/

Aave Narrows V3 Core as Borrowers Put V4 to Work

WEEKLY DIGEST

Aave Narrows V3 Core as Borrowers Put V4 to Work

Aave Narrows V3 Core as Borrowers Put V4 to Work

Aave’s proposed V3 Core collateral overhaul highlights the risks of leveraged staking trades, while V4’s growing loan demand and Sentora’s RLUSD-heavy hub show how borrowing is evolving across its markets.

Aave’s proposed V3 Core collateral overhaul highlights the risks of leveraged staking trades, while V4’s growing loan demand and Sentora’s RLUSD-heavy hub show how borrowing is evolving across its markets.

Sentora Research

Sentora Research

On V3 Core, two borrowers drawing WETH from the same pool pay the same rate even if one posts a liquid asset and the other posts a token whose value depends on a restaking strategy. A borrower who repeatedly pledges that token and borrows more ETH keeps the extra staking yield. If the token falls against ETH and cannot be sold quickly enough, the suppliers who funded the loan bear the shortfall. Their rate does not distinguish between the two trades.

Therefore, LlamaRisk’s November 10 proposal shows why the unwind matters as much as the choice of assets. As of November 9, rsETH and osETH backed $862.3M of WETH debt, much of it in positions leveraged 12 to 14 times. In LlamaRisk’s stress test, a 2 to 5 ve percentage point cut in liquidation thresholds would make most of that collateral liquidatable at once. A ten-point cut for LBTC, by contrast, put less than $100,000 of collateral at risk of liquidation.

The distinction is not just ETH versus BTC. Rocket Pool’s rETH backed $35.4M of debt against $81.8M of collateral, mostly for stablecoin borrowers. That is 43 cents of attributed debt per dollar of collateral, against roughly 92 cents for osETH. Rocket Pool asked for a separate assessment. Its objection illustrates what the blue-chip label leaves out: how borrowers actually use an asset can matter as much as who issues it.

The Snapshot vote endorsed stopping new exposure before changing the terms on existing borrowers. Borrowers could repay, refinance on V4 or substitute collateral that stays on Core. A narrower collateral list need does not mean a smaller lending pool. On November 19, risk stewards recommended lifting Core’s USDC deposit cap from $3.5B to $5B. Deposits had filled 99.6% of the cap, but borrowers were using only 74.3% of the USDC. There was $896M available to borrow. What Core lacked was room for another large depositor.

Where V4 Borrowing Is Growing

On August 22, the markets in Aave’s V4 held $496.3M of deposits and $174.0M of loans. By November 19, deposits had tripled to $1.51B and loans had grown 3.42 times to $595.5M. That difference matters: if deposits had grown without borrowers, a larger market would have left a smaller share of each dollar earning interest. Instead, the share on loan rose from 35.1% to 39.4%.

The pattern held across the governance window. From November 5 to November 19, loans grew 16.3% versus 13.5% for deposits. Roughly three quarters of the $83.2M increase in loans had already appeared by November 12. The balances cannot tell us which borrowers refinanced from V3, which were new to Aave, or how much of the change was due to token prices.

The Sentora Hub on Ethereum

Sentora as an Aave v4 curator sets collateral and risk parameters for its Ethereum hub under permissions Aave governance can revoke. Between November 9 and 19, deposits grew from $100.9M to $170.9M, while loans rose from $41.9M to $68.6M. Over the full fortnight the hub accounted for about a third of the increase in outstanding V4 loans in the same series. 

All $68.6M of debt was denominated in RLUSD, against $74.2M of RLUSD supplied. That makes the share on loan 92.5% for RLUSD, versus 40.1% for the hub.