How Much MEV Is There on Ethereum? What the Data Actually Shows

By Merunas Grincalaitis, founder of MEV DAO and a developer who builds MEV, sniping and trading bots on Ethereum. Last updated: Oct 30, 2026.

TL;DR: Nobody has one clean number for MEV on Ethereum, but the published estimates are in the same range: roughly $1 billion or more since the Merge in September 2022. Flashbots measured about 526,207 ETH (around $1.1 billion) of value paid to validators between September 2022 and June 2024, and Sorella Labs put total MEV revenue since September 2022 at $1.26 billion including CEX-DEX arbitrage, according to a 2025 ESMA report. Most of that value doesn't stay with the bots. It flows to block builders and validators.

"How much is MEV?" sounds like a question with a simple answer. It isn't, and the reason is useful to understand if you run bots, trade on DEXes or are thinking about paying someone to build an MEV strategy. Different trackers measure different things, cover different strategies and price ETH at different moments. Below I go through the numbers I trust, where they come from, and what they actually say about the business.

What does "how much MEV" actually measure?

There are at least three different numbers people call "MEV", and they can differ by a lot. Before comparing any figures, check which one you're looking at:

Extractable MEV: the theoretical value someone could have taken by reordering, inserting or censoring transactions. Nobody can measure this fully.
Extracted MEV (or MEV profit): what searchers actually captured, after gas and payments to builders. This is what early dashboards tried to count.
Realized extractable value (REV) or MEV revenue: what was paid out along the block supply chain, mostly to validators through MEV-Boost. This includes tips and priority fees, so it's bigger than searcher profit.

On top of that, every tracker classifies transactions with its own rules. A sandwich hidden inside a complex multi-hop route, or an arbitrage whose other leg happened on Binance, can easily be missed. The ESMA report makes this point directly: the published figures "are not necessarily reflective of the actual MEV revenues" because of these limits.

How much MEV was extracted before the Merge?

At least several hundred million dollars, and the counts are explicitly lower bounds. The term itself comes from the 2019 paper Flash Boys 2.0, which originally called it "miner extractable value" because miners ordered blocks before Ethereum moved to proof of stake.

The first large public measurement was Flashbots' MEV-Explore v0. After scanning every block from January 1, 2020, Flashbots found at least $314 million of extracted MEV across more than 1.3 million transactions, plus about $4.5 million in gas wasted on failed MEV attempts. Monthly extraction grew from about $180,000 in January 2020 to $57 million in January 2021. Flashbots called its figure a lower bound because it only covered about eight major DeFi protocols and their clones.

A later academic study, Remeasuring the Arbitrage and Sandwich Attacks of MEV in Ethereum, puts the total at around $675 million extracted before September 2022, which is when the Merge happened.

How much MEV is there on Ethereum since the Merge?

About $1 billion to $1.3 billion in revenue from September 2022 to early 2025, depending on the source. The ESMA risk analysis published on July 1, 2025 collected the main estimates in one place:

Source (as cited by ESMA)	What it measures	Period	Figure
Flashbots Transparency Dashboard	Value paid to validators in MEV-Boost blocks (REV)	Sep 2022 to early Jun 2024	526,207 ETH, about $1.1bn
EigenPhi	MEV revenue (arbitrage + sandwich)	Dec 2022 to Jan 2025	$963mn
EigenPhi	MEV profit (arbitrage + sandwich)	Dec 2022 to Jan 2025	$417mn
Sorella Labs	MEV revenue incl. CEX-DEX arbitrage	Sep 2022 onward	$1.26bn
MEV-Explore v0 (Flashbots)	Extracted MEV, lower bound	Jan 2020 to early 2021	$314mn+

Two things stand out to me. First, the gap between EigenPhi's revenue ($963 million) and profit ($417 million) shows how much of the gross goes to gas and builder payments. Second, ESMA notes that Flashbots' REV figure is only about 0.12% of the roughly $930 billion traded on Ethereum DEXes over the same period. MEV is a large absolute number and a small slice of the volume.

Which types of MEV make the most money?

Arbitrage has been the biggest category, and a few operations take most of the revenue. ESMA's EigenPhi-based figures show arbitrage MEV averaging $20.2 million a month in revenue against $16.8 million for sandwiches, with arbitrage also costing less to run.

