Rethinking Decentralization through Sui

Key Takeaways

• Following the 2025 Cetus hack, the Sui network executed asset freezes and recoveries at the protocol level—an effective technical response that nonetheless sparked criticism and debate over core blockchain principles like censorship resistance and decentralization.
  

• The incident raised fundamental questions such as “Why is decentralization important?” and “Is absolute decentralization always good?”—and Sui demonstrated a pragmatic stance, prioritizing ecosystem stability over ideological purity.
  

• For such actions to carry legitimacy, they must be grounded in clearly defined, pre-established principles rather than ad hoc responses; moving forward, Sui must develop a more flexible and transparent governance framework to determine when and how the network should intervene.

1. Background - The Cetus Hack and the Response of Sui Validators

1.1 The Cetus Hack and Its Root Cause

On May 22, 2025, a major hacking incident occurred at Cetus, the largest decentralized exchange in the Sui ecosystem. The hack amounted to approximately $223 million, making it the largest DeFi-related exploit of 2025. The root cause was a vulnerability in the overflow check logic of an open-source library called inter_mate, used by Cetus’s CLMM smart contract. Specifically, the checked_shlw function mistakenly validated inputs based on a 256-bit limit instead of the intended 192-bit limit. The attacker exploited this flaw by inputting a small amount of tokens while tricking the system into believing that a much larger amount of liquidity had been provided, allowing repeated and fraudulent withdrawals.

This was not an issue stemming from Sui itself or from Move on Sui, but rather from a third-party library that Cetus had chosen to use. The truly contentious events began afterward.

1.2 Network and Validator Response, and the Question of Censorship Resistance

After the hack and the theft of over $200 million, the Sui validators swiftly took action to prevent the stolen funds from leaving the network. They froze the hacker’s wallet and the associated assets at the network level. This was not a newly introduced feature—it utilized Sui’s pre-existing transaction denial mechanism

(https://github.com/MystenLabs/sui/blob/9006185009afa2fe239fec771a7d067437814a42/crates/sui-transaction-checks/src/deny.rs#L170).

Here, the first controversy emerged—the question of censorship resistance.

Even though the funds in question belonged to a hacker, some have argued that the coordinated freezing of assets by the validators within such a short time violates one of blockchain’s core values: censorship resistance. The concern is that, if in an emergency, a coalition of validators and the foundation can unilaterally freeze assets or reject transactions, then can the network truly be considered censorship-resistant?

By definition, censorship resistance means that no one—regardless of power or position—should be able to interfere with the network. From this standpoint, Sui’s rapid intervention to freeze the hacker’s assets leaves it vulnerable to criticism regarding both censorship resistance and decentralization.

1.3 Ongoing Discussions on Chain Centralization and Asset Reversion

But the story doesn’t end there. If the stolen funds were frozen, the next logical step would be to return those funds to their rightful owners. To this end, Cetus proposed a governance resolution

(**https://sui.scan.space/vote).

The proposal suggests transferring the frozen assets from the hacker’s address to a multi-signature wallet controlled by Cetus, the Sui Foundation, and OtterSec.

However, the proposal itself sparked additional debate, as it arguably violates one of the fundamental principles of blockchain—immutability of ownership. The concern is that a small group (validators or the foundation) could effectively reassign asset ownership through collective agreement.

While the proposal doesn’t involve a direct transaction rollback, it does authorize a special transaction to recover the stolen assets from the hacker’s wallet. As such, many argue that this still constitutes a “soft rollback,” since it effectively alters the network’s state post-incident. The governance proposal is currently undergoing a vote within the community.

2. Takeaway - What Must Blockchain Protect, and What Can It Let Go Of?

2.1 Is Decentralization, god that failed?

Decentralization has become idolized in the blockchain industry. There is little concrete discussion or meaningful debate about why decentralization is necessary—decentralization itself is revered like a god. For some, the more decentralized a chain is, the more valuable it is perceived to be. But we must ask ourselves again: why did we support decentralization in the first place, and why did we feel the need to emphasize its importance? While everyone may have different views, I believe decentralization is merely a tool. What we ultimately aim to achieve through it is self-custody—the ability to fully own and control our own assets. This idea closely resembles the kind of society envisioned by classical liberals or libertarians. In those philosophies, freedom is not the ultimate goal, but rather a means to an end. What they ultimately seek is property rights (of course, interpretations differ depending on whether one is a utilitarian or deontological liberal; I align with the latter here).

