Web3 & DeFi

What Is MEV in Crypto? Lessons From the $25M Ethereum Case

MEV is the profit bots make by reordering transactions, and for DEX traders it mostly shows up as sandwich attacks. Here is how it works, what the $25M Ethereum MEV acquittal decided, and how to protect your swaps.

Navixa12 min read

Illustration: two small robot trading bots holding the slices of a sandwich with a single coin squeezed in the middle, standing on a chain of blocks next to a judge's gavel

MEV, short for maximal extractable value, is the profit someone can make by choosing which transactions go into a block and in what order. On a decentralized exchange, it mostly reaches ordinary traders as the sandwich attack: a bot sees your pending swap, buys just before you, and sells just after you, so you pay a worse price and the bot keeps the difference.

MEV was in the news this week for a different reason. On November 16, 2026, a federal judge in Manhattan acquitted two brothers accused of taking about $25 million from sandwich bots in roughly 12 seconds. The ruling answers a narrow legal question, and it is easy to read too much into it. This guide explains what MEV is, what the case did and didn't decide, and what actually protects your trades.

📌 Key Takeaways
  • MEV is profit from choosing the order of transactions in a block. For DEX traders it mostly shows up as sandwich attacks.
  • On Nov 16, 2026, Judge Jessica Clarke acquitted Anton and James Peraire-Bueno in the $25M Ethereum MEV case: prosecutors did not prove a material lie, which wire fraud requires.
  • The ruling does not make MEV legal in general. It says this conduct was not wire fraud. EU regulators treat MEV as possible market abuse.
  • Sandwich bots ran 60,000 to 90,000 attacks a month on Ethereum between Nov 2024 and Oct 2025, according to EigenPhi data.
  • Your protection is technical, not legal: a tight slippage tolerance, protected routing, limit orders and deep pools.

What is MEV?

Ethereum's developer documentation defines MEV as “the maximum value that can be extracted from block production in excess of the standard block reward and gas fees”, by including, excluding or reordering transactions in a block. The term was coined in a 2020 research paper as "miner extractable value". It later became "maximal" extractable value, because validators and other parties, not only miners, can decide the order.

MEV exists because a blockchain has no single queue. Pending transactions wait in a public "mempool" where anyone can read them before they are confirmed. The European Securities and Markets Authority (ESMA) puts the contrast with traditional markets plainly in its July 2025 report on MEV: stock exchanges enforce a fixed order, usually first come, first served, while blockchains leave the ordering to the parties that build blocks.

On Ethereum today, three groups are involved:

  • Searchers run bots that scan the mempool for profitable opportunities and bid to have their transactions placed in a specific spot.
  • Builders assemble those bids and other transactions into full blocks.
  • Validators propose the block. Many run MEV-Boost software, which collects blocks from builders through middlemen called relays and picks the one that pays the most.

That relay layer matters for the court case below.

Not all MEV hurts you

MEV is not one thing. Some forms keep DeFi working; others take value straight from users.

TypeWhat the bot doesEffect on regular traders
ArbitrageBuys a token where it is cheap and sells it where it is dear, often in one transactionMostly helpful: keeps prices in line across exchanges
LiquidationsRaces to close under-collateralized loans on lending protocols for a feeMostly helpful: keeps lenders solvent, painful for the borrower
Back-runningTrades right after a large swap to pull a pool's price back in lineMostly neutral: keeps pool prices current
Front-runningCopies or jumps ahead of a profitable pending tradeHarmful: you get a worse price or lose the opportunity
Sandwich attackBuys right before your swap and sells right after itHarmful: you pay more, the bot keeps the gap

Ethereum's documentation makes the same split: rational searchers fix price gaps that DeFi protocols depend on, while sandwiched users get worse execution.

How a sandwich attack works

Diagram of a sandwich attack: in one block, the bot buys first, your swap fills at a worse price, and the bot sells right after, with a sketch of the token price stepping up twice and then down
The order of transactions in a sandwich attack. A diagram, not real data.

