$50 million instantly shrinks to $36,000 — a transaction harvested by MEV

Written by: Tanay Ved

Translation: Shanyouba, Jinse Finance

Key Points

A $50 million USDT exchange transaction initiated by a user through the Aave interface was routed through multiple DeFi platforms and completed within a single block, ultimately resulting in only about $36,000 worth of AAVE tokens received.

The transaction was routed to the SushiSwap WETH/AAVE trading pool, which had only about $75,000 in liquidity. This caused extreme price impact and allowed nearby MEV bots to earn millions of dollars in profit.

This event reveals how large orders, on-chain liquidity shortages, and routing mechanism failures can lead to poor transaction execution. It also highlights the urgent need to establish better risk protection mechanisms and design smarter transaction routing systems in permissionless blockchain markets.

A Nearly $50 Million Loss on a Swap Transaction

One of the core features of blockchain and decentralized finance (DeFi) is its openness and permissionless nature. While these traits bring new efficiencies and participation methods, they can also lead to protocol exploits, oracle pricing errors, or operational mistakes.

On March 12, a user executed a swap worth $50.4 million, losing 99.9% of the initial value. The user exchanged $50.4 million USDT on the Ethereum mainnet Aave interface and only received about $36,000 worth of AAVE tokens. In permissionless blockchain systems, large transactions encountering liquidity shortages exemplify the on-chain risks of transaction execution.

This article will analyze the mechanics of this swap, the flow of funds, and the lessons for executing large trades with limited on-chain liquidity.

Fund Flows and Operational Mechanics Behind the Transaction

The user’s initial goal was simply to swap $50 million USDT for AAVE tokens, but the process was complex. It involved multiple protocols and asset conversions, from the lending protocol interface to trade aggregators, decentralized exchanges, and entities involved in maximum extractable value (MEV) arbitrage.

The transaction was first initiated as a collateral swap on the Aave interface, with the user holding an aEthUSDT position. aTokens are interest-bearing deposit receipts on Aave, and aEthUSDT represents USDT deposited on Aave V3. To complete the collateral conversion, the Aave interface routed the order to CoW Protocol—a third-party decentralized trading aggregator that finds optimal trade paths across liquidity platforms using various market makers and algorithms.

Within the same block, the user’s aUSDT position funds flowed through Aave, CoW, Uniswap, SushiSwap, and ultimately to MEV-related arbitrage entities: Coin Metrics ATLAS platform and Talos Research.

In this transaction, the winning solver first exchanged the redeemed USDT for WETH on Uniswap V3, then transferred the WETH into the low-liquidity SushiSwap AAVE/WETH pool. This was the core step causing significant value loss. Using Coin Metrics’ ATLAS platform, we can trace the fund flow across block #24643151, clearly identifying the points of value degradation.

Decentralized Exchange Routing

aUSDT Burn & USDT Redemption: The user’s 50,43 million aEthUSDT was transferred to CoW Protocol’s GPV2 settlement contract, which redeemed the tokens via Aave V3, burning the aTokens and releasing $50.43 million USDT from the reserve pool.

USDT to WETH on Uniswap V3: The CoW settlement contract transferred the $50.43 million USDT into the USDT/WETH pool on Uniswap V3, ultimately receiving 17,957.81 WETH, worth about $38.2 million.

WETH to AAVE on SushiSwap: The 17,957.81 WETH (roughly $38.2 million) was transferred into the SushiSwap WETH/AAVE pool, which at that time had only about $73,000 in liquidity. Facing such a large swap, the pool could only provide about 331 AAVE tokens, worth roughly $36,000. Most of the value was lost at this stage.

This block also triggered numerous MEV arbitrage actions. MEV bots monitor pending transactions to exploit large or predictable orders for profit. In this case, bots performed a “sandwich attack,” executing reverse trades before and after the user’s transaction to profit from price deviations in the SushiSwap pool.

Related MEV Arbitrage Operations

Front-running: An MEV arbitrage bot (wallet address: 0x06cf…5ef) borrowed 14,175 WETH via Morpho’s flash loan, then swapped on Bancor to get 128.57 AAVE tokens. This pushed up AAVE’s price before the user’s transaction entered the SushiSwap pool.

Back-running: After the user’s order was executed at the inflated price, the bot immediately sold the 128.57 AAVE tokens for 17,912 WETH, worth about $4.09 million. After repaying the Morpho flash loan (14,175 WETH), remaining profits were split: approximately 13,087 WETH (around $2.99 million) paid as block inclusion fee to Titan Builder, and about 4,824 WETH ($1 million) kept by the arbitrage bot.

All these steps—from flash loans, front-running, back-running, to loan repayment—were completed atomically within the same block in a single transaction.

Where Did the Funds Go?

The transaction’s value was redistributed among other participants in the block. As shown above, about $27 million in block rewards (gas fees) went to Titan Builder, roughly $10 million was earned by the MEV arbitrage bot performing the sandwich attack, and the user’s final AAVE position was worth only about $36,000.

Large Transactions and Liquidity Shortages

The approximately $39 million value loss can be traced back to the SushiSwap WETH/AAVE pool, an automated market maker (AMM) liquidity pool under SushiSwap. At the time of the trade, the pool’s total liquidity was only about $75,000, yet it was asked to process a nearly $37 million trade—almost 500 times larger. Such a scale mismatch caused severe price impact. As the pool’s asset reserves fluctuated sharply, candlestick data showed that the AAVE token’s price surged from about $118 (0.054 WETH per AAVE) to roughly $306,000 (139.9 WETH per AAVE).

This incident also raised questions: Why was such a large trade executed as a single transaction routed to a liquidity pool with such limited liquidity? Such trades should have been intercepted from the start. Aave’s post-incident report cited market liquidity shortages and the user’s explicit acceptance of a 99.9% price impact warning. CoW Protocol’s analysis indicated that the event resulted from multiple routing failures: its price verification system failed due to “GAS limit issues,” rejecting better-priced routes, and ultimately routing the large trade to a pool with only $75,000 in liquidity.

Lessons from the Event

In summary, the core issue was not protocol vulnerabilities or malicious exploits, but the combination of large trades and on-chain liquidity shortages causing extreme price impacts. This event offers several important lessons for the structure of on-chain markets and the broader crypto industry:

Scalable Risk Protection Mechanisms

For large trades, on-chain risk warnings are insufficient. The permissionless nature of DeFi means protocols typically do not prevent economically harmful trades. However, this incident shows the necessity of implementing filters for trade execution, tiered validation of trading pools, or trade size limits. Finding a balance between maintaining open market access and protecting user assets remains a key challenge.

Execution Strategies for Large Trades

Large swaps like this $50 million transaction should be split over time using methods like Time-Weighted Average Price (TWAP) or algorithmic trading across different platforms. This approach reduces market impact and minimizes MEV arbitrage risks. Such strategies are well-developed in traditional finance and are increasingly important on-chain.

The Role of Data

Real-time data on market depth, cross-platform liquidity, and slippage is crucial for pre-trade validation. Embedding these data points into routing and risk management systems can help prevent similar incidents, especially as on-chain markets grow and fragment.

Trade-offs of MEV and Permissionless Trading

As demonstrated in this block, open mempools are vulnerable to sandwich attacks and other MEV arbitrages. The industry needs to optimize design by offering private or protected transaction channels, MEV-aware routing systems, and giving users options to enable or disable risk protections—balancing the benefits of permissionless liquidity with the need to reduce trading risks.