Two brothers took about $25 million from Ethereum trading bots in roughly 12 seconds. On Monday, U.S. District Judge Jessica Clarke acquitted Anton and James Peraire-Bueno of wire fraud and related conspiracy charges over the April 2023 exploit. The brothers had replaced transactions that drew the bots into trades with transactions that sold the tokens instead. Clarke found that prosecutors had not proved a material misrepresentation, a false claim important to the alleged victims’ decisions. Her ruling limits the government’s attempt to prosecute this kind of manipulation as wire fraud without proving a lie.
How the brothers took $25 million from trading bots
The bots used sandwich trading. They placed trades around another user’s transaction to profit from its effect on a token’s price. According to Clarke’s ruling, the brothers targeted those bots through MEV-Boost, software that lets Ethereum validators use specialized builders to assemble blocks of transactions. A relay passes those blocks from builders to validators. The brothers used a flaw in that relay to inspect the bots’ trades before committing to a block.
- The Peraire-Bueno brothers: They placed bait transactions designed to attract particular sandwich bots. The bots then bundled trades around those transactions.
- The relay: Its flaw let the brothers see the bots’ bundled trades before the brothers committed to a block. By then, the bots had acted on the transactions the brothers had placed.
- $25 million: The brothers replaced their bait with transactions that sold the tokens. The bots lost about $25 million.
The exploit took place on April 3, 2023. Prosecutors charged the brothers in May 2024, describing their conduct in the indictment as a wire fraud and money laundering scheme. A jury could not reach a verdict at their November 2025 trial. Clarke’s acquittal ends the criminal case without a second trial.
Why Clarke rejected the fraud charges
Prosecutors argued that the brothers’ conduct falsely presented them as honest Ethereum validators. Clarke found that none of the government’s theories established a material misrepresentation. She wrote that the brothers’ use of MEV-Boost may have frustrated how its operators intended it to work. Prosecutors still had to prove a false representation that mattered to the alleged victims’ decisions, and Clarke found they had not done so.
The question before the Court is not whether the Peraire-Buenos’ conduct was right or moral, or even whether it violated uncharged crimes.
Jessica Clarke, U.S. district judge
The conspiracy charge tied to money laundering depended on proceeds from the alleged wire fraud, so it fell with the fraud counts. Clarke’s ruling leaves the bots’ roughly $25 million loss outside these criminal convictions. It also rejects this prosecution’s attempt to treat compliance with MEV-Boost’s intended operation as a duty enforceable under wire fraud law.
What Flashbots patched after the exploit
On April 21, a person claiming responsibility for the April 3 attack contacted researchers and Flashbots about a separate way to manipulate Ethereum blocks. Flashbots’ disclosure says the person proved control of an Ethereum address associated with the claim. The person then explained how a validator could request a block early, inspect its transactions and try to publish a competing block. That proposed strategy involved an early request for a block; the April 3 exploit used a relay flaw to inspect bundled bot trades.
We were able to confirm through the Flashbots relay logs that this strategy has not been used.
Flashbots disclosure
Flashbots tested a mitigation on April 23 and deployed it to its mainnet relay that day. Under the change, the relay waited until a validator’s assigned time slot before returning a block that the validator had requested early. Other relays deployed the patch over the following hours and days.














