Written by: Haotian
Many people are puzzled. After the Sui official stated that @CetusProtocol was hacked, the validator network coordinated to 'freeze' the hacker's address, recovering $160 million. How did they achieve this? Is decentralization really a 'lie'? Below, let's analyze this from a technical perspective:
Cross-chain bridge transfers: After the hacker successfully attacked, they immediately transferred some assets like USDC to Ethereum and other chains through a cross-chain bridge. This part of the funds can no longer be recovered because once they leave the Sui ecosystem, the validators are powerless.
Still on the Sui chain: A considerable amount of stolen funds remains in the Sui addresses controlled by the hacker. This part of the funds has become the target of 'freezing.'
According to the official announcement, 'a large number of validators identified the stolen fund addresses and are ignoring transactions from these addresses.'
—How exactly is this implemented?
1. Transaction filtering at the validator level—simply put, the validators collectively 'turn a blind eye':
Validators directly ignore transactions from the hacker's address at the transaction pool (mempool) stage;
These transactions are technically completely valid, but they just won’t be packaged onto the chain;
The hacker's funds are thus 'house arrested' in the address;
2. The key mechanism of the Move object model—the object model of the Move language makes this 'freezing' feasible:
Transfers must go on-chain: although the hacker controls a large amount of assets in the Sui address, to transfer these USDC, SUI, and other objects, a transaction must be initiated and packaged by the validators;
Validators hold the power of life and death: if validators refuse to package, the objects will never move;
As a result: the hacker nominally 'owns' these assets, but in reality, has no way to access them.
It's like having a bank card, but all ATMs refuse to serve you. The money is on the card, but you can't withdraw it. With ongoing monitoring and interference from SUI validator nodes (ATMs), the SUI and other tokens in the hacker's address will be unable to circulate, and these stolen funds are now effectively 'destroyed,' objectively playing a 'deflationary' role?
Of course, besides temporary coordination among validators, Sui may have preset a denial list function at the system level. If this is indeed the case, then the process might be: the relevant authority (such as the Sui Foundation or through governance) adds the hacker's address to the system deny_list, and validators execute according to this system rule, refusing to process transactions from blacklisted addresses.
Whether through temporary coordination or execution according to system rules, it is necessary for a majority of validators to act in unison. Clearly, the power distribution of Sui's validator network remains overly centralized, as a few nodes can control key decisions across the entire network.
However, the problem of validator centralization in Sui is not an isolated case for PoS chains—most PoS networks, from Ethereum to BSC, face similar validator centralization risks, but Sui has made the issue more apparent this time.
—How can a supposedly decentralized network have such strong centralized 'freezing' capabilities?
What's more troubling is that the Sui official stated they would return the frozen funds to the pool, but if the validators indeed 'refuse to package the transaction,' theoretically, these funds should never move. How did Sui manage to return them? This further challenges the decentralization characteristics of this chain!
Is it possible that, aside from a few centralized validators refusing transactions, the officials even have system-level super permissions to directly modify asset ownership? (Sui needs to provide further details on the 'freezing')
Before revealing the specific details, it is necessary to discuss the trade-offs regarding decentralization:
Is it necessarily bad to sacrifice a little decentralization for emergency response interference? If faced with a hacker attack, is it what users want for the entire chain to be completely powerless?
What I want to say is that while no one wants money to fall into the hands of hackers, this move raises greater market concerns: the freezing criteria are entirely 'subjective': what counts as 'stolen funds'? Who defines it? Where are the boundaries? Today it's freezing hackers, tomorrow who will it be? Once this precedent is set, the most core anti-censorship value of public chains will be utterly bankrupt, inevitably causing a loss of user trust.
Decentralization is not black and white; Sui has chosen a specific balance between user protection and decentralization. The key issue lies in the lack of a transparent governance mechanism and clear boundary standards.
Most blockchain projects are making this trade-off at this stage, but users have the right to know the truth, rather than being misled by the label of 'complete decentralization.'