Community Trust ScoreVerified
Aave wants out of six chains. The protocol’s risk management service LlamaRisk has put forward a proposal to wind down V3 markets on Sonic, Scroll, zkSync, Metis, Soneium, and Aptos — pulling the plug on $98.1 million in supplied assets and $15.6 million in outstanding debt in the process. It’s a big cleanup, and the numbers behind it are hard to argue with.
LlamaRisk, working alongside other Aave service providers, put together a proposal to offboard 50 low-use reserves and 21 matured Pendle principal token listings. Eleven separate deployment areas are caught up in the plan. The document filed is an ARFC — basically a detailed precursor to a formal Aave Improvement Proposal — so it lays out the full picture of what’s being suggested without locking anything in yet. No onchain vote has happened. No execution. Just a very clear signal of where the protocol wants to go.
The Aptos numbers are rough.
Liquidity on Aptos has dropped 94% over six months. Quarterly revenue there came in below $1,000. The V3 market on Aptos only launched 11 months ago, which makes the exit pretty striking — less than a year in, and the numbers already look like a write-off. Reserves on Scroll, zkSync, Metis, and Soneium have already been frozen, and LlamaRisk is recommending the same treatment for Sonic and Aptos to bring them in line.
Where the $2 Million Floor Came From
None of this came out of nowhere. Back in December 2025, a temp check on Aave’s governance showed overwhelming support for two specific changes: raising reserve factors on underperforming instances, and setting a $2 million annual revenue floor for any new deployments going forward. That vote basically set the table for everything happening now. Then in April, an accelerated process added Scroll to the list of affected protocols, moving fast enough to push through a direct proposal to freeze all Scroll reserves without waiting for the slower standard governance track.
So by the time LlamaRisk filed the current ARFC, the direction was already pretty well established. The community had already said it wanted to tighten standards. The freeze on Scroll, zkSync, Metis, and Soneium was already done. What’s left is getting community sign-off on applying the same logic to Sonic and Aptos, and formally offboarding the 50 reserves and 21 Pendle listings flagged in the proposal.
Aave founder Stani Kulechov weighed in on the broader picture. Per Kulechov, the moves are meant to cut economic and technical risk in line with the new Aave Risk Framework and Technical Asset Listing Framework. He was also clear that Aave’s multichain expansion isn’t being reversed — it’s being refocused. Continuous risk assessments across all assets and deployments will stay in place. That’s the line, anyway.
Avalanche Launch and the Refocusing Argument
Aave recently launched on Avalanche, which makes the “refocusing, not retreating” framing easier to sell. The protocol isn’t pulling back from multichain. It’s just getting more selective about where it shows up and how long it stays if things don’t work. The updated risk framework — released in June — spells out criteria around asset risk, bridge risk, monitoring, and chain risk, and it includes specific guidelines for when to wind down reserves or full deployments. That framework is probably what made the Aptos decision feel inevitable once the liquidity numbers came in.
The revenue floor is a real change in how Aave thinks about expansion. Before, the protocol was fairly aggressive about deploying on new chains, probably because the cost of doing so seemed low. But running a market that generates less than $1,000 per quarter while carrying millions in supplied assets and the associated technical overhead — that’s not really a free option. It’s a slow drain. The $2 million annual floor forces any new deployment to clear a bar that the Aptos market never came close to clearing.
And it’s not just revenue. LlamaRisk’s framework also looks at bridge risk and chain-level monitoring capacity. Some of the smaller or newer chains Aave deployed on don’t have the same infrastructure depth as Ethereum mainnet or major L2s. That matters when things go wrong — and in DeFi, things go wrong.
Community Vote Still Pending
The proposal is still waiting on community approval before anything gets executed. Aave’s governance structure means stakeholders have to sign off, which is both a check on the process and a potential slowdown. Given how the December temp check went, opposition seems unlikely — but unclear yet whether the timeline will be fast or slow. The ARFC stage is just the beginning of the formal path.
Fifty reserves and 21 Pendle principal token listings on the chopping block. Six chains. $98.1 million in supplied assets being retired. And a revenue floor that probably means any future deployment has to prove itself a lot faster than Aptos did.
The Aptos market, for the record, generated less than $1,000 in quarterly revenue.
Hub: AAVE price, news, and analysis
Frequently Asked Questions
Which blockchains is Aave planning to exit under this proposal?
Aave’s proposal covers six chains: Sonic, Scroll, zkSync, Metis, Soneium, and Aptos, with all V3 reserves on those networks flagged for retirement.
Why is Aave leaving Aptos so quickly after launching there?
Aptos liquidity fell 94% over six months and the market generated less than $1,000 in quarterly revenue — well below the $2 million annual floor Aave’s community endorsed in December 2025.
