MEV stands for Maximal Extractable Value, the extra value a validator or sequencer can capture by reordering, inserting, or censoring transactions. Head to the Binance Official Site to check spot prices for on-chain DEX assets, download the Official Binance App to try on-chain trading, and if you're on iPhone, read the iOS installation guide first. The short answer: Ethereum mainnet has the largest MEV footprint (an annualized $1.0-1.5 billion), Solana's MEV has been standardized through Jito (an annualized $400-600 million), and MEV on L2s is still small but growing fast. For everyday users, MEV mostly shows up as slippage losses from sandwich attacks, which drain roughly $500 million a year from retail traders.
MEV can be split into five categories based on how it's extracted.
| Type | Mechanism | Who's hurt |
|---|---|---|
| Arbitrage | Price gaps across DEXs | No one (a normalized profit) |
| Sandwich attacks | Front and back a user's trade | Retail traders |
| Liquidations | Triggering lending-protocol liquidations | Leveraged users |
| Front-running | Spotting a mempool trade and inserting ahead | High-priority transactors |
| Back-running | Acting right after a significant transaction | The arbitrageur itself (no victim) |
Sandwich attacks do the most damage to ordinary users. When a user submits a large swap on a DEX, an attacker watching the mempool inserts a same-direction trade ahead of it with higher gas (pushing the price up), the user's trade fills at a worse price, and the attacker immediately swaps back for a profit. The whole sequence costs the user several extra percentage points of slippage.
Ethereum mainnet has by far the largest MEV market. Data from Flashbots, a research outfit dedicated to MEV, shows the following.
| Year | Total MEV ($100M) | Sandwich attack share | Validator share of MEV revenue |
|---|---|---|---|
| 2021 | 7.5 | 35% | 90% (PoW miners) |
| 2022 | 6.8 | 30% | 85% (late PoS) |
| 2023 | 9.2 | 28% | 72% (MEV-Boost live) |
| 2024 | 12.5 | 25% | 70% |
| 2025 | 14.3 | 23% | 68% |
MEV extraction on Ethereum mainnet mostly runs through infrastructure like Flashbots Protect and MEV-Boost. MEV-Boost is a relay service built by the Flashbots team that lets validators "sell" block-building rights to specialized searchers, who bundle MEV arbitrage transactions while validators earn a fixed fee. This PBS (Proposer-Builder Separation) design makes the distribution of MEV profit far more transparent.
MEV on Solana has been standardized through the Jito Block Engine. Jito turns MEV extraction into a formal process: searchers submit bundles (groups of related transactions) through Jito's auction system, the highest bidder wins the right to have its bundle included, and validators pack blocks according to Jito's rules.
Jito's core innovation is redistributing MEV revenue back to stakers. 95% of MEV income goes to Jito validators and their stakers, with the remaining 5% going to the Jito DAO. That design lets ordinary SOL holders share in the MEV upside too.
| Solana MEV data | Value |
|---|---|
| Total 2024 MEV | $520 million |
| Cumulative Jito MEV distributed | $500 million (through April 2026) |
| Sandwich attack share | 18% |
| Arbitrage share | 60% |
| Validator-searcher split | 95:5 |
Solana's MEV market is roughly a third the size of Ethereum mainnet's, but it's growing faster. On-chain DEX volume on Solana already surpassed Ethereum mainnet in 2025, and its MEV volume is projected to catch up to or even overtake Ethereum's in 2026.
MEV on L2s works differently than on L1 — most L2s run a single sequencer (Arbitrum, Optimism, and Base are all operated by a single company's sequencer), giving that sequencer full control over transaction ordering.
| L2 | 2024 MEV ($10K) | Sequencer operator | User protection mechanism |
|---|---|---|---|
| Arbitrum | 800 | Offchain Labs | FCFS (first-come-first-served) |
| Optimism | 450 | Optimism Foundation | Private mempool |
| Base | 600 | Coinbase | FCFS |
| zkSync Era | 220 | Matter Labs | FCFS |
Arbitrum orders transactions First-Come-First-Served — trades are processed in the order they reach the sequencer, which in theory eliminates sandwich attacks. In practice, though, network latency still gives large players and high-frequency traders an edge, so FCFS can't fully eliminate MEV.
Optimism's private mempool keeps transactions hidden before they're packed into a block, in theory removing the "spot a trade, then insert" attack vector. But a single sequencer can still prioritize transactions that pay higher gas fees.
Polygon PoS, a sidechain with 100+ validators, sees roughly $250 million a year in MEV (2025). But Polygon PoS lacks standardized infrastructure like Flashbots, so extraction there is more fragmented.
Polygon zkEVM runs on a single sequencer operated by Polygon Labs, and its MEV volume is small (about $5 million a year), similar to other L2s.
MEV on BSC runs about $600 million a year. But BSC's 21-validator setup concentrates MEV heavily — the top 5 validators capture over 60% of total MEV revenue. BSC extraction runs through infrastructure like bloXroute and 48 Club.
Sandwich attacks hurt retail traders more on BSC than on Ethereum mainnet, because PancakeSwap's liquidity depth is shallower than Uniswap's, which lowers the marginal cost of running a sandwich attack.