This leads us to a fundamental question: What is the acceptable limit of freedom—or decentralization? Absolute freedom or decentralization can be interpreted as the absence of any external coercive force. But when freedom is taken to its extreme, it leads to anarchy; and when decentralization is taken to its extreme, a similar condition arises. In such a society, how can individuals protect their property from others? If my freedom infringes on someone else’s, is that still legitimate freedom? Likewise, if my decentralization infringes on someone else’s decentralization, can it still be considered legitimate? Just as absolute freedom paradoxically results in the erosion of others’ freedoms, absolute decentralization paradoxically leads us back to a need for centralization. This pattern has repeated itself throughout human history.

Even political philosophers who advocated radical freedom acknowledged that some freedoms are unacceptable. One such thinker is Murray Rothbard, a leading libertarian anarchist. In his book For a New Liberty, he identifies the Non-Aggression Axiom as the core of libertarian anarchism, clearly defining what constitutes an acceptable action and what does not:

If no man may aggress against another; if, in short, everyone has the absolute right to be “free” from aggression, then this at once implies that the libertarian stands foursquare for what are generally known as “civil liberties”: the freedom to speak, publish, assemble, and to engage in such “victimless crimes” as pornography, sexual deviation, and prostitution(which the libertarian does not regard as “crimes” at all, since since he defines a “crime” as violent invasion of someone else’s person or property).

Murray Rothbard <For a New Liberty>

Indeed, even within extreme freedom, there are principles and axioms. In any society where people live together—and especially in a community bound by economic interests—principles and rules are all the more necessary. That is what allows people to live alongside one another. Ultimately, any value only has meaning if it can be applied in practice. If the end destination of the freedom that libertarians so fervently advocate is a society akin to a jungle, then freedom can no longer be considered a praiseworthy value.

Let’s apply this to decentralization. People who support Ethereum often do so because it is, among all existing PoS-based blockchains, one of the most decentralized and censorship-resistant. (Of course, the DAO hard fork in 2016—which rolled back the chain’s history—was a clear exception and should be set aside from this discussion.) Over the past decade, Ethereum has experienced numerous major hacks—such as the Ronin Bridge and Poly Network exploits—but even then, Ethereum never froze assets or intervened at the chain level. In a way, this is because such restraint aligns with the core values Ethereum seeks to uphold. It is likely that Ethereum will continue to refrain from freezing assets in the future (though we can never say anything with absolute certainty—human behavior is unpredictable). For this reason, Ethereum could be described as an environment where decentralization is implemented in its most extreme form—a society where no scenario exists in which decentralization is considered unacceptable. While Ethereum and the EVM continue to improve in terms of security, hacks can happen at any time, and the only real way to protect users is through proactive auditing measures.

Now, imagine someone is pitching a blockchain platform to institutional clients. Would they prefer a chain that says:

“Our blockchain is entirely censorship-free, but that means if a hack occurs, recovery is impossible.”

or one that says:

“Our blockchain can, in rare cases, allow censorship through network-wide consensus—but only under clearly defined principles and criteria (such as violations of others’ property). In all other cases, we do not censor.”

Which of these two do you think institutions would prefer?If entities with business objectives are using a blockchain to develop services or implement specific features, they won’t necessarily prioritize decentralization for the sake of censorship resistance—like Ethereum does—as their highest value.

2.1.1 Idolization Undermines Diversity

This brings us to a fundamental question: Is it better to live in a society where decentralization is considered a supreme value but coercive force is allowed when someone infringes on another’s property? Or is it better to live in a society where decentralization is the only supreme value, and thus coercive force is never permitted under any circumstances? I would argue—with confidence—that most reasonable people would choose the former. Even political philosophers who advocate extreme libertarianism refrain from claiming that freedom alone is the ultimate moral good, precisely because such thinking becomes detached from reality and drifts into utopian fantasy.

