LlamaRisk proposes Aave V3 cap changes after high Core utilization
LlamaRisk proposes Aave V3 cap changes, including increasing USDC borrow capacity from 2.7 billion to 4 billion, to address high Core utilization across multiple deployments.
USDC borrow capacity on Aave V3 Core leads LlamaRisk’s latest parameter filing: a proposed increase from 2.7 billion to 4 billion, adding 1.3 billion in nominal borrowing room. Published November 20, the Aave Governance Forum proposal also targets USDT, synthetic-dollar assets, bitcoin-backed collateral and WETH’s interest-rate model, attributing each recommendation to user behavior, on-chain liquidity and position health.
The November 20 filing has a predecessor. The filing follows a November 19 LlamaRisk proposal that addressed a different set of Aave V3 reserves.
Core market changes
LlamaRisk reported Core USDC borrow-cap utilization at 98.5%, against reserve utilization of 63.1%. Holding current borrowing constant, the implied amount borrowed is roughly 2.6595 billion; divided by the proposed 4 billion cap, that produces a post-change utilization of 66.5%.
Core USDT is tighter. At 93.8% of its 2.7 billion borrow cap and 85.6% reserve utilization, the proposal would raise the ceiling only to 2.8 billion. The same constant-borrowing math — 2.7 billion × 93.8% ÷ 2.8 billion — lands at 90.5% utilized. That’s meaningfully less breathing room than the USDC change delivers.
Three more supply caps would expand. USDe was at 91.5% of 750 million; a 1 billion cap would bring implied utilization to about 68.6%. sUSDe was at 98.3% of 210 million, and the proposed 400 million cap implies 51.6% on the same basis. WBTC was at 97.8% of its 36,000-unit cap; at 50,000 units, the unchanged-balance figure works out to 70.4%.
The filing also proposes cutting the Core supply cap for BTC.b from 40 to 30 and for eBTC from 100 to 85. Neither reserve has exhausted its current limit — BTC.b sat at 69.2% utilization (27.68 of 40 units) and eBTC at 79.5% (79.52 of 100) — but the lower ceilings would constrain new deposits regardless.
On WETH’s interest-rate model, LlamaRisk proposes moving variableRateSlope1 from 2.20% to 2.10%. At 83.6% utilization, the filing states the resulting variable borrow rate would be 1.91%, down from 2.00% under the current setting.
Cuts spread across deployments
The filing runs wide across deployments. The filing lists current borrow-cap utilization of 68.4% for USDT0, supply-cap utilization of 59.5% for USDe, 74.0% for sUSDe and 18.6% for WXPL.
Mantle receives cuts across five assets — syrupUSDT, sUSDe, WETH, WMNT and wrsETH — with current supply-cap utilization at 75.6% for syrupUSDT, 1.3% for sUSDe, 1.0% for WETH and 17.9% for WMNT. High- and low-utilization reserves land in the same tightening batch.
Other proposed reductions cover LINK and ARB on Arbitrum, BTC.b on Avalanche, LBTC on Base, WETH, weETH and ezETH on Linea, and xOKSOL and xSOL on X Layer. Utilization across those assets ranges from 7.5% for xOKSOL on X Layer to 74.0% for ezETH on Linea; Linea WETH sits at 40.4% supply-cap utilization and 70.2% borrow-cap utilization.
The proposal notes that rsETH and wrsETH are frozen on multiple Aave V3 instances — Core, Prime, Plasma, Mantle, Arbitrum, Avalanche, Base, Linea and MegaETH — and that tBTC is frozen on Core.
A frozen reserve accepts no new supply or borrowing. Existing positions can still be withdrawn, repaid and liquidated. Holders of affected frozen assets can’t open or increase positions; existing borrowers retain full repayment and liquidation exposure.
What changes for positions
The utilization data reflects reserves that were close to their stated caps: USDC borrow capacity at 98.5%, sUSDe supply at 98.3%, WBTC supply at 97.8%. If governance adopts the proposed settings, USDC and USDT borrowers alongside USDe, sUSDe and WBTC suppliers gain room to operate.
Until governance acts, every figure above describes a proposed setting. For related context on lending-market cap decisions, Compound’s proposal to zero supply caps on seven staking-token listings was covered previously here.