Insights from the DeFi Crash: Thinking About the Future Multi-Chain Landscape from a Liquidation Perspective

This article is from the Baihua Blockchain, with the original title: "The Crash Brings a Great Test to DeFi, Possibly Signaling the Future Direction of Multi-Chain Patterns for Ethereum, Polkadot, and Others."
You must still vividly remember the disaster on May 19; for those in the circle, it is hard to forget this day, as it can be compared to the epic waterfall of March 12.

You may have also heard about the awkward situation XVS faced on May 19, where XVS experienced an "epic" liquidation, and this liquidation was essentially almost unrelated to the crash on May 19.

Here's what happened: It is well known that Venus allows XVS to be collateralized, with a borrowing rate of 80% for assets like BTC or ETH, while the liquidity of XVS is actually quite poor compared to mainstream coins. A certain whale discovered this "loophole" and executed the following operations:

1. Spent tens of millions of dollars in a short time, pulling the price of XVS from $70 to $140.
2. Collateralized XVS at a high price to borrow 4,000 BTC and tens of thousands of ETH.
3. Made off with the funds.
4. Let the price of XVS rapidly decline and began liquidation.
5. Due to insufficient liquidity, XVS could not be liquidated in a short time, causing the platform to incur over a hundred million dollars in bad debts.

According to a community member, over a million XVS waiting to be liquidated were listed at low prices, and it took the liquidators (most likely bots) several hours to slowly buy them all.

This led us to ponder: In the current multi-chain universe, can each chain truly operate a complete DeFi financial system independently? Including exchanges, borrowing, stablecoins, and synthetic assets.

Given the current state of blockchain, the answer is likely "suspended."

01 Ubiquitous Liquidation

Liquidation is something you usually wouldn't notice at all because it rarely happens. However, when extreme market conditions arise, liquidation is everywhere.

Take the four basic components of DeFi: exchanges, borrowing, stablecoins (specifically referring to DeFi-native crypto-collateralized stablecoins like DAI, LUSD, etc.), and synthetic assets. You will find that, apart from exchanges, the other three have liquidation mechanisms as their core components.

In simple terms, no liquidation, no DeFi. Each extreme market event, such as March 12 or May 19, can be seen as a test of the liquidation mechanisms of various projects and even the overall "health" of the DeFi system.

Different platforms have different liquidation methods; for example, MakerDAO uses an asset auction mechanism, while Compound and AAVE allow liquidators to directly take over the debtor's position at a discounted price. The recent stablecoin newcomer Liquity LUSD has even set up a triple liquidation mechanism (stability pool, debt warehouse transfer, global liquidation). Interested friends can look it up themselves; I won't elaborate here.

However, regardless of the liquidation mechanism, it cannot escape these three basic steps:

1. Liquidated assets are sold at a discount.

2. Liquidators buy the discounted debt.

3. Liquidators sell the debt they bought at a discount.

It is not difficult to see that in the entire liquidation system, the role of the liquidator is the core of the core.

02 A Simple Review of the 312 MakerDAO Liquidation Failure

Like miners, liquidators are also a group of people driven by economic incentives. After all, the liquidated assets are sold at a discount, as long as you grab them and sell them within a short time, it is almost a guaranteed profit.

They constantly monitor projects like MakerDAO, Compound, AAVE, and Liquity for assets that meet liquidation conditions or are close to liquidation, and when liquidation occurs, they swarm in. It's like a group of vultures watching a dying rabbit, waiting for it to fall so they can see who dives in faster and grabs it more accurately.

You can probably guess that this kind of thing cannot be done by humans with the naked eye; it is certainly countless pre-written bot programs that are constantly monitoring.

Ethereum is a dark forest, where not only various arbitrage bots lurk but also various liquidation bots are always on standby. Many liquidation bots even use advanced tools like flash loans for zero-risk liquidation (buying discounted assets with their own money and then trying to sell them at the original price, which carries the risk of "insufficient funds" during rapid price drops).

So why did MakerDAO end up with a zero-dollar price winning the auction on March 12, accumulating millions in bad debts? The main reason is that at that time, ETH congestion caused gas fees to exceed 1000 gwei (1 ETH = 10^9 Gwei), leading to two consequences:

1. In MakerDAO's auction mechanism, the most congested were many bids above zero that could not be packed due to insufficient gas fees.

2. More critically, many liquidation bots stopped under such high gas conditions. This is understandable; bots are profit-driven, and when gas fees are too high, exceeding the asset's discount price, or when the gas cost of each failed liquidation is too high, many bots will "shut down."

This is similar to how the price drop of BTC below the "shutdown price" of many mining machines leads to miners shutting down. It is said that on March 12, only three liquidators participated in the MakerDAO auction, and each (or bot) won the auction at a price of zero.

03 The Liquidation Depth Issue Exposed by XVS

March 12 exposed the issue of MakerDAO encountering "liquidators buying discounted debt" on ETH, while May 19 exposed the issue of XVS encountering "liquidators selling the debt bought at a discount" on BSC.

