Mistakes in Sui concentrated liquidity after the Cetus exploit

The Cetus recovery and liquidity risks

Cetus volume returned to the top 10 DEXs shortly after its relaunch, but I suspect much of this activity involves users reclaiming funds following the one-month waiting period after the oracle manipulation exploit. That attack, which used fake tokens like $BULLA to distort price feeds, drained $260 million from the Sui ecosystem. While the Sui Foundation froze $162 million in stolen funds, attackers laundered the rest through various bridge pathways. One user reported receiving 86% of his SUI holdings from the vault, but another user noted a 27% gap in a supposedly unaffected liquidity pool. I find the community debate over Sui’s decentralization compelling because the network validators used their superior voting power to freeze the hacker’s wallet. The Cetus relaunch happened after the team allocated 15% of the total CETUS supply, including all unvested tokens, to compensate users who lost assets during the oracle manipulation attack that drained $260 million from the ecosystem. To support the protocol, the Sui Foundation provided a $30 million loan. The native CETUS token dropped 40% in the aftermath of the exploit, reflecting massive market anxiety. SUI price struggled to break past a resistance zone of $3.87 despite the rising TVL. Now, with the launch of native USDC on Sui, users can avoid the risks of bridging assets, but they still face the complexities of concentrated liquidity. You should check your pool balances daily if you use these narrow ranges.

Mysticeti speed and MEV extraction

Mysticeti reduces consensus latency by 80% to 390 milliseconds on the Sui testnet compared to the Narwhal-Bullshark algorithm used during the mainnet launch a year ago. This speed does not stop the extraction of value by block producers. Block producers still order transactions to capture arbitrage or sandwich user swaps. I would skip the hype around speed if it does not address the fact that users transacting on raw AMMs remain exposed to sandwich extraction. While the SUAVE project aims to restructure MEV at the architectural level, most users still face the reality of bots exploiting price impact. Mysticeti uses an optimized version of BFT consensus to process shared object transactions and minimize communication between validating nodes. This new protocol also reduces the CPU requirements for validator nodes. The engineering team expects to promote the implementation to the mainnet in a few months with the support of validating nodes. I ask how a faster consensus layer helps if the order flow remains easy to intercept.

Impermanent loss in narrow ranges

Providing liquidity in narrow ranges on Cetus amplifies the risk of impermanent loss for every participant. When prices deviate beyond your set range, your position becomes idle and earns zero fees. I consider the mismatch between fee income and price divergence the most dangerous part of providing liquidity for any active participant. If the price of an asset appreciates and moves outside your range, your position becomes 100% of the stablecoin or the depreciated token. If you expect a volatile asset to move 2x, you should budget at least 6% for impermanent loss before fees. If you provide 10 ETH and 10,000 USDC when ETH is 1,000 USDC, and ETH moves to 2,000 USDC, your value is 28,142.1 USDC. This results in an impermanent loss of 1,857.9 USDC when compared to holding the original assets. This occurs because LPs are effectively short gamma, meaning they lose money whenever the market moves in either direction. You must account for the fact that a 4x move in price causes a 20.0% loss and a 5x move causes a 25.46% loss.