> ## Content Index > Fetch the complete content index at: https://news.curve.finance/llms.txt > Use this file to discover other available public pages before exploring further. # What Curve's H1 2026 lending data shows about liquidations under stress - URL: https://news.curve.finance/what-curves-h1-2026-lending-data-shows-about-liquidations-under-stress/ - Published: 2026-09-16T07:44:49.000Z - Updated: 2026-09-16T07:44:49.000Z - Description: H1 2026 put DeFi lending systems under sustained stress. Curve’s data shows how LLAMMA’s Liquidation Protection can give borrowers more time to respond, turning liquidation from a single cliff event into an ongoing risk-management process. - Author: Curve - Tags: Llamalend, Research ### **Key Takeaways** The H1 2026 data highlights three broader observations about how DeFi lending behaves during periods of market stress. - **Liquidations are driven by multiple forms of market stress** The first half of 2026 demonstrated that liquidation events are rarely driven by a single catalyst. Market declines, protocol exploits, liquidity shortages, and infrastructure failures each created distinct episodes of stress that affected borrowing markets in different ways. As lending markets become more interconnected, liquidation systems increasingly need to operate reliably across a wider range of scenarios than simple price volatility alone. - **Liquidation Protection give borrowers more time to respond** Curve's data indicates that a borrower’s position does not necessarily close when liquidation begins. Under LLAMMA, loans can remain open while collateral is progressively converted. As prices move through the liquidation range, the direction of these conversions can change. This can rebuild the borrower’s collateral exposure, although it does not necessarily restore the position to its original economic value. During H1, Curve’s lending markets saw 704 Soft Liquidation episodes with a median duration of 14.5 days, illustrating that liquidation can function as an ongoing risk-management process rather than a single irreversible event. - **Borrower outcomes deserve greater attention alongside protocol resilience** Protocol solvency will always remain the primary objective of any lending system. Efficient liquidations protect lenders and reduce bad debt, making them an essential part of DeFi infrastructure. At the same time, borrower experience increasingly deserves consideration as lending markets mature. Metrics such as response time, recovery rates, and the economic cost of managing positions through volatility can complement traditional risk indicators and provide a more complete understanding of liquidation performance. **The findings have also been covered by* [***CoinDesk**](https://www.coindesk.com/tech/2026/09/08/hundreds-of-defi-loans-on-curve-stayed-in-liquidation-for-weeks-without-dying)**, which focused on how long Liquidation Protection can remain active while positions undergo Soft Liquidation, and how those positions can remain open after entering the liquidation range.* --- The first half of 2026 has tested nearly every part of the DeFi lending ecosystem. Falling asset prices, several high-profile security incidents, and infrastructure failures combined to produce repeated periods of market stress, forcing lending protocols to process large waves of liquidations while continuing to protect lenders and maintain solvency. From a protocol perspective, these events demonstrated that modern DeFi lending infrastructure has become increasingly resilient. Liquidation engines continued operating during periods of elevated volatility, helping prevent the accumulation of bad debt even as market conditions deteriorated. Liquidations are typically evaluated through the lens of protocol safety. The primary question is whether unhealthy positions can be closed efficiently before they threaten lenders or the protocol itself. This remains an essential requirement for any lending system. Yet borrowers experience the same events differently. For them, liquidation performance also depends on what happens to their positions after they approach or enter liquidation, and what options remain available before collateral is permanently closed. This distinction has become increasingly relevant as DeFi lending continues to evolve beyond simple collateralized borrowing toward more capital-efficient strategies involving yield-bearing assets, liquidity positions, and increasingly sophisticated collateral structures. Using market-wide data alongside Curve's own lending statistics, this report examines how liquidation systems behaved during H1 2026, considering both protocol resilience and borrower outcomes. ## **Market Stress Put DeFi Lending to the Test** The first quarter of 2026 marked one of the most challenging periods for DeFi lending since the market recovery began in 2024\. Rather than being driven by a single event, stress accumulated through a combination of declining crypto prices, major protocol exploits, liquidity withdrawals, and isolated infrastructure failures that affected multiple lending markets simultaneously. According to Galaxy Research, the value of outstanding crypto-collateralized lending [declined](https://www.galaxy.com/insights/research/crypto-leverage-report-q2-2026-defi-lending-decline-futures-open-interest) by 16.78% during Q2 2026, falling to $56.16 billion. Outstanding loans across DeFi lending applications contracted even more sharply, decreasing by nearly 27.6% over the quarter to $20.4 billion, marking a third consecutive quarter of contraction. Security incidents added further pressure during Q2, triggering more capital outflows. One of the most significant catalysts in this regard was the LayerZero/KelpDAO exploit, which took place in April and had substantial downstream effects across lending markets. Within two weeks of the incident, more than $5.5 billion in stablecoin deposits left Aave, while approximately $3.1 billion in stablecoin borrowing was repaid. The protocol also experienced large withdrawals of wrapped Bitcoin and Ether as participants sought to reduce exposure amid heightened uncertainty. The consequences extended beyond simple declines in total value locked. As liquidity left lending markets, utilization rates rose sharply across major borrowing assets, temporarily making it more difficult for users to access liquidity or unwind positions efficiently. Galaxy's