@specialist78 makes a fair point. The current plan treats every vault share the same way — turns all of them into recovery claims — even though most of the vault is still solvent. Healthy holders carry an impairment they did not earn.
The split he proposes is intuitive: keep normal vault operation for the solvent part, only convert the impaired part into a tradable claim. But this runs into the same problem the current plan has. To let the “normal” 76.8% withdraw, the vault needs free crvUSD on hand. There is none today — it is locked inside active borrows. Freeing it up means either forcing borrowers out (the DAO will not do this), waiting for them to repay on their own (which is what we already have), or making it attractive for them to repay early.
Making it attractive for borrowers to repay early means one thing: the recovery pool has to have depth and a price close to fair value. Right now the pool is thin and the price is far from fair, so no borrower has any reason to act. Without depth, neither the current plan nor the split plan works. Both need the pool to behave like a real market.
So depth in the recovery pool is the one thing that has to be solved first — and worth thinking about who provides it and on what terms. The pool currently trades at a meaningful discount to par. An LP entering today is buying impaired claims at a depressed price, collecting swap fees while the market thickens, and holding a position that recovers as bad debt is paid down and spot moves toward fair value. This is not a gift to anyone. It is a position with carry and recoverable principal. Protocol-owned liquidity in other Curve pools follows the same logic.
A question I would put to @specialist78 and to @LlamaRisk: does this look fair on both sides? The DAO seeding liquidity in the recovery pool is not a subsidy — the DAO retains the capital as a market position, earns fees, and exits at whatever spot the recovery delivers. Affected holders get a working market to exit through. Borrowers get a price signal that makes early repayment rational. If the alternative is to ask depositors (impaired and unimpaired both) to absorb the entire impairment while the DAO waits for organic liquidity that may not come, is that the more honest split of responsibility? Or is a recoverable LP commitment from the DAO closer to a fair arrangement for everyone — including the DAO itself?
Picking up where the depth question is sitting. @saintrat (#3), @Hubert (#4), @specialist78 (#20), and @mo957qx8 (#23) all converged on the same point from different angles — without depth the pool is theoretical. The diagnosis is shared. The forward question is what actually thickens it on a timeline that matters.
Some on-chain news that is relevant here, separate from any DAO-side decision. Reading GaugeController as of block 25016303, the gauge 0xF429AeC167C92aCA16cD77aef54F196B1988cBA3 has its first non-zero weight scheduled for the 2026-05-07 epoch — 0.45%. That weight comes mostly from one voter (0x9B44…B029, ~27.6M veCRV, 12.62% of votes cast for this gauge so far).
That makes the next ~14 days a natural test of the “investors will arrive” hypothesis from the OP. Total CRV emissions are ~$74K/day across all gauges, so 0.45% weight = ~$334/day flowing into a pool currently around $1,760 of depth. Early money gets the fat APR before TVL fills and dilutes it. External LPs deciding on yield numbers do not need any DAO assumption to act on this.
If real depth shows up by 2026-05-21, the borrower side could front-run incentives — that is a valid concern and the dashboard LlamaRisk suggested in #17 becomes worth building. If depth does not show up, the design question reopens with cleaner data than we have today.
Sidebar — there are three live bad-debt threads in parallel right now (this one, 11029, the WFRAX market). Different cohorts, different mechanisms. Curve DAO has no playbook for these — they each become an ad-hoc 3-month forum debate. That is an observation, not a proposal.
@LlamaRisk just posted a proposal at sDOLA-long2 Repayment Funds Source Proposal to fund the sDOLA-long2 reimbursement (~$822K target, 27 affected borrowers from the March oracle exploit) from ~$1.14M of stranded L2 admin fees that have not yet been burned or distributed. Source-wise these are by-design DAO-revenue.
A question for everyone who has been following the CRV-long situation in this thread.
The CRV-long market has been distressed since October 2025. The recovery pool is operational since Vote #1400 closed on 2026-04-30, but it currently sits at around $63K TVL against ~$754K of bad debt and has no funding source attached. The sDOLA-long2 incident happened on 2026-03-02. A fully-scoped funding proposal landed on the forum two months later, drawing on the same kind of DAO-revenue source.
CRV-long has been waiting seven months. The new proposal does not mention this market or this thread anywhere in its text.
Is that fair?
Not asking rhetorically. Honestly interested in how people who have been tracking the CRV-long market see it. Should the same source carry both cases in some shape, or is there a principled reason one LlamaLend market gets fully funded from common DAO-revenue while a parallel one with a longer-running and comparable-size loss waits indefinitely on arbers and external LPs?
Weighing in on the /t/11069 thread directly on the scope question would be useful, since the proposal is moving toward vote with no acknowledgement of CRV-long in its text.
Not disclosing that a market carries bad debt is a serious failure. This information should be clear and visible to everyone — that way each person can assess the risk before depositing.
A recovery pool with no liquidity and no clear timeline is not a real solution for affected users. In practice it increases losses rather than reducing them.
@stryk asks exactly the right question and I would genuinely like to hear the answer — what is the principled reason CRV-long is being treated differently here?