Spark Shuts Down SparkLend Protocol on Gnosis Chain
Spark announced on September 14 the closure of the lending instance SparkLend deployed on the Gnosis Chain. Following the execution, all outstanding loans on this instance will immediately enter a liquidatable state, and users must repay and withdraw collateral before the deadline.
This is not a sudden shutdown but the final step of a phased exit. In February of this year, Spark froze all new deposits and loans on the Gnosis market and unified the reserve factor to 50%. The assets named at that time included WXDAI, sDAI, WETH, wstETH, EURe, USDT, USDC.e, GNO, and USDC. Existing positions could still be retained, and the execution on September 14 changed the status from "new deposits frozen" to "existing positions can be liquidated."
SparkLend is a fork of Aave V3 and was launched in September 2023 as the first cross-chain deployment on Gnosis. The governance cited low usage of this instance, minimal reserve factor income, and ongoing operational and risk monitoring costs. Public data previously showed that the vast majority of the protocol's supply and borrowing was concentrated on Ethereum, with Gnosis accounting for less than 1% for a long time, representing a high-maintenance, low-contribution marginal market.
The shutdown mechanism operates according to existing liquidation rules: a health factor below 1 can be called for liquidation by a third party, with a health factor between 0.95 and 1 allowing for a maximum of 50% of debt to be liquidated, and if it drops to 0.95 or below, it can be fully covered at once, with liquidators receiving discounted collateral. Therefore, positions that do not actively repay after the deadline do not need to wait for prices to drop; the protocol parameters themselves will turn the entire pool of positions into targets for liquidation.
At the same time, the focus of products has shifted back to the main market. Recently, Spark upgraded its lending products to an isolated market based on Morpho and continues to obtain liquidity from stablecoin issuance as a sub-DAO of Sky. Shutting down the Gnosis instance is a move to shrink the multi-chain presence to a level that can generate reserve income and collaborate with institutional custody and isolated markets on the Ethereum mainline.
In market mechanics, this is an active contraction rather than a run-driven one. The seller is the protocol governance, which changes the parameters to shift the marginal market from "available for borrowing and lending" to "only an exit channel remains"; the buyers are borrowers who must raise repayment assets within four days and professional liquidation bots waiting for liquidation discounts. Funds move from collateral and debt on Gnosis to stablecoins or base assets used for repayment, which are then withdrawn off-chain or migrated back to Ethereum. The beneficiaries are Spark's risk budget and operational bandwidth; the pressure is on leveraged accounts still tied to this instance and the local stablecoin circulation in Gnosis that relies on this pool's depth.
On-chain, after the freeze in February, new traffic has stopped, and the focus on September 14 is whether existing debt has been repaid within the window. Unpaid portions will concentrate into the liquidation queue after execution, with liquidation bonuses becoming a short-term arbitrage entry.
Source: Public Information
ABAB AI Insight
After the Maker system turned Spark into a sub-DAO of Sky, the path has been to "use stablecoin issuance rights to supply blood to blue-chip currency markets". Moving SparkLend to Gnosis in 2023 was the first stop in the Endgame multi-chain narrative, initially seen as a model for leaving single Ethereum settlement. Two years later, the same governance has shut down this station, indicating that cross-chain expansion has not rewritten the income structure: Ethereum still contributes nearly all supply, while Gnosis only retains operational, oracle, bridging, and reputation tail risks. Similar contractions have previously occurred on inefficient collateral, with tBTC, ezETH, etc., being phased out, and Gnosis has upgraded "asset downgrade" to "entire chain instance downgrade".
The capital movement is very direct: freezing new deposits, raising the reserve factor, and then opening the forced liquidation window is equivalent to using parameters to draw back attention and monitoring costs sunk in inefficient pools to the main market. The motivation is not that Gnosis immediately incurred bad debts, but that reserve factor income cannot cover ongoing maintenance, and oracle and bridging events on marginal chains can reverse-pollute the main brand. Money and people are turning to the Morpho isolated market, Anchorage third-party custody, and blue-chip pools of USDS/USDC on Ethereum, which is exchanging "multi-chain presence" for "auditable institutional credit channels".
In comparison to Aave's multiple shutdowns of long-tail deployments and Compound's liquidity incentive contractions on some side chains, the model is consistent: first using forks to gain coverage, then using TVL and reserve income for subtraction. In terms of industry position, on-chain lending has moved from "multi-chain storefronts" to a stage of "main chain concentration + isolated market tail control". Spark is still expanding institutional borrowing and stablecoin circulation, but for instances like Gnosis with less than 1% contribution, control has shifted from the growth committee to the risk budget.
The structural change belongs to capital concentration. The mechanism is: the marginal safety costs of multi-chain are paid per chain, while marginal income is collected from the main chain pool. When governance can directly change reserve factors and liquidation switches, inefficient instances do not need to wait for liquidation to be closed; they will be closed in an accounting manner. Pricing power thus shifts from "where there is a market, there is a valuation" to "only markets that can generate extractable reserves and do not increase tail radius are retained".
ABAB News · Cognitive Laws
Coverage is not a moat; reserve income is.
Multi-chain costs are paid per chain, while profits are collected from the main pool.
Once the shutdown window opens, liquidation rights price faster than governance rights.