In simple terms, when the $80 XVS was discounted to $70 and bought by the liquidator, the liquidator found that due to slippage and depth issues, if the volume was large, they could only sell it for $65 when they tried to sell it later, which became awkward…

This was also the original plan of the whale who took advantage of XVS. With enough XVS in hand, wanting to offload it, and the market depth unable to absorb their "dumping," they collateralized XVS to the system, effectively selling all their XVS at 80% of the highest price (the collateralization rate of XVS), while the buyers were all the liquidators plus the bad debts caused by the inability to liquidate.

This sparked much discussion, especially since XVS, as an asset with relatively low liquidity, has an 80% collateralization rate that is too high, which triggered this bad debt incident. If the collateralization rate were 50%, these issues might not have occurred at all.

Alternatively, if only high liquidity assets like BTC and ETH were accepted as collateral, such problems could also be avoided. So the question arises: even if only BTC and ETH are accepted as collateral, will liquidation still be an issue in the multi-chain universe?

04 Liquidation and the Endgame of Blockchain in the Multi-Chain Universe

You see, liquidators or liquidation bots are essentially driven by profit.

So you need to ensure that after they buy the discounted collateral, they can immediately find a platform (usually a DEX) to sell these collaterals at a price higher than their purchase price, ensuring the liquidators make a profit.

Previously, we only had Ethereum, and we had Uniswap; we never worried about the depth issues of WBTC, ETH, and some mainstream DeFi tokens, so whether it was Compound, MakerDAO, or AAVE, no one worried about what problems might arise in the step of "liquidators selling the debt bought at a discount."

However, the XVS incident has reminded us that even if we only accept BTC and ETH as collateral, in the multi-chain universe, do we have enough BTC and ETH?

You see, we now have ETH, EOS, TRX, BSC, HECO, Solana, Fantom, Cosmos, Polkadot, Avax… and these are just the main chains.

For L2 or sidechains, we have Loopring, Xdai, Matic, and just recently the Rollup family launched Arbitrum, with Optimism coming in a couple of months, and in a few months, we might also see ZkSync, Startware, Aztec…

It seems I haven't even counted some less popular main chains like Waves, Ada, the older generation of NEO, Quantum, IOST, etc…

If we believe that the endgame of blockchain, or at least in the next few years, will be such a multi-chain universe of competing factions.

What do you think is the likelihood of running all four major DeFi components: exchanges, borrowing, stablecoins, and synthetic assets on each chain? Just from the perspective of liquidation, it must satisfy:

1. There is a DEX on this chain with BTC, ETH, or mainstream collateral.

2. The LP pool of this DEX must be deep enough.

3. The LP of the DEX must also be sufficient. Indeed, many of these mainstream LPs on chains are provided by a few large holders; what if they withdraw their pools when liquidation occurs or before it happens?

By this standard, it is not an exaggeration to say that among the dozens of chains listed above, very few can actually meet the criteria.

If you believe that DeFi will be the future of blockchain (at least it certainly seems so at present), then based on the DeFi perspective, especially from the liquidation angle, the endgame of the multi-chain universe in a few years will likely be one of the following:

1. Coexistence of Multi-Chain Universes.

Theoretically, if the speed of entering and exiting L2 to L1 is fast enough, and there is a sufficiently complete cross-chain mechanism for the multi-chain universe, then when a black swan-level liquidation is triggered, the issue of insufficient depth in single-chain DEX can be solved by liquidation bots "arbitraging" from various other chains.

Because this, to some extent, shares the overall liquidity of the multi-chain universe, but it is only theoretical; with so many chains of different standards, achieving a quick and perfect connection is practically very difficult.

2. All Chains Converge to ETH.

One chain + one L2, ETH enters the L2 Rollup + 2.0 sharding era, where TPS and GAS are no longer issues; Rollup L2 completes the chicken dinner mission, leaving the strongest Rollup to KO all other Rollups, and depth is no longer an issue. From the perspective of liquidation, this should be the most "comfortable" model.

2.2 All Chains Converge to ETH.

One chain + multiple L2s, ETH enters the L2 Rollup + 2.0 sharding era, where TPS and GAS are no longer issues; in the Rollup L2 warring states era, fragmentation is severe, but methods to connect various Rollups are found, such as state channels or aggregators, and depth in this case is also acceptable. Liquidation will definitely be more complex than 2.1 but less than 1.

3. All Chains Converge to Polkadot.

From the perspective of liquidation, Polkadot has two major advantages: first, in terms of underlying design, ETH liquidation requires a TX to trigger, while Polkadot can achieve direct liquidation.

Second, Polkadot's standardized multi-chain with common interfaces is much more practical in terms of sharing depth across multi-chain interactions compared to the first type of heterogeneous cross-chain in the multi-chain universe, and it has more advantages compared to 2.2, with the biggest competitor being the 2.1 scenario.

Of course, this is just a speculation about the development of blockchain from the perspective of DeFi and liquidation, merely providing a thought process; please don't take it too seriously.

05 Conclusion

Regarding liquidation, let me mention one last thing: Uniswap V3 provides a capital efficiency that is N times higher than Uniswap V2 by offering granular control for LPs (liquidity providers), but if you think about it the other way around, in a large range, V3 theoretically also means N times lower depth than Uniswap V2.

When a black swan extreme event occurs and large-scale liquidation happens… it increasingly feels like Wall Street: the higher the efficiency, the bigger the bubble.