LlamaRisk, the risk service provider for Aave, has recommended winding down Aave V3 deployments on six networks: Sonic, Scroll, zkSync, Metis, Soneium, and Aptos. The proposal, posted July 29 as an ARFC (Aave Request for Comment), argues that support costs for these markets exceed the revenue they generate. The plan would put $4.1 million in debt on a staged exit path.
The scope of the proposal
The ARFC covers 25 lending reserves across the six networks, with a total of $12.8 million supplied, based on LlamaRisk data from July 28. Separately, the same ARFC targets 50 individual reserves and 21 matured Pendle principal tokens across 11 deployments, representing $85.3 million supplied and $11.5 million borrowed. That second batch is not part of the six-market exit but is included in the same forum discussion.
Why the exit is being considered
The economic case is straightforward: revenue doesn't cover costs. Sonic, Scroll, and zkSync each generate less than $5,000 in quarterly revenue. Metis, Soneium, and Aptos each bring in less than $1,000. The proposal cites oracle, monitoring, and operational support costs but does not put a dollar figure on the shortfall. LlamaRisk's recommendation is to wind down these markets before they become a net drain on the protocol.
How the wind-down would work
The plan starts with freezing every reserve on the six networks. Supply and borrow caps get cut to 1. For reserves that have debt, the reserve factor jumps to 99% and the base variable borrowing rate is set at 5%. Unborrowed reserves don't get those rate changes. The 99% reserve factor means nearly all interest flows to the Aave treasury, leaving almost nothing for suppliers. LlamaRisk expects lower yields to encourage withdrawals, which raises utilization and pushes borrowers to repay. The freeze stops new supply, new borrowing, and use as collateral, but existing positions stay open. Further unwind steps would be handled case by case.
Later steps could include raising interest rate curves or reducing liquidation thresholds for selected collateral. Deployment oracles could be replaced with fixed-price adapters. Users who remain after the first step face possible rate, collateral, and oracle changes.
Current status and next steps
The forum thread was still under discussion as of July 31. The starting point: every listed reserve on Sonic and Aptos is active, while every listed reserve on Scroll, zkSync, Metis, and Soneium is already frozen. Aave's proposal lifecycle moves from ARFC to a community Snapshot vote, then to an on-chain Aave Improvement Proposal. No timeline has been set for the next stage. The community is still debating whether to proceed with the wind-down or adjust the terms.
