protocol to Celo on behalf of the community. This proposal is providing background information on Celo and why Aave v3 should be deployed on Celo. Specific asset specifications and risks are not yet highlighted and will be done pending approval from the Aave community.

Celo is a mobile-first, carbon-negative, EVM-compatible blockchain. Specifically, we propose launching Aave v3 on the Celo platform to:

  • Bring Aave’s infrastructure to protocols on Celo focusing on real-world use cases and financial inclusion.
  • Expand Aave’s reach to the 6B smartphone users around the world utilizing the Celo blockchain’s mobile-first design.
  • Increase Aave’s global presence and bring in new users currently not utilizing web3 technology through
  • In the medium to long-term, create mechanisms for natural capital-backed assets such as tokenized carbon credit to serve as collateral for borrowing on Aave given Celo’s ReFi focus

Motivation

Currently, borrowing and lending on the Celo Platform (“Celo”) is possible with Moola and Chee Finance. Moola, the primary borrowing and lending protocol, focuses on CELO and Mento stable assets (cUSD, cEUR, and cREAL) and recently experienced a hack. There is strong demand in the Celo ecosystem for a blue-chip capital efficient liquidity protocol that supports CELO and Mento stable assets and in the future will add non-CELO assets (particularly ReFi assets). Additionally, a blue-chip liquidity protocol would support the recent deployments of Curve and Uniswap v3 on Celo to support overall DeFi activity. We seek to be purposeful and rigorous in identifying a need for deploying notable protocols onto the Celo blockchain. For Uniswap v3, it was a capital efficient DEX for natural capital assets that required concentrated liquidity. For Curve, it was a DEX with multi-coin pools for stable assets that would allow people and projects to better bridge on/off ramp onto the Celo blockchain (e.g. swap cUSD for USDT to offramp). For Aave, it is a borrowing and lending protocol built for long-term growth with an emphasis on security that will support current Celo-based asset needs and future non-CELO and Mento assets. This is crucial to bringing greater DeFi projects to Celo’s desktop and smartphone users around the world.

Celo focuses on real-world use cases of DeFi in which Aave can serve as a key building block, especially given Celo’s mobile emphasis which will bring in new Aave users. One example is utilizing Aave’s credit delegation feature. The Celo Foundation and Mercy Corps Ventures recently launched a pilot in rural Kenya utilizing credit delegation for employer-based lending. The Celo Foundation and Mercy Corps Ventures are currently looking to scale this solution as part of the Social Impact Collective and require a robust protocol—something we hope to specifically partner with the Aave community on scaling. Aave helps promote new use cases such as this, by allowing for crypto assets to be utilized in community-based financial models. Two other DeFi focused examples on Celo include GoodGhosting and Spirals, both of which have expressed interest in using Aave for savings and climate impact respectively for everyday users.

Additionally, with the launch of the Climate Collective, the Celo ecosystem features a growing group of ReFi projects (e.g. Toucan, Senken, Thallo, and Spirals). In the future, as these ecosystems grow through further engagement with carbon registries (e.g. Verra and Gold Standard), as well as define new carbon credit standards through the mass-coordination tool of blockchain technology, Aave will serve as a centerpiece for the growth of ReFi allowing users to access liquidity with their natural capital-backed assets. This in turn supports longer term holding of natural capital-backed assets, a more liquid marketplace for these assets, and growth of ReFi across web3. By creating a more robust and liquid ecosystem for natural capital assets, the Celo community has the vision of creating natural capital-backed currencies that would lead to a growth in preserved natural resources. To do this, Celo has a long-term goal of adding natural capital-backed assets to the Mento Reserve as part of the Climate Collective.

Overall, we believe Aave’s mission of creating the infrastructure for anyone to serve as a depositor or borrower aligns closely with Celo’s vision of real world DeFi and ReFi.

About Celo

Celo’s mission is to build an open financial system that creates the conditions of prosperity for all. To execute this mission, Celo’s technology features include:

  • Mobile First: A mobile-first EVM-compatible layer 1 blockchain that is accessible to anyone with a smartphone. Celo has a fast and secure PoS consensus protocol with 1 block finality that aggregates validator signatures using the SNARK-friendly BLS12-377 curve. This choice enables mobile friendly SNARK-based light clients such as Plumo 1, as well as SNARK proofs that can prove asset ownership off-chain. The network includes a decentralized phone verification protocol that can be used by dApps for lightweight sybil resistance and to make it easier for wallets to use hashed phone numbers as identifiers in an interoperable manner.
  • Assets: Gas can be paid with CELO or ERC20 tokens (that have been allow-listed through Celo’s on-chain governance). This would make it possible for Aave users to pay for transactions with Mento stableassets (e.g. cUSD, cEUR, cREAL). The CELO asset supports the ERC20 interface natively, which does away with the need for a wrapped counterpart like WCELO. The network includes a few stablecoins that have been allow-listed as a gas currency (cUSD, cEUR, and cREAL).
  • Regenerative Finance: Carbon negative by allocating 0.1% of epoch rewards to offset carbon—to date, 3,569 tons of carbon has been offset, which is over 8 times Celo’s carbon footprint.