analysis found that WETH utilization remained above 99% for nearly thirteen days following the exploit, illustrating how quickly market-wide stress can constrain normal lending activity even when protocols continue operating as designed. Importantly, not every liquidation event during the period was caused by falling asset prices. In March, a configuration error affecting Aave's risk oracle [resulted](https://finance.yahoo.com/news/aave-liquidates-27m-oracle-glitch-113252143.html) in approximately $27 million of forced liquidations despite relatively stable market conditions. A temporary pricing discrepancy incorrectly identified healthy borrowing positions as undercollateralized, triggering automated liquidations before users had an opportunity to react. Although the protocol later confirmed affected users would be compensated, the incident highlighted another source of liquidation risk that receives comparatively less attention: infrastructure failures can be just as significant as market volatility. Taken together, these events illustrate three distinct categories of stress that increasingly shape DeFi lending markets: - broad market declines that reduce collateral values; - infrastructure failures, including oracle or pricing incidents; - protocol-specific events that trigger sudden changes in liquidity conditions. Each presents different operational challenges, but all ultimately test the same underlying question: how effectively lending protocols manage positions as borrowers approach liquidation. ## **Liquidation Efficiency Is Only Part of the Question** For any lending protocol, reliable liquidations are a fundamental safety mechanism. When collateral values decline, unhealthy positions must be resolved quickly enough to prevent losses from spreading to lenders or creating protocol-level bad debt. Without an effective liquidation process, lending markets become increasingly fragile during periods of volatility, ultimately reducing confidence in the protocol itself. Recent analysis published by Aave illustrates this point well. The protocol highlighted its ability to process approximately $4.6 billion in historical liquidations while maintaining protocol solvency and protecting suppliers. From a risk management perspective, this represents an important measure of operational resilience rather than simply a record of liquidation activity. Efficient execution remains one of the core responsibilities of any lending protocol. Yet protocol resilience and borrower outcomes are not necessarily the same thing. A liquidation that executes exactly as intended may represent a successful outcome for lenders while still imposing significant costs on borrowers. Once a position crosses its liquidation threshold, users often have limited opportunity to respond before collateral is sold. During periods of elevated volatility or network congestion, that window can become even narrower. This distinction suggests that liquidation architecture should be evaluated using more than a single measure of efficiency. Four dimensions are particularly relevant: - **Protocol solvency and bad debt:** whether liquidations resolve unhealthy positions before losses threaten lenders or the protocol. - **Execution reliability:** whether the liquidation mechanism continues to operate effectively during periods of volatility, liquidity stress, or infrastructure disruption. - **Time and options available to borrowers:** how positions evolve after entering liquidation and what opportunities borrowers retain to repay debt, add collateral, or otherwise manage their position as market conditions change. - **Borrower costs:** the economic impact of liquidation, including collateral losses, liquidation penalties, fees, and other costs associated with resolving or recovering a position. Together, these measures provide a broader framework for assessing liquidation performance from both the protocol and borrower perspectives. These considerations become increasingly relevant as DeFi lending expands beyond simple borrowing against passive collateral. Yield-bearing assets, liquidity provider positions, and other productive forms of collateral allow capital to remain active while securing loans, but they also introduce more dynamic risk profiles. ## **What Curve's Data Shows** This discussion becomes all the more meaningful when viewed through the actual borrower perspective and their results. This is where Curve’s data provides an important lens. Curve's lending system is built around LLAMMA, whose Liquidation Protection replaces the traditional liquidation cliff with a liquidation range. Typically, in lending protocols, once a position crosses the liquidation threshold, collateral is seized, and the transfer is final. LLAMMA provides this protection through Soft Liquidation, a transitional stage before a position reaches Hard Liquidation. As prices move through predefined liquidation bands, collateral is progressively converted into the borrowed stable asset while the loan remains open. 0:00 /0:28 1× How the collateral composition of LLAMMA bands changes based on the collateral price. Importantly, LLAMMA’s Liquidation Protection does not eliminate liquidation risk or guarantee a particular outcome. Instead, Soft Liquidation replaces what is usually an immediate binary outcome with a process that can give borrowers more time and flexibility to respond while their position remains active. Conversions within the liquidation range are not lossless, and Liquidation Protection does not guarantee that Hard Liquidation will be avoided. Observations collected across Curve's lending markets in H1 2026 show how Liquidation Protection works in practice through Soft Liquidation.. A third-party tracker [recorded](https://nikcrv.github.io/curve-liquidations-protect-dashboard/) 704 episodes in which this protection was active, involving 602 unique borrower addresses. Of these, 476 episodes began during H1 itself, affecting 402 borrower addresses and representing approximately $69.8 million in episode-level collateral exposure. ![](https://storage.ghost.io/c/84/bd/84bdf1ce-3211-4e56-85ff-6bf76cbb2657/content/images/2026/09/data-src-image-8c4a0f01-ad63-4f67-a106-2c6255bf7fcb.png) This data also shows that the duration of Liquidation Protection varied considerably across Soft Liquidation episodes: while the largest share of observed episodes lasted between one and seven