When you swap on a DEX, you set a slippage tolerance: the most the price may move against you before the trade is cancelled. A sandwich bot uses exactly that room. In Uniswap's words, “sandwich bots prey on wide slippage tolerances by pushing the price just enough to trigger your swap”. It buys first, your swap fills near your worst acceptable price, and the bot sells into the higher price your own trade created.

How common is it? A Cointelegraph Research study built on EigenPhi data, published December 4, 2025 and covering November 2024 to October 2025, found:

  • 60,000 to 90,000 sandwich attacks a month on Ethereum, a steady pace through the whole period.
  • The value extracted fell from nearly $10 million a month in late 2024 to about $2.5 million in October 2025, even as DEX volume grew.
  • Trader losses of roughly $60 million a year, with block builders collecting most of the bots' take as fees.
  • About 70% of all attacks came from a single bot operator known as "jared".
  • About 38% targeted stablecoin, wrapped-token and liquid-staking pools, so even "boring" swaps were hit.

The study suggests the decline may partly reflect more traders using MEV protection, though it does not prove that. Risk also differs by token: a study by Uniswap Labs researchers found the chance of adversarial slippage was about 80% higher when trading the memecoin PEPE than when trading USDC.

The $25M case: what happened

Timeline of the Peraire-Bueno case: the exploit in April 2023, arrest and charges in May 2024, the separate Mango Markets ruling in May 2025, a four-week trial ending in a mistrial in October to November 2025, and the judgment of acquittal on November 16, 2026
Key dates in United States v. Peraire-Bueno.

According to prosecutors, in April 2023 Anton and James Peraire-Bueno, two MIT-educated brothers, set a trap for sandwich bots. As Decrypt summarized the case, they placed bait transactions designed to lure specific bots, used a bug in a MEV-Boost relay to see the bots' bundled trades before the block was final, and swapped their bait for trades that sold the tokens instead. The bots lost about $25 million in roughly 12 seconds.

ESMA's report described the alleged scheme as "a sandwich attack on a sandwich attack", and pointed out the irony: the case targeted the people who tricked the MEV bots, not the bots that extract value from ordinary users every day.

The brothers were arrested in May 2024 and charged with wire fraud, conspiracy to commit wire fraud and conspiracy to commit money laundering, charges carrying up to 20 years in prison each. Their four-week trial ended in a mistrial in November 2025 after the jury could not agree. On November 16, 2026, Judge Jessica Clarke granted their motion for a judgment of acquittal on all three counts, which ended the case without a second trial.

What the judge decided, and what she didn't

Wire fraud requires a material misrepresentation: a lie that matters. The judge found the government had not proven one. Per the court coverage:

  • She called Ethereum a "trustless, highly adversarial environment with no terms of use."
  • She wrote that "the question before the Court is not whether the Peraire-Buenos' conduct was right or moral", but whether the evidence could support wire fraud.
  • The scheme "may have frustrated the spirit" of MEV-Boost, she found, but that was not enough to make it fraud.
  • In a footnote, she said the prosecution "raises serious due process concerns", without ruling on that point.

She also leaned on a 2025 decision from the same courthouse. In May 2025, Judge Arun Subramanian overturned the convictions of Avraham Eisenberg, who had drained more than $100 million from the Mango Markets protocol, finding among other things no proof of a false statement on a platform with no rules against what he did. Prosecutors appealed that ruling in January 2026, arguing it would unsettle traditional understandings of fraud.

Reactions split along familiar lines. Coin Center's Peter Van Valkenburgh called the acquittal "excellent news", warning that the prosecution's "honest validator" theory could have let prosecutors police how validators behave on an open network.

So is MEV legal?

There is no clean answer, because no law defines MEV. ESMA's report says it directly: "MEV is not legally defined and there have been very few enforcement actions or litigations specifically targeted at MEV activities."

In the US, the Peraire-Bueno ruling settles one point: on a permissionless network with no rules, exploiting other bots' trades was not shown to be wire fraud, because no one was lied to. It does not say the conduct was legal under every other law, and the judge contrasted Ethereum with regulated exchanges, whose written rules can forbid deceptive orders. The Mango appeal may also shape how far this reasoning reaches.