The core impact of MEV on retail traders is slippage loss. Research from Eigenphi and Flashbots shows that in 2024, sandwich attacks cost retail traders roughly $150 million on Ethereum mainnet and about $80 million on BSC.
The table below estimates expected MEV losses for swaps of different sizes across chains.
| Swap size | Ethereum mainnet (USDC/ETH) | Solana (Raydium) | BSC (PancakeSwap) |
|---|---|---|---|
| $1,000 | $0-3 | $0-2 | $0-5 |
| $10,000 | $5-25 | $3-15 | $10-50 |
| $100,000 | $30-200 | $20-150 | $50-300 |
| $1,000,000 | $200-1,500 | $150-1,000 | $300-2,500 |
Note: actual losses depend heavily on the liquidity depth of the swap pair, gas prices, and network congestion. Obscure tokens can see MEV losses far higher than the ratios above.
For ordinary users, the main protection tools are as follows.
Flashbots Protect: the standard protection for Ethereum mainnet users. Flashbots Protect offers a dedicated RPC that sends your transaction straight to the Flashbots Builder, bypassing the public mempool. Sandwich attackers who never see the transaction can't insert around it.
MEV Blocker: a free MEV-protection RPC built by the CoW Swap team, working similarly to Flashbots Protect but at no cost. Users just switch RPCs in MetaMask to turn it on.
Cow Protocol: a DEX built on batch auctions. Every user's trades in a block settle in a batch at a single uniform price, structurally eliminating sandwich attacks. Its liquidity depth is shallower than Uniswap's, though.
Jito (Solana): using a Jito-only RPC ensures your transaction is packed through the Jito Block Engine, avoiding the risk of getting sandwiched as a retail trader.
1inch Fusion / 1inch Pathfinder: 1inch's aggregated trading mode routes through a private mempool by default, cutting MEV losses for small and mid-sized trades.
Every major chain is researching "protocol-level MEV protection."
Ethereum: PBS (Proposer-Builder Separation) and MEV-Boost have standardized MEV extraction. The EIP-7732 proposal goes further by separating block building from proposing, distributing MEV profit more fairly.
Solana: the Jito model already returns 95% of MEV revenue to stakers, making it the most transparent MEV distribution mechanism today. Jito is an off-protocol component, though, and may eventually be integrated directly into the Solana protocol.
Polygon zkEVM: plans to introduce an "encrypted mempool," where transactions stay fully encrypted before being packed and only validators can collectively decrypt them, structurally eliminating front- and back-running.
Cosmos app chains: Skip Protocol, an MEV infrastructure layer on Cosmos, lets app chains opt into "protocol-level MEV," automatically routing all MEV revenue back to the chain's token holders.
Q: Can MEV be avoided entirely?
Not entirely, but it can be greatly reduced. Ordinary users who rely on protection tools like Flashbots Protect, MEV Blocker, or Cow Protocol can cut sandwich-attack losses by 80%+. But "victimless" MEV like cross-DEX arbitrage and liquidations will always exist.
Q: Do I need to worry about MEV when trading on an L2?
MEV volume on L2s is small, but a single sequencer is a latent risk. Arbitrum's FCFS ordering and Optimism's private mempool both push the MEV risk on any single trade close to zero. Small and mid-sized trades (under $10,000) on L2s generally don't need to worry about it.
Q: Is Jito necessary on Solana?
Not necessary, but recommended. A Jito-only RPC lets your trades share in the 95% MEV split (if you stake SOL) while cutting your risk of being sandwiched. Phantom wallet integrates Jito by default.
Q: Do MEV attackers face any legal consequences?
Technically, no. MEV extraction isn't illegal in most jurisdictions because it's essentially a legitimate use of the public mempool. Regulators haven't yet classified MEV extraction as illegal activity.
Q: Is MetaMask's default RPC safe?
MetaMask's default RPC offers no MEV protection. If you're making large swaps (over $5,000), it's strongly recommended to switch to a Flashbots Protect RPC or MEV Blocker RPC inside MetaMask.
Q: Is MEV on BSC worse than on Ethereum?
Proportionally, yes — MEV hurts retail traders more on BSC. That's because BSC's liquidity is shallower, users tend to trade larger amounts, and BSC lacks mature protection tools comparable to Flashbots Protect.
Q: Is there MEV risk when trading on a CEX?
Centralized exchanges (like Binance) don't have a public mempool, so sandwich attacks in theory don't happen there. But a CEX acting as its own market maker could extract value from order flow ("internal MEV") in ways users can't observe.
Q: Can on-chain trades be fully private?
Partly. Privacy chains like Aztec, Penumbra, and Namada use ZK proofs to achieve full transaction privacy, but their liquidity is extremely thin. For most users, the realistic option is using private-mempool services like Flashbots Protect to achieve "privacy from MEV attackers."
MEV is an inherent risk of on-chain DeFi trading that can't be fully eliminated. Ordinary users should use protection tools, but even so, large trades will still face some degree of MEV loss. This article is compiled from public data and does not constitute investment or trading advice. Please read our Risk Disclaimer.
Data sources: Flashbots Research, Eigenphi, Jito Foundation, and the EigenPhi MEV database (April 2026 data).