days, a substantial number of positions remained under protection within the liquidation range for several weeks or even months. The median observed duration was 14.5 days, while the 75th percentile reached 38.9 days. Moreover, many of the longest-running episodes were still active when the dataset was captured, indicating that their final duration would ultimately be longer. All of this ultimately points to the fact that many borrowers’ positions remained within the liquidation range for highly prolonged periods rather than immediately progressing to a final outcome. To validate these observations, the Curve team independently reviewed 12 high-value Soft Liquidation episodes representing $15.2 million of tracked collateral exposure. In every case, the analysis confirmed that collateral was actively converted within LLAMMA, supporting the conclusion that the data reflected genuine Soft Liquidation activity. ## **How Liquidation Mechanisms Differ** Liquidation mechanisms differ across major DeFi lending protocols, particularly in how positions are resolved once they become eligible for liquidation. | **Protocol** | **Liquidation mechanism** | **What happens to the borrower’s position** | | -------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Aave V3 | When the health factor falls below 1, a liquidator repays part or all of the eligible debt and receives collateral plus a liquidation bonus. | Collateral transferred in an executed liquidation cannot be recovered. In the case of a partial liquidation, the remaining position stays open. | | Compound III | When an account is eligible for liquidation, the protocol absorbs the debt using its own reserves and the collateral is transferred to the protocol. The borrower receives the collateral's base-asset value, less the liquidation penalty. | The original collateral is not restored after absorption. Any remaining value is held in the base asset. | | Curve / LLAMMA | Entering the liquidation range activates Liquidation Protection through Soft Liquidation, progressively converting collateral while the loan remains open. | While Liquidation Protection remains active, the borrower retains time to repay or otherwise manage the position, while conversions can occur in either direction as prices move through the liquidation range. | Liquidation design continues to evolve across the sector: Aave V4, for example, is introducing a more granular partial-liquidation mechanism, illustrating how major lending protocols are developing different approaches to resolving unhealthy positions. At the same time, Liquidation Protection through **Soft Liquidation should not be interpreted as a lossless process or a guarantee that Hard Liquidation will be avoided**. Borrowers may still incur AMM fees, conversion and rebalancing losses, accrued interest, path-dependent effects, or eventually reach Hard Liquidation. Price movements can also change the direction of conversions without restoring the original economic value of a position. 0:00 /0:28 1× This loan continuously entered and exited liquidation protection and stayed in it for quite some time (around 4 hours). The user constantly monitored its health and repaid some debt as soon as health got closer to 0. 0:00 /0:05 1× This liquidation occurred during an extremely volatile market event. Liquidation Protection gave the user around 40 minutes (200 blocks) to act before the position was Hard Liquidated. Taken together, the data suggests that liquidation performance can be evaluated using more than a single end-state metric. Instead of treating liquidation as a single irreversible event, LLAMMA introduces a liquidation range in which positions can continue evolving as borrowers respond to changing market conditions. That additional dimension of “time” provides another way of assessing how liquidation mechanisms function during periods of market stress. Dive into the inner workings of Liquidation Protection [Go to Curve Docs ](https://docs.curve.finance/user/llamalend/liquidation-protection/how-it-works) ## **Conclusion** The first half of 2026 demonstrated that liquidation systems must remain resilient across different forms of market stress, from price volatility and liquidity shocks to security incidents and infrastructure failures. While efficient liquidations are still essential to maintaining protocol solvency and protecting lenders, as DeFi lending continues to mature, protocol safety alone is no longer sufficient to evaluate the quality of a liquidation system. Curve's H1 2026 data reinforces this perspective by showing that Liquidation Protection can turn liquidation into a process rather than a single event, and that protocol design can influence how positions evolve before they reach their final outcome. Under LLAMMA, borrowers entering the liquidation range receive this protection through Soft Liquidation: they retain time to respond while collateral is progressively converted, with conversions able to change direction as prices move through the liquidation range. This can restore collateral exposure, but does not mean the Soft Liquidation process itself was lossless. This demonstrates that liquidation design affects not only protocol safety, but also how positions evolve during periods of market stress. This approach is also what lies at the basis of the next stage Curve has taken in expanding its lending infrastructure. With the recent launch of [Llamalend V2](https://cryptobriefing.com/curve-llamalend-v2-optimism-launch/), Curve’s focus now extends beyond liquidation protection alone and toward creating lending markets that are more flexible and capital efficient. Support for productive collateral such as LP tokens, isolated markets with market-specific risk management, and more flexible collateral and borrow asset combinations are designed to bring lending and liquidity provision closer together while enabling users to put capital to work more efficiently. As DeFi lending evolves, protocols will increasingly be evaluated not only by how effectively they liquidate positions, but by how they help users manage risk before liquidation becomes unavoidable. Improving capital efficiency, expanding the range of productive collateral, and designing liquidation mechanisms that better accommodate periods of market stress are all part of that progression. Borrow now with LLAMMA [Go to Llamalend markets ](https://www.curve.finance/llamalend/)