In the EU, the direction is different. ESMA's 2024 consultation on market-abuse rules under MiCA named MEV as an example that "may suggest the existence of market abuse", and its 2025 report says regulated firms may need to detect MEV and report it as suspicious activity. The same report adds that MEV "may go against the principles of fairness and integrity" that underpin orderly markets.

⚠️Warning

The ruling protects no one's swap. If a bot sandwiches your trade, a confirmed transaction on a public blockchain can't be reversed, and the bot is usually an anonymous address. Your protection against MEV is how you trade, not a court.

How to protect your swaps from MEV

None of these steps removes MEV completely; Uniswap itself says "no single tactic can get rid of malicious MEV forever." Combined, they shrink the room bots have to work with.

  1. Keep slippage tolerance tight. The lower it is, the less a bot can push your price before your trade cancels. Uniswap's web app suggests a value that is usually between 0.1% and 5%, based on gas and trade size. Too tight, and swaps fail, and you still pay gas on a failed swap.
  2. Use protected routing. Private transaction services keep your swap out of the public mempool. Flashbots Protect, for example, sends transactions to a private mempool where they are "hidden from frontrunning and sandwich bots", and doesn't include transactions that would fail. Some wallets have this built in; Uniswap says its wallet protects all Ethereum mainnet swaps this way. ESMA notes private routes come with their own trade-offs, such as more reliance on a few operators.
  3. Prefer limit orders or intent-based swaps. With a limit order, the price is fixed in advance, so there is little room to squeeze. Intent systems such as UniswapX send your order to competing fillers instead of the mempool. If you're new to order types, our guide to market, limit and stop orders explains the trade-offs.
  4. Trade in deep pools. Uniswap warns that "low-liquidity pools are prime targets for MEV bots." A large swap in a thin pool moves the price more, which is exactly what a sandwich needs.
  5. Be extra careful with memecoins. Thin, volatile tokens carry more adversarial slippage, as the PEPE study above showed.

For a step-by-step walkthrough of swapping itself, see how to trade on a DEX, and our MetaMask wallet guide covers the wallet side.

What about centralized exchanges?

On a centralized exchange, your order goes to the exchange's own matching engine, not to a public mempool, so this kind of block-ordering MEV doesn't apply. You still pay costs: trading fees, the spread, and slippage on market orders. That is why Navixa's backtests include them: the Price Action strategy, for example, is tested on Binance with 0.05% commission and 5 ticks of slippage on every trade. Our guide to crypto trading fees breaks those costs down.

Quick Quiz
You swap on a DEX with a 5% slippage tolerance. What does a sandwich bot do with that room?

FAQ

Is MEV the same as front-running?

No. Front-running is one kind of MEV. MEV also covers arbitrage and liquidations, which help keep DeFi prices and lending protocols in order. A sandwich attack combines a front-run and a back-run around one victim's trade.

No. It found that prosecutors had not proven wire fraud, because they had not shown a material lie. The judge said she was not deciding whether the conduct was right or moral, and she distinguished regulated markets, where written rules can ban deceptive orders.

Can I get my money back after a sandwich attack?

Usually not. A confirmed swap on a public blockchain can't be reversed, and the bot is typically an anonymous address. Prevention, through slippage settings and protected routing, is the only practical defence.

Does MEV happen on other blockchains?

Yes. ESMA's 2025 report found MEV widespread on Ethereum and growing on some other chains. Any network where pending transactions can be seen and reordered before they are confirmed can have it.

The bottom line

MEV is built into how blockchains order transactions, and for DEX traders its most costly form is the sandwich attack. The $25M Ethereum case ended with a narrow finding: tricking MEV bots on a network with no rules was not proven to be wire fraud. It did not make MEV legal everywhere, and it changed nothing about the bots that sit between you and a fair fill. Keep your slippage tight, route swaps privately, and use limit orders where you can. If you trade on centralized exchanges, the costs to watch are different, and our guide to crypto trading fees covers them.

Was this article helpful?

Ready when you are

Get your seat on TradingView

Pick a plan and your invite-only strategy is activated within 24 hours.

  • Full reports published
  • Every coin we tested, losers included
  • Cancel anytime