Aave’s markets are moving again after the Oct. 29 emergency freeze triggered by the Binance hot-wallet exploit, but the reopening comes with strings attached that every borrower and lender needs to understand. On Oct. 30, risk provider LlamaRisk confirmed it had executed lower borrowing ceilings and temporary interest-rate changes across Aave V3 and V4, restoring access to markets that had blocked all new supply and borrowing for roughly a day. Withdrawals, repayments, and liquidations stayed enabled throughout the freeze.
The catch is collateral. Under the rollback plan, USDT, USDC, USDT0, POL, WAVAX, sAVAX, XAUt, PAXG, and OP, including wrapped and bridged versions, now carry zero collateral borrowing power for new positions. On V3 that means a zero loan-to-value setting that also applies in E-Mode; on V4 the equivalent is a zero collateral factor. You can still supply, withdraw, borrow, and repay these tokens once market restrictions lift, but they cannot secure new loans.
Why the freeze happened at all
The trigger was the exploiter behind the 6.3 billion USD Binance theft. LlamaRisk’s initial assessment found the exploiter held a position on Aave that was overcollateralized and subject to normal liquidation, which meant no immediate bad debt for the protocol. But the position included a 19,762 WETH loan, and the risk team moved fast to contain any leverage the attacker might build before being liquidated. Hence the emergency freeze on Oct. 29, followed by the structured reopening.
How the new caps were sized
LlamaRisk’s methodology starts with the larger of borrowing before the exploit and borrowing at observation time, then adds roughly 2%, with rounded ceilings per asset. It targets borrowing paths that had more than 10 million USD borrowed before the incident and more than 10 million USD of spare capacity. The result preserves room to return to pre-exploit borrowing levels while cutting off the headroom above them.
The numbers are large. On the biggest paths, USDT now has a listed cap of 2,560 million USDT against 2,504.7 million borrowed before the exploit, leaving 55.3 million of room. USDC shows a 2,270 million cap against 2,215.9 million borrowed, leaving 54.1 million. Core USDC’s cap sat roughly 248.4 million above actual debt as of Oct. 30 because repayments had already pulled borrowing below pre-exploit levels. Core WETH’s ceiling of 1.79 million WETH covered then-current debt of 1,749,730 WETH.
LlamaRisk estimates the cap cuts remove roughly 3.8 billion USD of unused borrowing capacity in total. Queued changes on the GHO stablecoin would cut Core’s ceiling from 150 million to 105 million GHO and Prime’s from 85 million to 35 million, a combined 95 million GHO reduction.
Liquidity is the other constraint
Caps are only half the story. Aave withdrawals are limited to available, unborrowed liquidity, and during the freeze some reserves were nearly drained. At 00:26 UTC on Oct. 30, LlamaRisk put Core USDC utilization at 97.8%, with several other reserves displayed at 100%. A supplier trying to exit a fully utilized pool waits for repayments or new supply to free cash, no matter what the caps say.
To relieve pressure, LlamaRisk also executed temporary Slope2 reductions, lowering the steep portion of the borrowing-rate curve above optimal utilization. The idea is to slow interest growth for borrowers near full utilization without changing base rates. One wrinkle: while LlamaRisk described a universal 10% borrowing ceiling as the objective, its own modeled table lists 20% for Core DAI, 34% for Core USDtb, 25% for Monad mUSD, and 19% for Ethereum Core frxUSD, and Plasma GHO’s second slope appears as 5.25% in one summary and 11.5% in the specification. The execution update does not clarify which settings shipped, so borrowers should check their specific reserve rather than trust the headline number.
What changes for existing positions
The rollback plan unfolds in four waves, and only the V4 collateral-configuration wave was explicitly marked as executed in the Oct. 30 update. On V3, liquidation thresholds remain unchanged, so the zero-LTV reduction does not alter existing health factors by itself. On V4, the mechanics are stricter: any new borrow, collateral withdrawal, or collateral toggle moves an existing position onto the zero-factor configuration.
Aave’s V3 documentation adds one more condition that will surprise some users: zero-LTV reserves must be disabled as collateral or fully withdrawn before you can withdraw other assets from the same account. In other words, holding newly zeroed tokens can lock the rest of your portfolio until you deal with them.
The takeaway
The freeze and staged reopening show both the strength and the friction of decentralized lending governance under stress. Aave avoided bad debt from the Binance exploiter and preserved system solvency, but the cost is a market where major stablecoins and several altcoins no longer work as collateral for new borrowing, and where the emergency settings shipped with inconsistencies borrowers have to navigate themselves. If you hold affected assets on Aave, check whether your reserve’s caps, rates, and collateral settings have actually changed before assuming the old rules still apply.
The detail nobody mentions: on V3, zero-LTV reserves must be disabled as collateral or fully withdrawn before you can pull other assets from the same account. People are locked out of their own portfolio without realizing.
And on V4 it is stricter. Any new borrow, collateral withdrawal, or toggle moves an existing position onto the zero-factor config. Sitting completely still is the only safe move until waves finish.
the v4 trap is nasty. touch your position once, even a toggle, and the zero factor config grabs the whole thing
Core USDC at 97.8% utilization during the freeze is the scarier number. Caps mean nothing when the pool is drained. Slope2 cuts soften it but you still wait on repayments to exit.
LlamaRisk described a universal 10% borrowing ceiling but its own table lists 34% for Core USDtb and 25% for Monad mUSD. Check your specific reserve instead of trusting the headline number.
97.8% utilization while caps were still on means the unfreeze is theater. everyone rushed the exit and the pool is nearly empty
reopening the markets but keeping usdc and usdt at zero ltv is basically trust me bro lending lol
zero LTV on USDT and USDC collateral is wild. the safest assets on the platform treated like memecoins after one hack
its temporary until they re-risk, but yeah if you had a leveraged POL position you woke up to liquidation city
withdrawals and repayments stayed on the whole time at least. imagine the bank run if those had been blocked too