MEV (Maximal Extractable Value)
Executive summary
Maximal Extractable Value, originally framed as Miner Extractable Value in 2019, describes the surplus that block producers and their delegates can capture by ordering, including, or censoring transactions inside a block. By 2026, MEV has matured from a frontier curiosity into a multi-hundred-million-dollar institutional industry with a layered supply chain: searchers, builders, relays, validators, and orderflow auctions. Ethereum cumulative MEV-Boost payments crossed nine-figure annual run-rates, while Solana built parallel infrastructure through Jito, and Layer-2s wrestled with their own variants. The dominant tension is no longer whether MEV exists but who captures it: protocols, users, validators, or specialized intermediaries. Regulatory scrutiny, encrypted mempools, and proposer-builder separation reform are reshaping the landscape faster than retail users perceive.
Origin and mechanism
MEV emerges from the basic property of blockchain block production: whoever controls the ordering of transactions in a block can choose which to include, exclude, and where to place them. In 2019, Daian, Goldfeder, and Breidenbach formalized the concept in Flash Boys 2.0, observing arbitrage bots competing in a priority gas auction so intense that gas prices became a leading indicator of profitable trading opportunities. The canonical taxonomy splits extraction into atomic and statistical categories. Atomic MEV resolves entirely within a single block: DEX-DEX arbitrage that closes a price gap, liquidations on Aave or Compound that seize undercollateralized positions, and sandwich attacks where a searcher front-runs and back-runs a victim swap to extract slippage. Statistical or non-atomic MEV plays out over multiple blocks: long-term basis trades, oracle-update timing, and CEX-DEX arbitrage where the searcher cannot atomically guarantee profit. The economic primitive is simple: a public mempool plus deterministic block-by-block ordering plus AMMs with constant-product price impact create a queue everyone can see, and whoever pays the highest priority fee or gives the validator a private kickback wins. The mechanism is invariant across chains; only the institutional wrapper differs.
The MEV supply chain in 2026
Modern MEV on Ethereum runs through a four-tier pipeline. Searchers detect opportunities in pending transactions or block state and craft bundles of ordered transactions. Builders aggregate searcher bundles plus public mempool transactions into candidate blocks, optimizing for total payment to the proposer. Relays connect builders to validators and enforce the auction's commit-reveal property: validators cannot peek at block contents before committing to propose, preventing them from stealing searcher profits. Validators (proposers) sign whichever block pays them the most. By 2026, this stack moved over 90% of Ethereum blocks, with cumulative MEV-Boost payments to validators crossing the multi-billion-dollar mark since launch. The builder market has consolidated dramatically: a small handful of dominant builders such as Beaverbuild, Titan Builder, and rsync routinely produce 80%+ of blocks, raising legitimate concerns about builder-level censorship and centralization. Solana runs a structurally different model through Jito, where validators run modified clients that auction off block-space slots; Jito tips emerged as a meaningful contributor to validator yield, and Jito Labs' JTO token captures part of that flow. Layer-2s have followed divergent paths: rollups with centralized sequencers (Arbitrum, Optimism, Base) currently capture all MEV at the sequencer, while based rollups and shared sequencer designs (Espresso, Astria) attempt to redistribute it.
Sandwich attacks, JIT liquidity, and user harm
Sandwich attacks remain the most visible user-facing MEV. A searcher observes a large incoming swap on an AMM, front-runs it with a buy that pushes price upward, lets the victim's swap execute at the worsened price, and back-runs with a sell. The victim absorbs the slippage as a tax to the sandwicher. By 2026, retail-facing aggregators (1inch, CowSwap, UniswapX) deliberately route around vulnerable pools or use signed off-chain orders that close the front-run window. Just-in-time liquidity (JIT) is a more subtle form: a searcher mints a concentrated liquidity position immediately before a large swap, captures a disproportionate share of the swap's fees, then burns the position. JIT is debated as either parasitic to passive LPs or beneficial because it tightens spreads. The empirical answer depends on the pool: high-fee pools see JIT crowd out organic LPs, while low-fee blue-chip pools tolerate it. CowSwap's batch auction mechanism explicitly socializes the surplus that would otherwise flow to sandwichers back to the swap originators, and by 2026 had grown to multi-billion-dollar monthly volume on the strength of that mechanism. mev-blocker, an RPC endpoint that shields user transactions, became the default endpoint for several wallet integrations and rebated meaningful capture back to its users.
Flashbots, SUAVE, and the search for protocol-level MEV
Flashbots remains the pivotal organization in the space, having shipped MEV-Boost in 2022, MEV-Share in 2023, and the SUAVE testnet in 2024. MEV-Share inverted the traditional model by giving users the option to selectively reveal their transactions to searchers in exchange for a share of any MEV those searchers extract from them, formally turning MEV into a user-side rebate channel. Adoption was modest in 2024 but accelerated in 2025-2026 as wallet integrations matured. SUAVE (Single Unifying Auction for Value Expression) is the more ambitious bet: a separate decentralized chain dedicated to expressing user preferences and producing blocks for any rollup or L1, with an encrypted mempool and decentralized builders. SUAVE mainnet launched in stages through 2025, and by Q2 2026 had begun producing blocks for several Ethereum L2 testnets, with mainnet integration on Optimism's Superchain projected for late 2026. The political question is whether SUAVE replaces or augments the existing builder oligopoly. Competing efforts include encrypted mempools through Shutter Network's threshold-encryption approach (deployed on Gnosis Chain) and TEE-based fair ordering through projects like Flashbots' BuilderNet.