Ecosystem: Celo is one of the fastest growing DeFi ecosystems with 3.9M addresses, up to 1M daily transactions, and over 174M transactions since the network launched. Celo activity is now back to levels during fall 2021 with ~500K daily transactions and TVL in CELO terms being higher than any prior period at 175M CELO. Additionally, Celo has been chosen by Kickstarter for its upcoming decentralized product (to date, Kickstarter has successfully supported over 214,000 projects with over $6B in funding and 20M backers) and Deutsche Telekom which will attract new users to the ecosystem.

Real World Use Cases: The Celo ecosystem utilizes TVL for real world applications. Several of these real world use cases include the following

  • UBI: impactMarket currently supports over 17,000 beneficiaries around the world.
  • Climate: Toucan protocol offers on chain carbon credit retirements. Toucan has launched their assets on the Celo blockchain and expects to support retirements in December.
  • Real world lending: EthicHub supports 21 smallholder farming communities in Mexico, Brazil, and Honduras for uncollateralized lending.
  • Commerce: Nuzo is a shop-to-earn and eCommerce marketplace in Kenya which supports over 40,000 users through stablecoin payments, offramps to mobile airtime, and reward tokens.
  • Native mobile offramps: Through FiatConnect, an API standard for web3 wallets to integrate with cash in/cash out providers, Celo users will be able to natively onramp and offramp through mobile wallets. The initial activation for Valora wallet users in ~11 countries is expected to happen in early Q1 2023.

Security: The chain consensus is Proof of Stake with validators taking delegation. There is no multisignature contract that can change the network. On-chain governance proposals and hard-forks are the only ways to make changes to the network. The Smart Contract was audited by OpenZeppelin and a Security Audit was conducted by Trailofbits. There is also a formal verification of Celo governance protocols by Certora

Bridge Infrastructure : Token bridges include Portal (Wormhole), Allbridge, and Satellite. Cross-chain messaging partners include Wormhole, Hyperlane, LayerZero and Axelar.

Rationale for the Aave community/ ecosystem: The Celo ecosystem is well suited for building DeFi applications due to its mobile compatibility and focus on building out real world DeFi and ReFi use cases. The Celo foundation has also allocated 0.5% of CELO circulating supply as incentive for Aave users to support the adoption and bootstrap growth on Celo. This deployment will bring new users to the Aave platform and reach more users from the Celo community. The Celo Foundation can also support Aave DAO in setting up a verified node for the Celo network (subject to technical requirements being met and voting by the community) for which Celo can provide a grant.

Next Steps

The first step to this will be a snapshot vote for the Aave community to determine deployment on Celo.

Aave operations rely on the Chainlink oracle and The Graph indexer. The Graph indexer is already available on Celo. The Chainlink SOAK test is completed and we are waiting for the Chainlink team for implementation. Pending approval from the Aave community and completion of Chainlink integration, Aave v3 will be deployed on Celo given that Celo is fully EVM-compatible. We believe that approval from the Aave community will help Chainlink prioritize deployment on Celo following their earlier communications on planning to deploy on Celo.

Upon a successful deployment, the Celo Foundation will launch a rewards program with 0.5% of CELO circulating supply of rewards to support the adoption and bootstrap growth of Aave on Celo. We see Aave as a core long- term partner, we propose a multi-year rewards program (3-4 years) rather than a large upfront volume of rewards. We believe that a high level of Celo rewards upfront will result in mercenary capital that damages the ecosystem in the long term, and in the current market will also lead to increased downward pressure resulting in reduced APR for Aave users on Celo. We’re open to feedback and thoughts from the Aave ecosystem on this.

The Celo Foundation can also support the Aave DAO to set up a verified node for the Celo network (subject to technical requirements being met and voting by the community) for which Celo can provide a grant. Additionally, the Celo Foundation will partner with the Aave community to build out real world use cases of Aave including a credit delegation platform for community-based finance.

Celo is ready to onboard Aave to its ecosystem and looks forward to the opportunity to build a long-term, sustainable, and mutually-beneficial relationship. Thank you for your consideration and we welcome questions and suggestions.

