Aave is preparing to remove dozens of underused lending markets in one of the DeFi protocol’s larger housekeeping efforts to date. A governance proposal filed on July 29 recommends deprecating 50 individual low-adoption reserves and winding down six full deployments, together covering roughly $98.1 million in supplied assets and $15.6 million in outstanding debt, as part of a portfolio-wide effort to reduce the protocol’s risk surface.
The changes are not yet live. They are set out in an Aave Request for Final Comment (ARFC), a late-stage governance post that precedes a binding on-chain vote, authored by DeFi risk service provider LlamaRisk.
Aave Founder Stani Kulechov summarized the proposal publicly on July 30. In a post on X, he shared, “After a comprehensive review, Aave is deprecating 50 low adoption asset reserves across multiple deployments. In addition, Aave is orderly winding down deployments on Sonic, Scroll, zkSync, Metis, Soneium, and Aptos, covering another 25 asset reserves.”
What the proposal covers
Per the ARFC, the deprecation splits into two parts.
The first is a set of 50 individual reserve removals, plus 21 matured Pendle Principal Tokens (PTs), fixed-maturity tokens whose lending markets serve no further function once they mature, spread across eleven Aave V3 deployments. Together these hold $85.3 million in supply and $11.5 million in debt, according to LlamaRisk’s figures.
The second is the full wind-down of six smaller deployments, which include Sonic, Scroll, zkSync, Metis, Soneium and Aptos, that adds another 25 reserves holding $12.8 million in supply and $4.1 million in debt. Combined, the two parts reach the $98.1 million supply and $15.6 million debt figures Kulechov cited, and touch 75 reserves in total alongside the 21 matured PTs.
A separate, companion ARFC handles a further group of long-tail reserves flagged for elevated Chainlink price-feed risk. Those assets are counted and specified in that document, not this one.
Why Aave is doing it
LlamaRisk frames the removals as a portfolio-level exercise under the Aave Risk Framework rather than a reaction to any single failing asset. The reasoning it gives is operational cost: every listed reserve carries a fixed maintenance load, an oracle to maintain, risk parameters to monitor, and a liquidation path that must work reliably, regardless of how small the market is. Where a reserve’s activity no longer justifies that load, the proposal winds it down.
Most reserves in scope are assets whose usage has stayed below, or fallen under, the level the framework requires for a standalone listing. The proposal also captures several structural cases: bridged tokens such as USDC.e and USDbC are removed where the native asset is already listed, so the same token is not carried twice; MaticX is included because its issuer, Stader, is sunsetting the token; and the 21 Pendle PTs are past maturity.
The six deployments being retired
For the whole-market deprecations, LlamaRisk’s stated rationale is that deposits have fallen to a point where the revenue each deployment generates no longer covers the cost of running it.
The proposal reports six-month declines in deposits of 74% on Sonic ($28.9M to $7.6M), 86% on Scroll ($16.1M to $2.2M), 88% on zkSync ($7.2M to $844k), 79% on Metis ($1.4M to $297k) and 95% on Soneium ($3.2M to $173k), alongside a 94% drop in available liquidity on Aptos ($18.0M to $1.0M). Each of the six, per the document, now produces under $5,000 per quarter in protocol revenue, and under $1,000 on the three smallest.
The largest positions in scope
On the reserve-by-reserve side, the two biggest balances are Bitcoin liquid-staking wrappers on Ethereum Core: FBTC at $11.1 million and eBTC at $5.3 million in supply, which LlamaRisk says were listed for collateral demand that did not materialise at scale. Their supplied balances have fallen from roughly $38.9 million and $33.5 million respectively over six months, according to the proposal.
The single largest in-scope balance is a block of matured Pendle PTs on the Plasma deployment at $32.2 million, which the document says now only await withdrawal.
How the wind-down works
The mechanics are the same across most reserves. The default action is to freeze the market, which blocks new deposits, new borrowing and use as fresh collateral while leaving existing positions open, and to reduce its supply and borrow caps to 1, effectively stopping new activity. On reserves that carry borrowing, the proposal also raises the reserve factor, the share of borrow interest routed to the Aave treasury, which lowers supplier yield and encourages withdrawals. For the six full deployments, every reserve is frozen with caps cut to 1 and, where borrowing exists, the reserve factor is raised to 99% and the interest-rate model’s base rate set to 5%.
LlamaRisk notes that a large share of the reserves in scope are already frozen, borrow-disabled or capped, meaning the proposal in many cases formalizes a wind-down already underway. Positions already open in affected markets can remain, and any effort to unwind lingering collateral positions is to be handled case by case.
What happens next
The proposal is at the ARFC stage and would need to pass a subsequent on-chain governance vote before the parameter changes take effect. LlamaRisk discloses in the document that it is a DeFi risk service provider funded in part by the Aave DAO, states it is not directly affiliated with the protocols reviewed, and says it received no compensation from them for the assessment.
Also Read: Uniswap Activates v4 Protocol Fees as Adams Rebuts LP-Cut Claims
