Fewer than 9% of loans on Aave, the largest decentralized lending protocol, now hold roughly half of all the debt on the platform, according to an August analysis from Galaxy Digital. That small cohort is running close to the same trade: borrow against liquid staking versions of Ethereum, use the proceeds to buy more, and repeat. Galaxy’s August 7 snapshot counted 19,073 loans on Aave V3 Core after standard filters. The group using Aave’s high-leverage “E-mode” setting carried a debt-weighted loan-to-value near 90% and an average health factor of just 1.06, leaving almost no cushion. With Ethereum up 29% over the past week, the size of that bet matters more now than a month ago.
A liquid staking token is a tradable receipt, such as Lido’s wstETH or Ether.fi’s weETH, that stands in for staked Ethereum and the rewards it earns, letting holders keep ETH exposure while using the token elsewhere.
Key takeaways
- Galaxy Digital found that fewer than 9% of Aave V3 Core loans, all using the E-mode leverage setting, hold about half of the protocol’s outstanding debt.
- The cohort borrows WETH against liquid staking wrappers, a levered bet on Ethereum’s staking basis, with an average health factor of 1.06 and debt-to-equity near 10.7 times.
- Galaxy estimates an 8% to 9% discount between the wrappers and ETH could push the average position toward liquidation, even if ETH’s own price holds.
- A May model showed a 10% weETH depeg would drop 205 accounts below a health factor of 1, concentrating forced selling in leveraged ETH staking.
Published: August 22, 2026, 09:00 UTC
How the leverage loop works
The trade is a recursion. A borrower deposits a liquid staking token as collateral on Aave, borrows WETH against it, converts that WETH into more of the staking token, and deposits again. Each pass adds leverage, and the yield from staking compounds across the layers. Galaxy found that Ethereum staking and restaking wrappers, including weETH, rsETH, and wstETH, make up about 66.2% of the collateral behind these loans, with weETH alone near 42%. WETH accounts for roughly 73% of the group’s debt.
Aave’s E-mode is what makes the high leverage possible. The setting extends larger borrowing limits when collateral and debt are expected to move together, so a 90% loan-to-value between two closely linked assets can carry roughly the risk of a conservative loan between unrelated ones. A health factor is Aave’s safety score for a loan, and once it drops below 1, the position becomes eligible for automatic liquidation. The E-mode cohort sits at 1.06 on average. The other 91% of borrowers run a health factor near 1.79, far more able to absorb a bad day.
What could break it
The danger is not that Ethereum falls. If weETH drops 10% and the WETH borrowed against it falls the same 10%, the health factor barely moves because both sides slide together. The real exposure is the exchange rate between the wrapper and the ETH it represents. If weETH, rsETH, or wstETH begins trading at a discount to ETH while the debt stays fixed in WETH terms, the collateral weakens relative to what is owed. That can happen even if ETH’s price never changes.
Galaxy puts the built-in cushion for the average E-mode position at about 5.7% before it reaches the liquidation line. Applied across the 66.2% of collateral held in Ethereum wrappers, that points to a broad discount in the high single digits, roughly 8% to 9%, as the level that could push the cohort’s average health factor toward 1. A separate May analysis from the firm modeled a 10% weETH depeg and found it would leave Aave with about $2.47 billion in debt against $2.42 billion in post-shock collateral, dropping 205 accounts below a health factor of 1. Such gaps are not hypothetical. Lido’s stETH traded at a discount of roughly 7% to ETH during the Terra collapse in June 2022, a dislocation that was temporary but would have liquidated positions this leveraged.
Why the aggregate numbers hide it
Galaxy’s wider second-quarter report describes crypto lending as shrinking in an orderly way. Total crypto-related debt fell 15.08% from the prior quarter to $73.2 billion, a third straight quarterly decline. The debt split between Aave’s E-mode and standard loans ran near 60/40 in favor of E-mode in April and has narrowed to about 50/50, but only because E-mode debt declined, not because the concentration eased.
The result is a smaller number of positions still large enough to carry real weight. During the same stretch, Ethereum climbed hard, with a short squeeze pushing ether up 18% in a single day earlier in the week. A rising market keeps these loops comfortable and their yields flowing. The question is what happens on the way down, when the collateral has to keep trading like ETH for the trade to hold. It is the same collateral-quality problem behind recent DeFi stress events, applied to leverage rather than code.
Borrowers watching a discount widen can add fresh collateral or repay part of the WETH they owe. Anyone who does neither and slips below a health factor of 1 gets liquidated, a permissionless process where outside actors repay the debt and seize the collateral plus a bonus. For now the basis is holding and the loops are intact. The data from Galaxy’s research and coverage from CryptoSlate mark where the pressure would land first.
Frequently asked questions
What is Aave’s E-mode and why does it matter here?
E-mode is a high-efficiency setting that lets Aave borrowers take larger loans when their collateral and debt are expected to move together. It powers the leveraged Ethereum staking trade, allowing loan-to-value near 90% and leaving the concentrated cohort with an average health factor of only 1.06.
Could this trigger liquidations if Ethereum keeps rising?
Not directly. The risk is a depeg between liquid staking wrappers such as weETH and the ETH they represent, not ETH’s price. Galaxy estimates an 8% to 9% wrapper discount could push the average E-mode position toward liquidation even while ETH holds steady.
How much of Aave’s debt is exposed to this trade?
Galaxy’s August 7 snapshot found the E-mode cohort makes up fewer than 9% of loans yet carries roughly half of all outstanding debt on Aave V3 Core, with about 66.2% of that collateral held in Ethereum staking and restaking wrappers.