10 Likes

Excited to partner here and get the ball rolling on this!

4 Likes

Hi! Nikhil from the Celo Foundation here. Excited to get this proposal up to start the discussion with the Aave community. I know it’s holiday season, so look forward to continuing the conversation into the new year! We’ve also engaged with Chaos Labs regarding the risk parameters and can move forward on that when appropriate pending support from the Aave community.

In the meantime, feel free to reach out if you have any questions and happy to hop on a call if needed.

2 Likes

Hi! Daniel from Plastiks. Looking forward to see this happening!

Rachel from GoodGhosting here - would love to see this!

1 Like

This sounds like a no-brainer to me!

The Celo DeFi ecosystem is expanding with more and more “blue chip” protocols, and Aave is in a good position to become the go-to lending market on Celo. The current solutions are not ideal (e.g. an Aave v2 fork that uses insecure oracles), and the demand for lending and borrowing is there.

The recent launch of Curve on Celo, also enables the creation of pools containing aTokens. I expect this to provide rather sticky liquidity for USDC, USDT, DAI and (potentially) cUSD on Aave.

Fyi, that 0.5% of CELO circulating supply to bootstrap growth is currently worth ca. $1.2M. This is a nice little bonus, which I believe will attract more users to Aave.

Elio Lopes from IntoTheVerse here. Looking forward to see this happen :grinning:

Francis from GoodGhosting here. This would be great indeed. I look forward to see this happening.

1 Like

Hi Celo team - thanks for coming to the forum.

It seems you are “partnering” with Penn Blockchain which is great. Can you go into more detail about the scope of your partnership - and what this means?

Two questions I’d love to gauge:

  1. Are you willing to guarantee a minimum ongoing total borrow TVL on Aave?
  • this can be achieved by pledging assets from the foundation - or business development from your end, working with protocols already deployed on Celo

  • with limited TVL on chain (~$94mm according to DeFi Llama), this will pay Aave, recouping development and analysis costs incurred by BGD, Chaos, Gauntlet, et al.

  1. Can you protect Aave on a new protocol with an insurance fund due to low existing liquidity?
  • this helps ease some of the hesitancy around deploying on a younger, less liquid chain - which affects supply and borrow caps and associated risks for onboarding this network

We are aware of the benefits an Aave deployment will bring to Celo and hope to ensure it is a mutually beneficial investment for the DAO.

Thanks to conversations with @benhoneill from Chaos Labs for inspiring this framework!

4 Likes

Hi all,

Roman here from the Mento Labs team. As shared in the post by @GCao, Mento powers the stablecoins native to Celo that you might have seen around - that is the Celo Dollar (cUSD), Celo Euro (cEUR) and Celo Brazilian Real (cREAL). It is the highest TVL project on Celo right now with a TVL (= Mento Reserve Collateral) of ~$142mm, see DeFi Llama and the Mento Reserve Homepage for more details.

Mento is very much focused on enabling real world use-cases of stablecoins and we have worked on building out the required partnerships as well as the required tooling over the last few years. I believe I speak for the Mento community if I say that we would be super excited to see Aave launch on Celo. There is a large need for a battle-tested borrowing/lending protocol by our partners and integrated protocols and it seems to me that there is a large potential to both sides if Aave would fill that void.

Hi @fig, appreciate the questions! Answering several of them and will need to come back on a few items in the new year.

  1. The partnership with Penn Blockchain and the University of Michigan is working together on the proposal in a way that aligns with the Celo and Aave ecosystems. Both student organizations are very familiar with the Celo ecosystem and have been helpful in the proposal drafting process.

  2. The Celo Foundation has helped seed liquidity in the past. We’ve found seeding liquidity to at times be more helpful than rewards but would love to get your thoughts on a few follow-up questions:
    a. Is there preference for seeding liquidity over rewards given that seeding liquidity ensures that the liquidity remains for people and projects to be able to access? Or are both rewards and seeding liquidity something you’re looking for?
    b. Is there a rough range you’re looking for or is it primarily to show serious commitment to Aave on Celo?

The intention here is to ensure a successful long-term collaboration between Celo and Aave. We also have protocols that would use Aave including Mento and GoodGhosting, who have commented in this post. I can connect with these teams among others and come back.

Also, just want to note that there’s a mistake in DeFi Llama and that the Mento Protocol TVL on Celo is $140M as Roman notes in the comments.

  1. Let me connect with the Celo Foundation team on the insurance fund and get back to you.
2 Likes

Hi @fig, circling back on our thread here and hope you had a happy new year!