Any extreme ideology eventually leads to fascism. The mindset of “my view is the absolute truth, and everything else is wrong” is precisely what gives rise to fascism. Ironically, the libertarians who preach radical freedom often resemble what one might call freedom-fascists—those who violently impose their idea of liberty. The same applies to decentralization. No one has ever reached a universal consensus on what degree of decentralization is best. Nor could such consensus exist, since people hold vastly different beliefs and values. Excessive idolization of decentralization silences those diverse perspectives and turns anyone who disagrees into a defector or a traitor. But if one truly supports decentralization, that must begin with respecting others—even if they support different levels of decentralization and live by their own principles.

The sanctification of decentralization may sound noble—but we also need to ask ourselves: Why have we come to revere decentralization so religiously in the first place? Too often, we become so immersed in the idea itself that we forget what originally led us to embrace it.

2.1.2 Break the Idol and Define Your Own Principles

If we agree that “absolute decentralization is not always right,” then the next step is to define our own boundaries for governance. That is: Under what circumstances can the chain intervene? And what principles and procedures (due process) must be followed? Just as even anarcho-libertarian societies rely on the Non-Aggression Axiom, blockchain networks too must have clearly defined principles. These should not simply be adopted from chains that treat decentralization as sacred. Rather, each chain should establish its governance boundaries and principles according to its purpose and identity—especially if we truly want blockchain technology to reach the broader public.

Take an example: If we agree on a principle that allows the use of coercion in cases where someone infringes on another’s property, then the next question must be: What is the threshold that justifies such intervention by the chain? For instance, should a network really freeze assets over a $20 token hack? In the case of Sui, validators and the foundation cooperated to freeze assets in this latest incident. But if no clear principles are established afterward, the network could be expected to intervene in every future hack. That is why it’s essential that the community, the foundation, and validators come together and define clear standards—so the network has a principled basis for when to intervene, or when to refrain, in future cases.

In this respect, one could argue that EOS was ahead of its time—by as much as seven years. It was the first (and perhaps last) chain to establish a constitution that reflected the principles and philosophy of its network from the very beginning.

2.2 What’s Disappointing About Sui’s Latest Proposal

Setting aside the philosophical debate around decentralization, I’d like to share a personal perspective on Sui’s recent governance proposal: in a word, it’s disappointing. Having followed and supported Sui for the past two years, I do recognize that this hack was truly a devastating incident. However, if the network and its ecosystem are to mature, the response must go beyond simply freezing the hacker’s assets and returning them. There must also be a clear and transparent identification of who holds responsibility for the incident—and appropriate accountability must follow.

If the hack was not caused by any vulnerability in the Sui network itself, nor in its programming language, but rather due to an external library used by Cetus, then why should the Sui network and its validators bear responsibility for those assets? This is a critical question. To draw a parallel: if the largest telecom company in a country gets hacked, should the state be the one compensating the victims? While there may not have been direct financial costs borne by the Sui network, it has nevertheless had to endure widespread criticism and reputational damage during the asset-freezing and recovery process.

While it’s important to have conversations around asset freezing and emergency responses, an even more fundamental issue needs to be addressed first: who is ultimately responsible for what happened, and how are they being held accountable? In this incident, the behavior of Sui validators and the Sui network resembled what one might call a paternalistic model of governance. If that’s the governance model Sui is aiming for, then so be it—but at least for me, it’s far from ideal.

What kind of governance does Sui actually want to embody? And how will it respond the next time an incident like this occurs? Who is the clear and rightful party responsible for this situation? And is that party truly fulfilling its responsibility?

2.3 Sui Needs a More Flexible Governance Framework

Source: Twitter

Another governance issue revealed by this incident is the loss of individual voting power within the Sui network. Sui uses a delegated proof-of-stake (dPoS) system, in which token holders delegate their tokens to validators, and validators exercise the associated voting rights on their behalf. This structure is very similar to today’s representative democracies. However, representative democracy emerged as a practical solution in the physical world because it was simply impossible for every citizen to participate in every legislative process or vote directly.

