The Liquidation Buffer: How DeFi Saver Helps Users Avoid Liquidation

At the end of January 2026, the price of Bitcoin and Ethereum declined by roughly 26% and 33%, respectively, in just six days.

Source: The Block

In the same period, Aave processed $429 million of liquidations across ~12.5K transactions, its highest since May 2021. Over the course of the liquidations, the Aave protocol did what it was designed to do. For borrowers on the other side of those transactions, however, that is precisely the problem.

Source: The Block

The Cost of Being Liquidated

For leveraged positions in a lending protocol, third-party liquidations are the most expensive way for a leveraged position to unwind. There are 3 main factors that cause this.

First is position size. On Aave V3, the default is that a liquidator can repay a borrower’s entire debt in whichever asset they target. A 50% cap, measured against the position’s total debt rather than one asset, applies while the position’s health factor sits above 0.95, the collateral and debt reserves involved each clear $2,000, and if the asset accounts for more than half of the position’s total debt. If any one of these factors is not fulfilled, the whole balance in that position can be cleared in a single call, as a partial liquidation must also leave at least $1,000 of both collateral and bad debt behind.

Second is the penalty. ETH collateral on Aave V3 carries a liquidation bonus of ~5%, seized from the borrower, that goes mostly to whoever repays the debt, with a portion also routed to the Aave treasury. Aave’s efficiency mode (E-mode) compresses that to 1% for tightly correlated pairs, while more volatile collateral is higher. Regardless, these rates are fixed in advance and applied in full irrespective of how long the position was underwater for.

Lastly, liquidity. A position that is being liquidated is essentially always sold into the thinnest, most unfavorable liquidity conditions possible. An example of this happened on February 5th, when on-chain depth at the time could not absorb standard close-factor sizes for a USDC loan backed by 355,093 AAVE, worth ~$28.4 million (about 2.3% of the token’s supply), so the position had to be liquidated in fragments of $500K to $1.5 million before Wintermute cleared the remainder in a single $10.1 million liquidation. In other words, secondary-market depth became the binding constraint on liquidation sizing.

With these factors in mind, a borrower at the brink of liquidation faces a choice of having to either pay a few basis points to defend their position, or hand over 5% of whatever is seized and lose the exposure anyways.

Why Manual Management Fails

A borrower manually managing a lending position has two levers, either hold a wide collateral buffer, which leads to capital inefficiency, or run the position tight and accept that any adverse move against their position while they aren’t paying attention might get resolved by a liquidator rather than by them.

In most cases, borrowers tend to take the first option, because the thought process, at face value, is that idle capital costs them close to nothing and is a far better option than facing increased risks of liquidation. However, the fact is that the first part of that option is not inherently true, as it does not take into account the opportunity cost of that idle capital, where it could be used in more productive avenues elsewhere.

A wide collateral buffer does not solve the rebalancing issue either, as positions can often drift across thresholds as interest accrues and collateral grounds lower with no rebalancing actions if the borrower forgets to do so.

Therefore, the gap that automation fills, beyond constant real-time monitoring and interventions, is being able to hold a ratio, meaning adjusting the position as often as the market moves it at a cost low enough to be worth doing each time. The tools below are built around that one function.

The DeFi Saver Toolkit

DeFi Saver’s tools are best understood by how they can make life easier for the user rather than the specifics of their on-chain actions. Set against the conditions described above, each one answers a different part of what makes a liquidation event expensive.

The trouble with every cost described above is that it lands at once the moment the health factor touches 1.0. One of DeFi Saver’s main tools is Liquidation Protection, which is built so that a position never gets there. On Aave v3, it is triggered when a position’s safety ratio breaks 105% and repays enough debt to lift the safety ratio 5% clear, to 110%, checked on a per-minute basis. 

Because all of that happens above the liquidation point, neither the close factor nor the ~5% penalty ever comes into play. It also means a user no longer needs to park stablecoins in a wallet to defend against a price drawdown at 3 in the morning, or to have to hold ETH across different chains for gas, or the luck required to land a transaction in the dire minutes approaching liquidation while many others are trying to do the same thing.

A crash or liquidation event also rarely arrives as a single dip. Aave v3 cleared $138 million of liquidations on February 5th and another $91 million the following day, so a position that survived the first down leg still had to survive the second.

Automated Leverage Management is the tool built for this, running the same repay logic against a trigger and target ratio that the user set themselves, as many times as the market breaks it. It works in the other direction as well, re-levering as the ratio recovers, though in a drawdown the value is less about the extra yield than about not needing to be present for each successive leg down.

Some positions are also better defended by leaving the protocol altogether. Aave v3 liquidates WETH collateral at a threshold of 83%, a fixed parameter that no borrower can negotiate, while Fluid’s main ETH/USDC vault sets the same threshold at 92% and charges a 1% liquidation penalty against Aave’s 5%. The Loan Shifter tool moves a position between the two in one flash-loaned transaction, while DeFi Saver charges no fees as long as the collateral and debt assets stay the same. For a position carrying $180K of debt against 100 ETH, that shift moves the liquidation price from $2,169 down to $1,957 without repaying a cent, which is ~10% more room on any position of that shape. If these actions were done manually, it would mean having to unwind the position and paying the slippage twice, while sitting uncollateralized in between.

Lastly, not every borrower will want to hand over execution of their positions. For them, the February 6th wick is the relevant problem, in that the decision gets made by an oracle inside a window they were never watching. The Notify tool covers cases like these, sending ratio alerts by in-app message, Telegram, email or even Blockscan chat at conditions the user chooses without taking any control over the position itself.

What Automated Defense Is Worth

To showcase what DeFi Saver’s automation tools are worth, we can take a look at a case that occurred between January 25, 2026 and February 9, 2026.

During this period, DeFi Saver’s automation had fired 382 repay actions across Aave V3 on Ethereum, Arbitrum, Base and Optimism. From those, 365 (~95.5%) were emergency interventions on positions already inside a 150% safety ratio.

Based on DeFi Saver’s report, its tools had defended over $446 million of at-risk collateral supporting $335 million of debt, with no protected position liquidated, and estimated avoided penalties at ~$8.38 million.

Looking deeper, one mainnet transaction on January 29th stood out, where against $7.16 million of debt, DeFi Saver’s automation swapped 697.49 WETH (~$1.965 million) and repaid $1.905 million. They charged a 0.3% fee, applied only to the swapped amount, costing just $5,896, while a standard liquidation at the 50% close factor with a 5% penalty would have taken ~$179K.

In other words, the automations saved about $173K, or 97% less than a liquidation would have cost, while the $5,896 fee worked out to just 8 basis points of the position size.

What This Means for Levered Borrowers

Adding automation into borrowed positions does not mean it reduces the risks of leverage, as repeated defensive repays sell collateral into weakness regardless.

Automation simply changes the failure mode as a position unwinds partially, on terms borrowers set in advance with relatively small fees, as opposed to being liquidated by a third party with less favorable conditions.

Readers inclined to test that now also have a way to do so. DeFi Saver recently released an Aave V3 position simulator that replays fixed positions against real hourly market history, and treats a position as liquidated if the hour’s lows would have breached it.