I had a chance to connect with our legal team. Unfortunately we cannot provide insurance funds both from a legal perspective as well as the decentralized nature of the Celo protocol. Happy to discuss further if needed. Let me know your thoughts though on the other items I mentioned above.

Hey @nraghuveera - thanks for the update.

Makes sense; we are grateful with your efforts and inquires with the internal teams.

Responding to #2, seeding liquidity seems of interest to our side - if you also mean guaranteeing borrow.

I will let the rest of the community chime in as it is not our sole discretion. The goal is to 1) to guarantee profitability for the technical costs incurred by the DAO via borrow 2) to show long-term alignment of Celo.

By seeding liquidity you achieve the supply side - but the revenue comes from the borrow demand.

Rewards may also work by helping subsidize interest rates for borrowing.


Tagging other community members for added detail:

@bgdlabs / @Llamaxyz for level of liquidity needed (aka revenue)

@ChaosLabs / @Pauljlei for risk considerations.

1 Like

Given that Celo is live for quite some time, studying the expansion of Aave to the network seems reasonable if the broad Aave community signals it.

@nraghuveera (or @PennBlockchain / University of Michigan) could you clarify which is the nature of the “partnership”?

2 Likes

Hey! Some good context on the nature of the partnership above. Launch Aave v3 on Celo - #12 by nraghuveera

1 Like

Hi all, hope everyone is well! Just an update that we’re communicating with various members of the Aave community. Let me know if you’d like to touch base and we haven’t yet!

Hi all, hope everyone is well! Just an update that we’re communicating with various members of the Aave community. Let me know if you’d like to touch base and we haven’t yet!

Gauntlet Risk Analysis

Simple Summary

Given the low level of DeFi activity, Gauntlet does not recommend that Aave deploy on Celo at this point in time. @ChaosLabs has reviewed this recommendation and has also arrived at the same conclusion.

Discussion

We evaluated the risk associated with deploying Aave on Celo by examining the liquidity of potential tokens and DEX liquidity/activity on the platform, as recommended by Gauntlet’s borrow and supply cap methodology. In particular, we analyzed factors such as:

  • overall total circulating supply of potential collateral assets on Celo
  • cost of incurring slippage on DEXes
    • It takes less than $1mil to incur 25% slippage on CELO, WETH, and WBTC.
  • liquidity and total volume of relevant DEXes

The picture painted by these numbers is also consistent with Aave’s asset risk methodology, which would assign a D risk rating to most of the assets we consider: CELO, cUSD, USDC, USDT, WETH, and WBTC.

Relevant Statistics

  • TVL: $126 million (as of 2023/03/07)

    The majority (68.7%) of Celo’s total value locked is on Mento.finance. This automated market maker (AMM) is primarily used for swapping Celo’s native stablecoins (cUSD, cEUR, cREAL) to and from CELO to help the stablecoins maintain their peg.

DEX Landscape

There are currently 4 decentralized exchanges on Celo (excluding Mento): Curve, Ubeswap, Uniswap, and Sushiswap. Their total value locked (TVL) is relatively low. Among them, Curve has the highest liquidity, primarily due to a cUSD/USDC pool with $20mil TVL and cUSD/USDC/USDT pool with $2.1mil TVL. All other liquidity pools have at most $600k TVL.

Exchange TVL
Curve $22.2mil
Ubeswap $3.16mil
UniswapV3 $1.59mil
Sushiswap $1.22mil

The top liquidity pools on Curve:

Ubeswap:

UniswapV3:

Sushiswap:

DEX Volume

The following table shows DEX volume over the past 1 day, 7 days, and 30 days.

All DEXs Ubeswap UniswapV3 SushiSwap
1D $147k $58.7k $77k $9.1k
7D $1.65mil $612k $892k $154k
30D $18.7mil $14mil $3.88mil $780k

Tokens

We show relevant statistics for CELO, cUSD, USDC, USDT, WETH, and WBTC, as these are the most likely assets to be used on an Aave deployment on Celo.

Token Circulating Supply ($) Net Top 5 Wallets ($)
CELO $432mil $131mil
cUSD $43.5mil $8.4mil
USDC $8.14mil $264k
USDT $575k $2.1k
WETH $1.58mil $171k
WBTC $596k $54.5k

Of the $8.14 million of USDC on chain, $7.2 million of the USDC tokens are in the cUSD/USDC Curve pool. Likewise, $573k of the $575k of USDT is in the cUSD/USDC/USDT Curve pool.