But with blockchain, this is no longer an issue—because the system isn’t bound by physical limitations. In fact, there are chains where not only validators but also individual stakers can participate directly in governance votes. Cosmos is a prime example. On Cosmos and Cosmos-based chains, users can delegate their voting rights to validators, but for specific proposals, they can override those delegations and vote independently. As a result, a validator’s decision does not automatically represent the vote of every delegator.

This makes Cosmos’s governance framework significantly more flexible. Sui, too, could take this opportunity to define the scope of its governance and adopt a system where individuals can directly vote on proposals within that defined scope. Doing so would allow Sui to evolve toward a more dynamic and inclusive governance model.

2.4 Blockchain as a Political Tool vs. Blockchain as a Platform

Source: Twitter

EigenLayer’s Kydo offered an interesting commentary on the recent asset-freezing incident in Sui. He noted that if someone felt uncomfortable with Sui’s decision to freeze the hacker’s funds, it likely means they view blockchain as a political tool (a tool of resistance). But if they were glad that the hack was prevented, it suggests they see blockchain as a technical tool.

Of course, the Sui network never explicitly stated at launch, “We are a platform, not a political instrument.” However, given that Sui does not pursue extreme decentralization, one could reasonably conclude that its goal is to build a platform that can be used by many, rather than serve as a tool for political resistance. From this perspective, Sui’s recent decision to freeze the hacked funds does not seem to contradict its identity.

That said, as I’ve repeatedly emphasized, blockchain is fundamentally a network of consensus—and any decision made on the network must be grounded in legitimacy. And legitimacy, in turn, must stem from clearly defined principles.

Through this incident, I hope Sui and its ecosystem will take the opportunity to establish clear, protocol-level principles—so that it may truly uphold rule by law, not rule by discretion.

3. Others’ Opinion

3.1 Jay from Four Pillars - “What Matters Most Is Establishing Procedural Legitimacy that Aligns with the System’s Vision.”

The fundamental reason any system exists is to fulfill its intended purpose. Yet no system in the real world can execute that purpose in a perfectly deterministic manner. This is especially true when human involvement is introduced, as errors in such contexts inevitably arise in stochastic ways.

From this perspective, blockchain can be seen as a novel approach aimed at complementing or redesigning existing systems. By creating a structure in which multiple participants act rationally under mutual oversight in pursuit of their own maximized utility, blockchain offers a “process-driven design methodology“ ****to help systems achieve their objectives more robustly.

However, we often witness a tendency to revere the values that blockchain-based approaches aim to uphold—such as censorship resistance or decentralization—simply because the approach itself is new. It's worth remembering that Bitcoin, the progenitor of modern blockchains, never even used the term “decentralization” in its whitepaper, nor did it attempt to rally support around a specific set of values. Its sole purpose was to establish a trustless, peer-to-peer system for value transfer.

Of course, different platforms may pursue different goals. IIn the case of the Sui smart contract platform, its core objective is to enable diverse assets to be reliably defined and interact seamlessly on-chain—unlocking use cases that were previously constrained in Web2 environments and expanding their potential horizontally. Therefore, it would be misguided to interpret Sui's chosen design approach through the lens of other blockchain paradigms or to apply their values wholesale. In this particular case, Sui prioritized the return of user assets and the stabilization of its ecosystem over strict adherence to censorship resistance—an arguably more aligned decision with its own long-term vision.

That said, I strongly agree with the author Steve's argument that ‘procedural principles’ reflecting a system's vision are essential for its sustainable operation. Such legitimacy provides a foundation for diverse stakeholders to understand and accept the system’s decisions, even when they are not in full agreement.

Whether it was Sui's decision not to censor the transaction in question, or its choice to assume responsibility for the actions of Cetus, both decisions could have been justifiable—if there had been procedural legitimacy behind them. As Sui seeks to build a more scalable ecosystem, strengthen its bond with an aligned community, and minimize operational friction, it will be imperative to establish mechanisms that secure legitimacy in Sui’s own manner, true to Sui’s unique identity—regardless of the decision at hand.

4. Resource

Related Articles, News, Tweets etc. :

• https://x.com/mattmaximo1/status/1927503611837210838

• https://x.com/0xkydo/status/1925603565260689609

• <For a New Liberty> - Murray N. Rothbard