skip to content

Who Sandwiches the Sandwichers? ‘MEV Bros’ Acquitted on Fraud Charges

acquittal thumbnail 2400x1340

This article originally appeared in Galaxy Research's weekly newsletter. Subscribe to get timely insights delivered to your inbox every Friday morning.

WHAT HAPPENED

On Nov. 16, Judge Jessica Clarke acquitted Anton and James Peraire-Bueno of wire fraud and related charges, ending the first criminal case built around MEV, the profit available to whoever decides the order of transactions on a blockchain.

In practice, that “whoever” means validators and the builders to whom they outsource. A whole industry of bots exists to pay them for favorable placement. The brothers set a trap for some of those bots, ones whose own business is jumping in front of other people’s trades, which makes for an unusual set of victims. Using a flaw in Ethereum’s block-building software, the pair saw what those bots were about to do, something the system was designed to prevent, and turned it against them. The brothers walked away with about $25 million in roughly twelve seconds, at the cost of a protocol penalty of about 1 ETH (worth ~$3,000 at the time).

OUR TAKE

The judge did not say any of this was fine. She said wire fraud requires a lie, and the government never showed that the brothers told one. This relatively narrow ruling leaves a lot unresolved. Below, we walk through what the court found, then the questions it raises about manipulation law, sandwich trading, validator obligations, and whether code really is law.

What the Court Found

Some background first, because this is not a simple case to wrap your head around. Validators are the participants chosen to propose new blocks on Ethereum. Most of them hand off the work of assembling those blocks to specialized builders, through intermediaries called relays. A relay keeps a block’s contents hidden from the network until the validator signs off on it. The targets here were sandwich bots, which buy just ahead of someone else’s pending trade and sell right after it. The brothers baited those bots with transactions built to look like ideal sandwich targets. The brothers’ own validator then signed off on a block with deliberately bogus fields. The relay checked that the signature was valid, but not those fields, and revealed the block’s contents. The brothers rebuilt the block, kept the bots’ purchases, and swapped out the bait for transactions that sold the same tokens into the prices the bots had just pushed up, leaving the bots holding the bag. Their validator was slashed about 1 ETH for the misdeed of proposing two blocks for one slot.

The government outlined what it thought were two lies, and the court rejected both. The first alleged lie was the bait transaction. The court found no evidence that placing a transaction in the mempool, Ethereum’s public waiting room, conveys an agreement to be sandwiched. Nor does a transaction promise to be part of anyone’s bundle, or to land in any particular spot in a block, or any given block at all. Exploiting another market participant's expectations, even intentionally and profitably, doesn’t necessarily amount to a lie. The court drew that point from the Mango Markets case, United States v. Eisenberg. Avi Eisenberg pumped the price of Mango’s token and borrowed over $100 million against his inflated position. Mango had no terms of service, and the court found that Eisenberg’s borrowing made no promises, explicit or implicit.

The second alleged lie was the signature on the bogus block. The government argued that by signing a block header with fields they knew were invalid, the brothers made a false statement to the relay. The court rejected the theory on two fronts. First, a validator’s signature is a commitment to publish the block or accept slashing, not a promise of anything else. Ethereum, and trustless blockchains more broadly, are systems where everyone expects validators to act in “their own rational self-interest—whatever that is.” Registering a validator with a MEV-Boost relay doesn’t come with any terms of use. Second, the relay wasn’t checking those fields at all. It would have revealed the block’s contents no matter what the brothers put there. The same bug that made the exploit possible also made the “lie” immaterial.

Was Wire Fraud the Wrong Charge?

Wire fraud requires deception. The Commodity Exchange Act, by contrast, separately prohibits manipulative trading and deceptive devices, so an artificial-price scheme can violate it without anyone telling a lie. Eisenberg’s wire fraud conviction fell for lack of a false representation, but his commodities fraud and manipulation convictions were vacated only because prosecutors brought them in the wrong venue, and the judge found the evidence on those counts sufficient. Judge Clarke, of the U.S. Court of the Southern District of New York, was similarly careful. She framed the question as whether the evidence supported wire fraud, not whether the brothers’ conduct was right, moral, or a violation of “uncharged crimes.” Neither ruling is a safe harbor for DeFi manipulation.

Are Sandwich Attacks Legal?

If manipulating prices through transaction ordering can violate the Commodity Exchange Act, sandwich trading itself starts to look exposed. The court observed that “no trader ... would agree to be sandwiched,” but it never decided whether the bots’ own strategy was lawful. The CFTC retains antifraud and anti-manipulation authority over spot commodity markets, digital commodities included.

Slashing Is a Price, Not a Prohibition

The validator half of the case raises a design problem more than a legal one. By reading the brothers’ signature as a commitment to publish a block or be slashed, the court treated slashing as a price rather than a prohibition. Ethereum’s guarantees hold only as long as deviating from expected behavior costs more than it pays, and here a 1 ETH penalty stood against a $25 million payoff.

Code Is Law?

Sort of. The Eisenberg case and the MEV brothers ruling both refuse to turn a developer’s intentions into binding promises without evidence. Terms and conditions could become the standard for protocols looking for legal recourse in the future, but a frontend popup gating access may not bind someone interacting directly with smart contracts. Following the code doesn’t protect you from anti-manipulation statutes either. The court rejected the idea that “undermining the spirit of a system alone constitutes fraud.” What it does say is that, for wire fraud purposes, expectations alone don’t create a promise, and absent an actual agreement, the code is the closest thing to one.

Code isn’t law, but it is the closest thing to a contract you get when nobody signs anything else. The brothers read that contract more carefully than anyone on Ethereum, including the people who wrote the relay. If your protocol’s security depends on validators not doing the math on a 1 ETH penalty against a $25 million payoff, you don’t have a security model.

You are leaving Galaxy.com

You are leaving the Galaxy website and being directed to an external third-party website that we think might be of interest to you. Third-party websites are not under the control of Galaxy, and Galaxy is not responsible for the accuracy or completeness of the contents or the proper operation of any linked site. Please note the security and privacy policies on third-party websites differ from Galaxy policies, please read third-party privacy and security policies closely. If you do not wish to continue to the third-party site, click “Cancel”. The inclusion of any linked website does not imply Galaxy’s endorsement or adoption of the statements therein and is only provided for your convenience.