# Arbitrum moves 8,425 ETH of the Binance haul without the key

> Arbitrum's Security Council froze 8,425 ETH from the Binance theft and showed that it can move other people's funds without the private key. It already did so in April.
> By José Maldonado · Ethereum · 2026-11-03

On October 30, Arbitrum's [Security Council](https://tech.yahoo.com/general/articles/arbitrum-security-council-move-wallet-213347208.html) froze 8,425 ETH and [74.559.805](https://ethereum-magicians.org/t/how-arbitrums-freezes-send-transactions-from-an-address-without-its-key/30175) ARB tokens that were in an address which trackers attribute to the Binance exploit, valued at around $6.3 billion. What matters is not the figure, it is the method. The council moved funds from someone else's wallet without holding its private key, using the governance capability that the network reserves for emergencies. It is the second time it has done so this year.

The first was in April, when the same body immobilized 30,765 ETH linked to the [KelpDAO](https://ethereum-magicians.org/t/erc-8415-and-the-binance-heist-what-can-be-frozen-and-why-rwas-need-an-asset-state-layer/30170)exploit. That decision already sparked an intense debate in the community, and this second one confirms that it was not an exception. What began as an exceptional measure is becoming a recurring tool, and that changes the way the neutrality promises of layer 2s have to be read.

The technical detail that explains the case is simple and has profound consequences. A network built on [Ethereum](https://observatorioblockchain.com/en/ethereum/hegota-62-propuestas-ethereum-focil-frame-transactions/) can modify the state of its accounts through an update approved by its governance. If that update changes the balance of an address, the result is a transaction that appears to come from a wallet without its owner having signed anything, because the signature is no longer the only path.

## How is a wallet moved without its key?

On a classic blockchain, the only way to move funds is to prove that you control the private key associated with an address. That rule is the basis of ownership and also its limit, because a mistake or a theft becomes irreversible. Layer 2s inherit the model, but they add a governance layer that can alter the recorded state.

Arbitrum's Security Council exists to act quickly in the face of critical failures, with powers that allow it to execute updates without going through a full DAO vote. The trade-off is obvious. A body with the ability to change the state of the network can use it for good, to freeze what was stolen, or for anything else it considers urgent.

The transparency of the process is the only intermediate defense. Decisions are recorded, the community can discuss them and the council members are known. That does not make them unquestionable, but it allows the use of power to be audited and discussed in public, something that does not happen on a fully centralized network.

## The second time is not an anecdote

Repetition is what turns this case into a trend. With two interventions in a year, the precedent stops being a response to a single incident and becomes an available procedure. Any attacker who moves funds through Arbitrum now knows that there is a real chance their haul will be frozen.

For the victims of a theft, that is a relief. For anyone who uses the network, it raises a question about ownership. If a third party can modify the balance of an address, the claim that code is law only holds as long as governance does not decide otherwise, and that condition does not appear in the marketing brochures.

The defense argument is reasonable. Without that capability, the stolen money would leave the network in minutes and there would be no way to recover it. The intervention has a clear humanitarian goal and rests on an on-chain investigation that identifies the funds precisely. Few would dispute the end.

The problem lies in the next step. Once it is accepted that the end justifies the means, the limit is defined case by case and by whoever has the power to decide. Cryptocurrencies have spent fifteen years promising that this limit does not exist, and every intervention makes it harder to defend.

## Who decides what a stolen fund is?

Attributing a theft is an interpretive process, not a mathematical fact. A tracker identifies addresses, groups movements and decides what part of the balance comes from the crime. It is expert work that can be wrong, and in Arbitrum that work becomes the basis for an irreversible action.

In the Binance case, the chain of custody is relatively clear because there is an FBI investigation, tens of millions frozen by stablecoin issuers and a public attribution to a North Korean group. In other cases there will not be as much documentation, and the decision to freeze will depend on the credibility of a private report.

That is where the risk of capture appears. If the council freezes funds under pressure from an exchange, a government or a stablecoin issuer, the network loses the characteristic that makes it different from a bank. It can remain useful and fast, but it stops being neutral, and neutrality was the main argument.

## The precedent that layer 2s do not want to discuss

No layer 2 wants to put this topic at the center of its pitch, because the honest answer complicates the sale. They all have some emergency mechanism, with different names and similar powers. The difference lies in how many times it has been used and how much information is published afterwards.

The comparison with Ethereum is also inevitable. The main network reversed the state of the chain once, in 2016, and that split the community in two. Since then the norm is not to repeat it. Layer 2s have normalized in a year what the layer 1 did once in a decade, and they have done so without a visible fracture because the volume and the governance are different.

That asymmetry has a practical explanation. Layer 2s are companies or foundations with identifiable teams, and their users accept terms of service that they rarely read. The expectation of neutrality is lower because the design was never completely neutral.

The classic exit is to move to Ethereum's main layer, where an intervention of this kind would require a community fracture rather than a council vote. The price is known: higher fees, less speed and a less comfortable experience for the end user.

Another option is to accept the risk and diversify. Keeping funds on several networks reduces the impact of a freeze, just as diversifying an exchange reduces the damage of a bankruptcy. It does not eliminate the problem, and it does limit the maximum loss.

## What remains to be decided

The case leaves three open questions. The first is whether the council will publish a formal criterion for intervening, with thresholds and procedures, or whether it will continue deciding case by case. The second is what mechanism exists to reverse a mistaken freeze. The third is who is accountable if the frozen funds belonged to a third party in good faith.

As long as there are no answers, every successful intervention strengthens the capability and weakens the promise. Users who value censorship resistance have slower and more expensive alternatives. Those who value speed and convenience have already accepted, perhaps without knowing it, that someone can move their balance without their key.

The question anyone keeping funds on a layer 2 should ask is not whether the team is trustworthy. It is what would happen if tomorrow one of their addresses were flagged by a private report, and who they could call to discuss it.
