MEV Brothers Pocket $25M in 12 Seconds, Walk Free: Judge Says They Didn't Lie—They Just Read the Code More Carefully Than Anyone

ETH
Wire Fraud AcquittalEthereum ValidatorsDeFi Manipulationsandwich attacksCommodity Exchange ActCode Is LawMEV
1 hour agoSource: blockweeks.com
MEV Brothers Pocket $25M in 12 Seconds, Walk Free: Judge Says They Didn't Lie—They Just Read the Code More Carefully Than Anyone

This article was compiled and organized by BlockWeeks

On November 16, Judge Jessica Clarke of the U.S. District Court for the Southern District of New York ruled that Anton Peraire-Bueno and James Peraire-Bueno were not guilty of wire fraud and related charges, bringing to a close the first criminal case built around MEV. MEV, or "maximal extractable value," refers to the profit that whoever has the right to decide the order of transactions on a blockchain can obtain.

In practice, this "who" refers to validators, as well as the block builders to whom they outsource block production. Around this chain, an entire bot industry has grown up: they pay validators and builders in exchange for transaction ordering favorable to themselves. And the trap set by the brothers targeted precisely one special class of these bots—sandwich bots, which make money by buying before others complete their trades and selling immediately after. In other words, the "victims" in this case were themselves engaged in the front-running business.

The judge did not say that all of this was fine. What she said was: wire fraud requires a "lie," and the government never proved that the brothers had lied. This relatively narrow ruling leaves a host of unresolved questions—about the legal boundaries of market manipulation, the legality of sandwich trading itself, the obligations of validators, and whether "code is law" actually holds.

First, let us explain the modus operandi clearly: how they set a trap for the "front-runners"

This is not a case that is easy to understand at a glance; some background is needed first. Validators are participants selected to propose new blocks on Ethereum. Most validators hand off the work of assembling blocks to specialized builders, with intermediaries known as "relays" brokering in between. Relays keep block contents hidden from the entire network until the validator signs off, after which they are made public.

The target in this case was sandwich bots: they buy ahead of other people's pending transactions and sell after the other party has pushed the price up. The brothers used a batch of carefully constructed transactions that looked like "perfect sandwich prey" to fish out these bots. Subsequently, a validator controlled by the brothers themselves signed a block in which a field had been artificially forged. The relay only verified whether the signature was valid, but did not verify those fields, and thus leaked the block contents.

The brothers then rebuilt the block, kept the bots' buy orders, replaced the bait transactions with transactions that "sold the same batch of tokens at the price the bots had just pushed up," and finally left the bots holding the bag. The whole process took about 12 seconds, and they took about $25 million; the cost was that their validator was slashed about 1 ETH for "proposing two blocks in the same slot"—about $3,000 at the price at the time.

The two "lies" alleged by the government: why both were rejected

The prosecution put forward two "lies," and the court accepted neither.

The first was the bait transactions. The court held that there was no evidence that "putting a transaction into the mempool (Ethereum's public waiting room for transactions)" amounted to a promise that one was willing to be sandwiched. Nor does a transaction promise that it will necessarily enter anyone's bundle, that it will necessarily land in a certain position in a block, or even that it will enter a particular block. Exploiting the expectations of other market participants—even intentionally, and even at a profit—does not necessarily constitute lying.

This view comes from the Mango Markets case (United States v. Eisenberg). Avi Eisenberg had pushed up the price of the Mango token and borrowed more than $100 million against an inflated position. Mango had no terms of service at the time, and the court accordingly found that Eisenberg's borrowing behavior had not made any express or implied promise.

The second was the signature on the forged block. The prosecution argued that the brothers knew the block header fields were invalid yet still signed, which amounted to making a false statement to the relay. The court rejected this on two levels:

First, a validator's signature is merely a commitment—either to publish this block or to accept slashing—and does not promise anything else. Ethereum, and more broadly trustless blockchains, are by nature systems in which everyone expects validators to act according to their own rational self-interest—whatever that may be. Registering a validator with an MEV-Boost relay also carries no terms of use.

Second, the relay did not verify those fields at all. No matter what the brothers put in those fields, the relay would still make the block contents public. In other words, the very vulnerability that made this attack possible also made this "lie" irrelevant.

The key distinction: wire fraud requires "deception"; market manipulation does not

The element required for wire fraud is deception. The Commodity Exchange Act (CEA) follows a different logic: it separately prohibits manipulative trading and deceptive devices, so a scheme to artificially create prices may violate the CEA even if no one lied.

Eisenberg's wire fraud conviction failed for lack of a false statement, but his commodity fraud and market manipulation convictions were vacated only because the prosecution had gotten the venue wrong; the judge found the evidence on those two counts sufficient.

Judge Clarke in this case was similarly cautious. She framed the question as "whether the evidence supports wire fraud," not "whether the brothers' conduct was correct, whether it was moral, or whether it constituted uncharged offenses." Neither ruling is a safe harbor for DeFi manipulation.

Sandwich trading itself is beginning to look dangerous

If manipulating prices through transaction ordering may violate the Commodity Exchange Act, then sandwich trading itself is in the crosshairs. The court wrote in its ruling: "No trader would agree to be sandwiched." But it never ruled on whether the bots' own strategy was legal. The CFTC retains anti-fraud and anti-manipulation enforcement authority over spot commodity markets (including digital commodities).

The "validator" half of the case is less a legal question than a design question. The court understood the brothers' signature as a commitment to "either publish the block or be slashed," effectively treating slashing as a "price" rather than a "prohibition." Ethereum's security guarantee holds only when the cost of deviating from expected behavior exceeds the benefit—and here, a 1 ETH slashing stood against a $25 million gain.

Code is not law, but it is the closest thing to a contract

The rulings in the Eisenberg case and the "MEV brothers" case both refuse to turn developers' intentions into binding promises without evidence. For protocols that wish to preserve a path to legal recourse, terms of service may become standard—but a front-end pop-up "agree" may not bind those who interact directly with smart contracts.

Acting according to the code will not protect you either; anti-manipulation statutes still apply. The court explicitly rejected the notion that "merely violating the spirit of a system constitutes fraud." What it actually said was: for purposes of wire fraud, expectations themselves do not constitute promises; and in the absence of any actual agreement, code is the closest thing to a promise.

Code is not law, but when no one has signed any other document, code is the closest thing to a contract you can get. The brothers read this "contract" more carefully than anyone else on Ethereum—including the people who wrote the relay. If your protocol's security rests on "validators won't do the math between a 1 ETH slashing and a $25 million gain," then you have no security model at all.