What did Spark, which made $1.3 billion against the trend after being shot in AAVE, do right?

When Aave faced a $10 billion massive withdrawal, Spark stepped in to absorb the massive liquidity.

The on-chain disaster triggered by the cross-chain漏洞 of Kelp DAO and LayzerZero tore the DeFi lending market into two clearly divided worlds.

“Poisonous” asset rsETH flooded into Aave, causing approximately $200 million in bad debt, with the entire network liquidity drying up and hundreds of billions of funds fleeing in panic.

However, amid the chaos, another lending giant, Spark, experienced its moment of glory. TVL (Total Value Locked) rapidly increased by $1.3 billion, and ETH deposit rates once soared to 130%, becoming the safe haven for whales to transfer assets.

A black swan event revalidated the rightful owner of the DeFi throne.

Aave Runoff Bloodbath, Spark Takes the Opportunity to Absorb $1.3 Billion

The cross-chain bridge of rsETH broke, and the entire Aave lending market was immediately paused.

Hackers used illegally minted rsETH to collateralize and borrow a large amount of WETH on Aave, draining the clean assets and leaving behind a pool of bad debt.

Panic spread rapidly like a virus: in just over three and a half days, Aave saw $15.1 billion in outflows, with total deposits dropping from $48.5 billion to $30.7 billion, about one-third of the funds fleeing; WETH utilization across multiple chains reached 100%; depositors could not withdraw, and liquidators had no funds to borrow.

The most notable move came from Sun Yuchen, who quickly withdrew 65,584 ETH, worth about $154 million, from Aave.

This “withdraw first” behavior created a herd effect in the market. For investors, even high annualized yields couldn’t offset the panic of being unable to withdraw principal.

As Aave became a liquidity exit for hackers, Spark became the escape route for users.

Spark’s TVL increased by $1.3 billion instead of decreasing, reaching a total of $4.74 billion. This money was a trust vote from the market, invested with real money.

Due to the surge in borrowing demand and highly scarce liquidity, ETH deposits on Spark experienced a spectacular spike, once soaring to an annualized rate of 130%, directly reflecting the high premium on safe assets.

Spark’s ability to absorb this demand is thanks to its unique ecosystem structure. Unlike Aave, it is the lending engine of the Sky ecosystem, backed by a large USDS reserve. As the liquidity front of Sky, Spark not only relies on external deposits but can also directly obtain stablecoin supplies through Sky’s credit lines.

This “central bank-level” liquidity backing allows it to always keep withdrawal channels open even during market shocks.

Abandoning the vanity of TVL, Spark removes rsETH from the platform

Spark avoided the rsETH disaster thanks to a contrarian decision made three months earlier.

On the same day, but with different fates. On January 29, two major lending platforms took opposite approaches to handling liquidity re-pledge tokens (LRT).

Aave went all out. The protocol officially launched rsETH E-Mode, allowing users to leverage borrow at a high collateralization ratio (LTV) of 93%. Aave’s goal was to attract an expected $1 billion in rsETH inflow, restore WETH utilization, and boost TVL and revenue.

Spark took a cautious retreat. The protocol, through governance action, Spell, completely halted new rsETH supply and gradually removed it from the asset list.

This move by Spark caused strong dissatisfaction among leveraged ETH users, who often repeatedly collateralized stETH or rsETH to arbitrage the spread. Spark’s delisting forced them to migrate their positions, most of which flowed to Aave, which had a more lenient policy and lower interest rates.

At the time, the community questioned whether Spark’s team was “too conservative” or “abandoning growth.” No one expected that this step might save the entire protocol later.

In retrospect, Spark’s strategic officer, monetsupply.eth, pointed out that the decision to delist rsETH was based on a safety-oriented tightening mechanism:

• Marginal cost vs marginal benefit: If maintaining a certain asset costs more than its risk-adjusted return to the protocol, that asset will be cleared;
• Concentration of risk exposure: rsETH utilization on Spark was extremely low, almost monopolized by a single wallet address, making risk difficult to diversify;
• User preference survey: The only whale user of rsETH expressed willingness to migrate to more mature collateral like wstETH or weETH, providing a smooth opportunity for asset cleanup.

It is this transparency and discipline in decision-making—“not blindly pursuing TVL”—that allowed Spark to avoid potential losses from hacker exploitation of rsETH.

Multi-layer risk control system: rate limits + interest buffers + isolation architecture

PANews believes that even without delisting rsETH, Spark’s architecture is sufficient to resist such risks. Compared to Aave’s pursuit of capital efficiency at the expense of safety redundancy, Spark has built a multi-layered deep defense system.

Spark implements strict rate-limited caps, meaning the amount of assets deposited and borrowed within a fixed period increases gradually. Even if rsETH had not been delisted, attackers could not deposit $290 million worth of collateral in a single event like on Aave. This design forcibly limits the maximum risk exposure of a single incident, constraining losses within manageable bounds.

Spark maintains relatively high interest rate ceilings over the long term. Under stable market conditions, higher borrowing rates discourage over-borrowing (expensive to borrow, so less borrowing), while attracting more depositors (depositors earn more). As a result, the pool always retains liquidity and won’t “run out,” preventing a liquidity crunch during market crashes.

When the utilization rate of the funds pool rises, Spark’s interest rate curve becomes steeper than Aave’s, leading to two significant consequences:

• Forced deleveraging: High interest costs push borrowers to actively seek liquidity to repay loans.
• Attracting additional liquidity: High annualized deposit yields quickly attract external arbitrage capital, breaking the deadlock of 100% utilization.

Spark’s modular isolation architecture offers strong controllability in risk management. When handling high-risk synthetic assets like USDe, Spark adopts a cautious approach, isolating them in specific primary risk pools. This ensures that even if a particular asset segment encounters issues, it won’t affect the main lending pools on the platform.

The liquidity migration from Aave to Spark signifies a shift in risk preference—from pursuit of high yields to prioritizing safety and stability.

Aave’s $10 billion outflow serves as a wake-up call to protocols chasing high capital efficiency. When safety margins are sacrificed, even minor external correlated risks can escalate into a global protocol crisis.