LZ Research paper argues slashing may not protect proof-of-stake networks

Proof-of-stake blockchains lean on a simple threat: misbehave and you lose your money. A new paper from LZ Research argues that threat may arrive too late to matter.

The paper is titled “Too Late to Slash: Coordinating a Risk-Free Equivocation Attack.” Researchers Hao Chung and Chen-Da Liu-Zhang uploaded it to arXiv on September 24, 2026, under identifier 2609.30509. Their core claim is that rational validators can coordinate an attack in a way that sidesteps the penalty slashing is supposed to impose.

What the paper actually claims

In a proof-of-stake system, validators lock up tokens as collateral for the right to help confirm transactions. Slashing is the punishment mechanism: if a validator is caught cheating, the protocol automatically destroys some or all of its staked funds.

The specific form of cheating the paper examines is equivocation — when a validator signs two conflicting blocks, effectively vouching for two incompatible versions of history at once.

Chung and Liu-Zhang introduce a coordination protocol for validators. According to the paper, this protocol makes it possible to solicit equivocation from validators without those validators risking losses if the attack fails to reach a critical threshold. If not enough validators sign on, nobody gets slashed. If enough do, the attack proceeds.

The game theory behind it

The authors frame their result in game-theoretic terms. They state that their proposed coordination strategy results in an ex post Nash equilibrium — a situation where the strategy stays optimal even after participants learn how things turned out.

The paper also argues the approach can produce positive gains, expressed as ε > 0, from an attack. The authors say this holds regardless of how small those gains are relative to the amounts staked. Conventional thinking holds that an attack only makes sense if the payoff exceeds the value of the stake at risk. The paper suggests even a minimal successful attack can come out ahead, because the risk side of the equation has been engineered away.

What it does and does not cover

The paper is theoretical. It does not directly analyze any specific proof-of-stake protocol or token.

Translating a game-theoretic result into a real-world attack also involves practical hurdles the model may abstract away. Validators would need to find each other, trust the coordination process, and act in concert.

Why slashing has been treated as essential

Slashing exists because proof-of-stake gave up the physical cost that secures proof-of-work systems. Miners burn electricity and hardware to produce blocks, which makes attacks inherently expensive. Validators, by contrast, mostly just sign messages. Signing a second, conflicting message costs almost nothing computationally. A 2023 report by a16z crypto emphasized the role of slashing as a mechanism to elevate the costs associated with corruption and collusion among validators, aiming to safeguard network integrity against both internal and external threats.

When researchers argue that the penalty can be dodged through coordination, they are poking at a foundational piece of that model. The title, “Too Late to Slash,” makes the point directly: by the time the protocol can punish anyone, the coordinated attack has either succeeded or quietly dissolved without consequence.

What this means for proof-of-stake networks

For investors and traders, the most immediate effect may be a reassessment of how much security slashing actually buys. The findings could also feed into how protocol designers think about validator incentives, as a deterrent that only activates after an attack is visible may need backup from mechanisms that make coordination itself harder or riskier to organize.

For stakers themselves, the practical takeaway is narrower. Slashing remains a real penalty for individual mistakes and uncoordinated misbehavior. The open question raised by Chung and Liu-Zhang is whether it holds up against validators who plan together and only commit once they know the numbers work in their favor.