DEX Slippage

Token Cost of 2% DEX Slippage Cost of 25% DEX Slippage
CELO $22k $572k
cUSD $6.1mil $9.63mil
WETH $2.4k $33k
WBTC $20 $1.9k

We use USDC as the target token when computing the cost of slippage. Though $1 worth of CELO should be redeemable for ~$1 in cUSD, users incur 2+% slippage costs for any CELO to cUSD swap over $30k on Mento.

To avoid large slippage, users can make large CELO to cUSD swaps but must do so via the Granda Mento mechanism, which requires them to create a governance proposal that locks their to-be-swapped funds during the proposal while being voted on by the community. Because swaps for liquidations require timely execution, we do not regard swaps carried out through the Granda Mento mechanism as a viable source of CELO/cUSD liquidity.

Swaps between cUSD and USDC are more liquid than the other assets because there is a $20mil cUSD/USDC pool on Curve, which is also the largest TVL pool on all of Celo.

cUSD Risks

Apart from the liquidity issues with the cUSD token on Celo, we have reservations about the long-term price deviation risk of cUSD. Celo block explorer shows that there is $43.5 million worth of cUSD, $20.3 million worth of cEUR, and $1.12 million worth of cREAL outstanding, which are all backed by a reserve consisting of a blend of CELO, ETH, BTC, DAI, and USDC. Of the reserves, USDC and DAI holdings make up $48 million (according to the USDC and DAI wallets). Although the native stablecoins on Celo are currently overcollateralized, there is still some risk of price deviation given how much of Celo reserves are comprised of CELO tokens.

2 Likes

Hi @Pauljlei,

We really appreciate the time you have taken on reviewing this post! Let us take a close look and share some thoughts in the coming days. We noticed already that there are a few things we should dig into that may not have been completely accurate. We’re excited to discuss these further!

4 Likes

Hi @Pauljlei,

Following up on our previous post, we once again appreciate the time you have taken on reviewing this post! We noticed a few things that you may want to consider in your analysis:

Regarding assets, we have not yet proposed which assets should be included because we are still working through ensuring that there would be enough liquidity to meet liquidations according to your methodology. We were initially thinking of proposing a very small set (e.g. CELO, cUSD and possibly USDC) with conservative borrow and supply caps that match the existing on-chain liquidity.

We have been waiting on doing the temperature check snapshot poll until we have the necessary liquidity in place. One exciting development around this is that the Mento protocol is proposing to return 120M unused CELO to the Celo Community Fund, which the Celo community has expressed interest in placing in Curve and Uniswap to boost liquidity (pending a governance process). More details here.

Regarding the liquidity between the CELO and cUSD pair, one thing that you may want to consider in your analysis is that the constant product based Mento DEX resets every 5 minutes based on oracle updates. This means any slippage in the DEX resets every 5 minutes, and so the slippage is significantly lower if multiple liquidations are happening minutes apart. Additionally, sophisticated liquidators can use multiple transactions when liquidating a large position to avoid incurring large slippage, which is something that market makers are already doing in the ecosystem for liquidation bots targeting Moola (an Aave fork on Celo). Even if liquidators don’t do this, market makers are aware that the DEX is reset every 5 mins so they frequently arbitrage it back to near CEX rates whenever it deviates so they can profit when the reset happens.

When factoring all of this in, the Mento protocol can provide around $6M liquidity per hour to hit 25% slippage (and still be reset afterwards), and up to $144M in DEX liquidity per 24 hours (subject to CEX CELO liquidity). This could also potentially be increased a little as the slippage parameters are governable (pending a Celo governance proposal).

Regarding cUSD risks, cUSD is backed 1:1 with USDC and DAI. Previously, Mento was utilizing an overcollateralized model involving CELO, BTC, and ETH. This structure was changed last year to instead be a 1:1 backing with USDC and DAI with additional buffer collateral in ETH, BTC, and CELO. In your analysis, we believe you may have not noticed that the USDC that lives in a Curve position is owned by the Mento reserve. This was easy to miss and so the Mento team has recently noted it in the Mento Reserve website more explicitly. When you factor that in, and that the cUSD in that pool is destined to be burned as per this discussion, you will find that the three stablecoins are backed by 107% USDC and DAI. For these stablecoins, the current outstanding supply is a total of ~$52.22M with the following breakdown

  • cUSD: $30.8M
  • cEUR: $20M
  • cREAL: $1.1M

The Reserve holds ~$55M DAI and USDC to back this with the following breakdown:

  • DAI: $31.6M
  • USDC: $23.7M

We’re happy to jump on a call to discuss these further in detail. Thanks!

2 Likes