The distribution is very uneven:

One study cited by ESMA (Wahrstätter et al., 2023) found that 20% of MEV operations accounted for 72% of total revenue between September 2022 and May 2023.
MEV spikes during market stress. ESMA notes that the USDC de-peg after Silicon Valley Bank's collapse in March 2023 pushed MEV revenue up about 8,000% over three days against that month's baseline, and the August 2024 sell-off caused a jump of about 700%.

That matches what I see running bots. Most days are quiet and competitive; a handful of volatile days produce a big share of the year's opportunities. If you want the strategy-by-strategy breakdown, I covered it in how MEV bots make money on Ethereum.

How much do users lose to sandwich attacks?

Tens of millions of dollars a year, though the trend is down. A sandwich attack is when a bot buys right before your swap and sells right after it, so you get a worse price.

Cointelegraph Research analyzed more than 95,000 sandwich attacks from November 2024 to October 2025 using EigenPhi data:

Monthly sandwich extraction fell from nearly $10 million in late 2024 to about $2.5 million by October 2025, even as DEX volume grew.
Traders still lost about $60 million a year, while attackers kept a profit margin of only about 5%. Builders captured most of the rest as fees.
Attacks stayed at 60,000 to 90,000 a month, and roughly 70% were tied to a single searcher, jaredfromsubway.eth.
The average profit per sandwich was just over $3.

The likely reason for the drop is that more swaps now go through private routes. If you trade on Ethereum, turning on MEV protection is still the single easiest way to stay out of these numbers.

Who actually keeps the MEV?

Mostly validators and builders, not the bots. Under MEV-Boost, searchers compete to get bundles into blocks, builders compete to win each slot, and the winning builder pays the proposer. Competition pushes searchers to bid away most of what they find. You can see it in the figures above: EigenPhi counted $417 million of profit out of $963 million in revenue, and sandwich bots in the Cointelegraph data kept only about 5%. ESMA also cites Heimbach et al., who estimate that 85% to 95% of blocks have been built through MEV-Boost since November 2022.

Block building itself is concentrated: ESMA points to reports that two builders, Beaverbuild and Titan, produced about 88% of Ethereum blocks in October 2024. You can watch current builder and relay shares on mevboost.pics.

The practical takeaway: a billion dollars of MEV does not mean a billion dollars of bot profit. A new MEV searcher is competing for the thin margin that's left after builders and validators are paid.

Why don't MEV numbers match between trackers?

Because they count different things, cover different strategies and convert ETH to dollars at different prices. The main reasons I see:

Definitions. REV includes validator tips; profit excludes gas and builder payments. These can differ by more than 2x.
Coverage. CEX-DEX arbitrage, where one leg is on a centralized exchange, is invisible on-chain. Sorella's figure tries to include it; most dashboards don't.
Classification. Heuristics miss complex sandwiches and mislabel some normal trades. The 2024 remeasurement paper exists because earlier rule-based methods produced false positives and negatives.
Pricing. A figure in ETH turns into very different dollar amounts depending on whether it's priced at the time of each block or today.
Dead dashboards. Flashbots stopped updating its Transparency Dashboard in June 2024, so newer totals rely on third-party trackers like EigenPhi.

When I quote an MEV number, I try to state the source, the period and whether it's revenue or profit. I'd treat any article that gives a single total without those three things with suspicion.

Frequently asked questions

How much MEV is extracted on Ethereum per year?

There's no official yearly figure. Flashbots' REV of about $1.1 billion over roughly 21 months after the Merge works out to several hundred million dollars a year in value paid to validators, with much less ending up as searcher profit.

Is MEV still profitable in 2026?

For a small number of well-built bots, yes, but margins are thin. Sandwich bots kept only about 5% of what they extracted in the year to October 2025, and most revenue goes to builders and validators.

What is the biggest MEV bot on Ethereum?

By sandwich count, jaredfromsubway.eth has been the best-known for years, linked to roughly 70% of Ethereum sandwich attacks between November 2024 and October 2025 in EigenPhi data.

Where can I see live MEV data?

EigenPhi tracks arbitrage and sandwich activity, mevboost.pics shows MEV-Boost relay and builder shares, and Flashbots' MEV-Explore covers the pre-Merge period.

What does MEV stand for?

Maximal extractable value: the value that can be captured by including, excluding or reordering transactions in a block. It was first called miner extractable value, before Ethereum moved to proof of stake.

The headline totals make MEV look like easy money, but the data says the opposite: it's a crowded, winner-takes-most market. If you have a specific strategy in mind and want it built properly, that's the kind of work I take on through custom MEV bot development at MEV DAO.