Aave wants to run a private credit desk for crypto institutions, and the way the DAO would fund it is drawing warnings. A September 24 proposal on the Aave governance forum would let institutions borrow dollars against Bitcoin and Ether held with a qualified custodian. The Aave DAO itself would finance those loans by borrowing stablecoins against a separate pool of its own crypto.
Aave Labs clarified the setup on September 30. An Aave Labs entity will sit on the other side of the contract as the legal lender. The DAO is asked to approve two funding routes: a $25 million issuance bucket for GHO, Aave’s own stablecoin, and up to $25 million of USDC or USDT borrowing against DAO assets. That is $50 million of total capacity for lending against BTC and ETH, not an obligation to lend it all.
Aave Labs reports roughly $300 million of indicated demand, including a $20 million lead BTC facility. Those are pipeline figures rather than commitments. The proposal does not disclose how much would actually be drawn at launch. Institutional interest in borrowing against custody-held crypto is real, as shown this week when Bitfinex moved 20,000 ETH worth about $53 million into Aave’s on-chain markets, but turning that interest into a fee business with borrowed money is a different trade.
Two collateral books, two separate debts
The structure puts crypto collateral on both sides of the trade. On the borrower side, an institution deposits Bitcoin or Ether with a qualified custodian, in the earlier V4 proposal Anchorage. That collateral secures the dollar loan under a Master Loan Agreement with the Aave Labs entity, and a three-party Account Control Agreement connects lender, borrower and custodian. Borrower collateral would never be rehypothecated.
On the funding side, the DAO would pledge its own WETH and WBTC, with AAVE permitted up to 50 percent of collateral at each pledge, and borrow USDC or USDT on Aave V3 to hand to the borrower. Two different asset pools secure two different debts. The DAO’s on-chain pledge has nothing to do with the borrower’s custody account.
That separation creates the uncomfortable scenario. A broad fall in crypto prices would squeeze both books at once. The customer’s custody collateral weakens just as the DAO’s pledged WBTC, WETH or AAVE loses value. The proposal itself flags the risk of AAVE weakening while BTC-backed loans come under stress. An institution might be paying on time, yet the DAO still has to top up its own funding collateral or repay, because its debt stands regardless of the loan book performing.
Borrower collateral cannot be treated as a backstop for the DAO’s position. Whether the lender could reach it depends on the security and enforcement terms of each facility, and the proposal leaves key details unpublished: the initial mix of DAO collateral, the size of its debt, health factors and the margin terms per facility. A record of enforcement under these arrangements is still to be reported.
When the spread disappears
The second problem is arithmetic. Aave Labs gives indicative borrower pricing of 6 to 8 percent APR against funding costs of about 4.5 percent, a spread of 1.5 to 3.5 points. Borrower coupons are fixed by contract and reprice only on 90 days’ notice. DAO funding pays the prevailing Aave V3 stablecoin rate, which moves daily.
Run the numbers and the cushion is thin. At 6 percent funding the spread is gone even when every borrower pays in full. At 7 percent the DAO pays out more to borrow than it earns on the loans. Losses without a single default are a real outcome, not a scare line.
Aave’s borrowing documentation adds its own pressure. A leveraged position needs a health factor above one, and collateral top-ups or partial repayment become necessary as prices fall. A decline that triggers margin calls for institutions can trigger them for the DAO in the same week. The custodian in the earlier proposal would handle institutional liquidations as an over-the-counter sale instead of dumping collateral on-chain, with proceeds settling the Aave position, but nothing of the sort protects the DAO’s own funding book, which is an ordinary leveraged position on Aave V3.
| Proposal term | Value |
|---|---|
| Total funding capacity | $50 million |
| GHO issuance bucket | $25 million |
| USDC or USDT borrowing limit | $25 million |
| Indicated demand | About $300 million |
| Lead facility | $20 million BTC |
| Borrower pricing | 6 to 8 percent APR |
| Funding cost | About 4.5 percent |
| Initial loan-to-value | 60 to 75 percent |
| AAVE share of collateral | Up to 50 percent |
Risk handling on the customer side sits with the custodian. It monitors the collateral, issues margin calls and liquidates when calls go unmet. Initial loan-to-value ratios of 60 to 75 percent mean a loan can survive a moderate drawdown before the liquidation machinery starts, though Bitcoin drops of 20 to 30 percent happen in this market, and higher leverage shortens the runway.
Aave founder Stani Kulechov framed the earlier custodied-lending proposal as a first of its kind: institutions can borrow against collateral held with a qualified custodian while tapping liquidity from Aave’s markets. The pitch is aimed at regulated firms that cannot drop Bitcoin into an ordinary on-chain money market. Anchorage, a federally chartered crypto bank, would hold the assets.
Governance comes next. The proposal moves through community feedback, a Snapshot vote and an on-chain AIP, with each funding authorization needing approval from two of three GHO Stewards. The finance committee, led by TokenLogic, would monitor the funding position after launch.
Timing is worth noting. The plan arrives during a downturn that has already forced layoffs at Anchorage, Gemini and Coinbase, and days after hackers used a third-party tool tied to Aave’s ecosystem to steal 114 ETH, about $540,000, in early October. Neither touches the proposal directly, but they sharpenboard-level scrutiny of anything that adds treasury risk.
The broader question is treasury risk. Aave’s DAO already holds large crypto positions, and this plan would gear them up. If the funding book can bleed money while every borrower is current, the community has to decide whether that trade is worth the fees. The vote will tell.