Who profits and who pays
Quantifying MEV flows by 2026, the rough waterfall on Ethereum is: searchers identify opportunities, build bundles, and pay between 80% and 99% of gross profit to builders via priority fees and direct payments; builders pay roughly 90%+ of that downstream to validators; validators pass it through to LSTs and stakers as yield uplift. The aggregate user tax was estimated by EigenPhi and Flashbots data at $500M-$1B annualized at peak market conditions, dominated by sandwich attacks and toxic CEX-DEX flow. Importantly, the largest MEV recipients in absolute dollar terms are the staking pools (Lido, Coinbase, Binance, Kiln) whose validator sets disproportionately benefit from MEV-Boost payments. Searchers operate on razor-thin margins because of the auction, with profitable searchers tending to be specialized firms running co-located infrastructure (Wintermute, Symbolic Capital Partners, Manifold). Solo retail searchers are functionally extinct in atomic arbitrage; the remaining niche is novel strategy types where the auction has not yet equilibrated. On the user side, the question of incidence is live: rebate mechanisms (MEV-Share, mev-blocker, CowSwap surplus) have started to recapture meaningful flow, but most retail users still bleed slippage they do not perceive.
Recent milestones (2024 to 2026)
The 2024-2026 window was unusually active. EIP-7251 (MaxEB) raised validator effective balance, indirectly affecting MEV-Boost economics by reducing validator count. Restaking on EigenLayer introduced new MEV-adjacent primitives: AVSs that promise pre-confirmations or fair ordering as cryptoeconomic services. The Pectra hardfork in 2025 included subtle changes to gas pricing that affected sandwich economics. On the orderflow side, Uniswap shipped UniswapX, a Dutch-auction filler-based protocol that essentially auctions user orders off-chain, with fillers competing on price. By 2026, UniswapX was clearing tens of billions of dollars monthly, with clear evidence that fillers were paying back surplus to users versus the public-mempool baseline. Robinhood, Coinbase, and other CeFi venues launched their own private mempool and orderflow auction products, sparking debate about whether centralized orderflow concentration creates new monopolies. Solana saw continued growth in Jito tips, briefly exceeding $100M monthly during peak meme-coin activity in late 2024, with steady-state in 2026 in the $30-50M monthly range. The collapse of selected meme-coin flow on Solana in 2025 illustrated how sensitive MEV revenue is to retail trading volume.
Risks, open questions, and second-order effects
The most-discussed risk is builder centralization. If 3-4 firms produce >90% of Ethereum blocks, they hold a de facto censorship lever; the post-OFAC-sanctions discourse in 2022-2023 made this concrete when several major builders began excluding Tornado Cash transactions. PBS reform, encrypted mempools, and inclusion lists are the technical responses. A second risk is the validator-side capture: large staking pools accrue disproportionate MEV revenue, accelerating stake concentration. A third is proposer-builder collusion, where a builder and proposer cut the relay out and split the surplus directly, breaking MEV-Boost's commit-reveal guarantee. Empirical instances of this have been documented but remain rare. A fourth is the new attack surface introduced by restaking and pre-confirmation services: an AVS promising fast pre-confs effectively sells access to ordering, which can be re-auctioned downstream. The deeper question is whether MEV is a permanent rent on user trades or an artifact of current AMM and mempool design that can be engineered down toward zero through better matching, encryption, and intent-based architectures.
Regulatory treatment and outlook through 2027
Regulators have begun to engage but have not converged on a model. Early U.S. enforcement actions in 2024 began drawing lines around what counts as market abuse in the MEV context; the CFTC has shown interest in MEV as a market-integrity issue. In Europe, MiCA does not directly address MEV, but the AMLR and DORA frameworks touch on the relay and builder layers. Looking ahead to 2027, three dynamics matter most. First, intent-based architectures (UniswapX, CowSwap, Across) and encrypted mempools (Shutter, SUAVE) will continue to chip away at the public-mempool MEV pool, redistributing capture toward users. Second, restaking and pre-confirmation marketplaces will create new MEV-adjacent revenue streams that look more like B2B contract terms than retail rents. Third, the unbundling of builders and proposers via execution tickets, MEV burn (a proposed mechanism to send a portion of priority fees to ETH burn), and PEPC (Protocol-Enforced Proposer Commitments) will move parts of the value capture into the protocol itself. The strategic outlook is that MEV does not disappear, but its institutional locus migrates from validators to either the protocol (via burn) or the user (via rebate). Searchers and builders consolidate into a handful of regulated, professionalized firms; retail-facing flow becomes increasingly intermediated by orderflow auctions; and the multi-billion-dollar annual extraction figure stops growing in absolute terms and starts redistributing in composition.
Watch points
- Ethereum builder market share concentration and any L1 PBS or MEV-burn proposal
- SUAVE mainnet adoption and its share of L2 block production
- Restaking-driven pre-confirmation marketplaces and their MEV implications
- Solana Jito tip volume and any fork or alternative auction infrastructure
TL;DR
MEV is now a measurable, multi-hundred-million-dollar industry on Ethereum and Solana, organized through a searcher-builder-relay-validator supply chain with major centralization and capture-redistribution debates ongoing into 2027.
Get DeFi Intel research in your inbox
Weekly long-form coverage of papers, incidents, jurisdictions, chains, tokens and the people building them. Free tier covers headlines; Pro adds the analyst-grade breakdowns.