Simple summary
Proposal for activating more Chainlink SVR feeds on Aave Ethereum, given the successful activation of the first batch after the Phase 1 proposal and the second batch on the Phase 2 proposal.
Motivation
The initial two Phases of Aave <> Chainlink SVR activation serve as progressive steps to cover pricing of different asset types with the SVR system: low-size BTC-correlated assets and AAVE+LINK on Phase 1; expanding to all other BTC-correlated assets, plus all ETH-correlated on a smaller pool like Aave v3 Prime with Phase 2.
In parallel, as can be observed on the periodic updates by Chainlink Labs (like this last), the Chainlink/Flashbots off-chain infrastructure and adoption of it by market participants (e.g., more searchers) have been improving steadily, reflected in progressively higher recapture rate.
This led us to propose this new Phase 3, focusing now on all other ETH-correlated assets and USDC on v3 Core after the “soaking” of the feeds on all ETH-correlated on Prime.
This proposal is done via Direct-to-AIP, as no fundamental aspect of Aave <> SVR changes, only implies a pre-approved expansion to new assets.
Specification
In terms of high-level specification, this new Phase follows the same architecture as what was described in the initial SVR Phase 1 post.
Regarding the assets proposed, the full list and rationale are the following.
Aave v3 Ethereum Core
The rationale of this sub-set of new assets is:
- Cover all ETH-correlated currently listed and USDC, given that both ETH/USD and USDC/USD feeds have been in production for more than 1 month without any problem.
- Leave other assets like governance tokens (e.g., UNI, MKR), lower-size stablecoins (e.g., USDS), or assets with more sophisticated layers (e.g., PTs) for a follow-up proposal, as they require extra due diligence and setup effort, while not carrying too big a percentage of liquidations volume.
Additionally, USDT and Ethena-related assets are not included either, for a combination of the previously commented sophisticated layers, lack of liquidations, and finally because of their interdependence (USDe priced based on USDT/USD, which has synchronisation implications for price feeds).
Similar to Phases 1 and 2, the SVROracleSteward contract active for Core will receive support for the new assets, providing redundant protection and allowing for a seamless transition to non-SVR if any very edge scenario arises.
Next steps
After some days in this forum and while we finish our internal security review procedures, we will proceed with the creation of an on-chain AIP of Aave <> Chainlink SVR Phase 3.