Intro
Liquidating LP collateral is not the same as liquidating a token balance. A standard lending market can often seize collateral, sell it, and restore solvency through a relatively direct path. LP backed borrowing is more complicated. The collateral may contain multiple assets, uncollected fees, concentrated range exposure, and venue-specific exit mechanics. If a position becomes unhealthy, the protocol has to do more than trigger a sale. It has to execute a controlled unwind. That is why Avana treats liquidation as an operational problem as much as a policy problem.
Context
At the policy level, liquidation begins when a position’s accrued debt exceeds its borrowing capacity. That is the line at which the loan can no longer be considered safely collateralized. But the real work begins after that threshold is crossed. The objective is not to seize everything as quickly as possible. The objective is to restore safety with the least necessary destruction of borrower value. Avana’s liquidation design is built around that principle. It prioritizes partial intervention, fee realization first where possible, and unwind paths that reflect the structure of the underlying LP rather than treating every position as a generic asset.
Model
In practice, liquidation is a runtime sequence. First, unhealthy accounts must be detected using the same risk-adjusted collateral values that the protocol itself relies on, not raw AMM spot state. Then execution liquidity must be sourced, often through a flashloan-style path, so the liquidator does not need to pre-fund the full unwind.
Value
Debt is repaid into the credit layer, the position is taken into controlled custody, claimable fees are realized where possible, and the LP is then unwound through the correct venue-specific path. Only after the underlying assets are recovered can they be routed into the debt asset, execution liquidity repaid, the liquidation premium distributed, and any residual borrower value returned. This is why LP liquidation is best understood as a controlled unwind workflow rather than a simple collateral seizure.
Risk
That distinction matters because different LP families do not share the same exit path. Fungible LPs are removed from the pool and decomposed into their underlying assets. Concentrated liquidity positions must be handled according to their actual inventory split at the time of unwind, which may be very different from what users assume if the position has moved mostly to one side. More custom or hook-based pools may require dedicated adapters before they can be considered safe enough for liquidation coverage at all. In Avana, that is not treated as an implementation footnote. It is part of the risk model. If the protocol cannot unwind a collateral type correctly, it should not pretend to support it safely.
Flow
A key principle in the design is minimal intervention. The protocol aims to restore target health rather than defaulting immediately to full liquidation. Partial liquidation is preferred because it reduces unnecessary borrower loss and limits disruption to the underlying pool. The framework computes the debt needed to restore a healthier position and targets that amount first. Full liquidation remains the fallback when solvency cannot be recovered through a smaller unwind, but it is not the starting assumption. This is one of the most important differences between a system designed around LP collateral and a system that treats liquidation as a blunt extraction event.
System
Another important principle is fees first. LP positions often contain claimable value that has not yet been realized. If fees can reduce the amount of principal that needs to be unwound, they should be used before the position is cut into more aggressively. That improves borrower outcomes and makes liquidation more economically precise. It also reflects the fact that LP positions are active financial objects, not passive token balances. A serious liquidation system should recognize the full structure of the position before it begins destroying principal.
Users
This is also why Avana Node and specialized liquidation runtimes matter. Liquidation remains permissionless, but LP collateral is harder for generic liquidators to service because the workflow is more involved. Nodes built for LP liquidation can watch active positions, refresh debt drift, index the relevant market state, simulate unwind paths, source execution liquidity, route assets correctly, and close the transaction atomically. In stressed conditions, this specialized execution layer acts as an important backstop against bad debt. The point is not to remove permissionless liquidators. The point is to make sure complex LP collateral still has reliable coverage when the market is least forgiving.
Outlook
Profitability and execution quality also matter more here than in simpler markets. Liquidators need to account for slippage, route depth, flashloan cost, and MEV exposure. Large or unusual unwinds may require private execution paths to avoid turning a recoverable liquidation into a destructive one. This is another reason Avana treats liquidation as an operator-facing discipline rather than a single function call with hidden assumptions. The policy layer may define when liquidation is allowed, but the runtime layer determines whether liquidation remains orderly in practice.
Takeaway
The broader lesson is simple. LP collateral is only credible if liquidation works under real conditions, not just in clean diagrams. A protocol has to know when a position is unhealthy, how much needs to be repaired, how to unwind the specific LP structure, how to route proceeds into the debt asset, and how to preserve borrower residual value once solvency is restored. That is the standard Avana is being built around. Liquidation should not be panic selling wrapped in smart contracts. It should be controlled resolution for an active onchain position.