MEV extraction is moving on-chain, and it's reshaping how block builders compete

MEV extraction — the practice of reordering, inserting, or censoring transactions within a block to capture value — used to be a dark art performed by searchers running private infrastructure and paying validators directly to win priority. In 2026, that value capture is increasingly happening through transparent, on-chain auction mechanisms, and the shift is quietly rewriting who profits from Ethereum's transaction ordering and by how much.

This matters beyond crypto-native circles because it determines execution quality for every on-chain trade, whether from a retail wallet swap or an institutional treasury rebalance. Worse MEV extraction means worse prices for ordinary users; better-designed on-chain markets mean users increasingly get a cut of the value their order flow creates instead of losing it entirely to extractors.

From Dark Forest to Proposer-Builder Separation

The original MEV problem — sandwich attacks, front-running, arbitrage extraction that taxed ordinary DEX trades — was largely unaddressed at the protocol level for years, forcing users into private RPC endpoints (like Flashbots Protect) just to avoid getting sandwiched. Ethereum's proposer-builder separation (PBS), which fully matured post-Merge, split the role of building a block (assembling the most valuable set of transactions) from proposing it (a validator simply picking the highest-paying block someone else built). This created a competitive marketplace of specialized block builders who compete on extraction efficiency, with validators auctioning off their block space to the highest bidder via relays.

PBS solved one problem — it decentralized the technical difficulty of building optimal blocks away from individual validators — but it didn't eliminate MEV. It industrialized it. A handful of sophisticated builders now capture the large majority of extractable value, and that concentration has its own competitive and censorship risks: several major relays have, at various points, censored transactions to comply with sanctions lists, raising exactly the neutrality concerns PBS was partly designed to route around.

What's Actually New in 2026

The current shift is toward moving the auction itself on-chain and encrypting order flow before it reaches builders, cutting off the information advantage that lets sophisticated searchers front-run visible mempool transactions. Encrypted mempool designs — using threshold encryption or trusted execution environments — keep transaction contents hidden until a block is committed, so a builder can't selectively reorder based on what a pending trade actually contains. This directly targets sandwich attacks, which depend entirely on builders or searchers seeing a trade before it lands.

Alongside encrypted mempools, intent-based architectures are gaining real production traction. Instead of submitting a fully-specified transaction, a user submits an "intent" — a description of a desired outcome, like "swap up to 1000 USDC for the best available ETH price within 30 seconds" — and a competitive market of solvers bids for the right to fill it optimally. This inverts the extraction dynamic: instead of searchers extracting value from a user's visible transaction, solvers compete against each other to offer the user the best execution, with the winning solver's margin coming from genuine routing efficiency rather than information asymmetry.

Order flow auctions (OFAs) are the third piece: wallets and applications now routinely auction their users' order flow directly to searchers and builders before submission, capturing a portion of the MEV that would otherwise be extracted invisibly and returning it to users as reduced fees or better execution prices. Several major wallet providers have integrated OFA mechanisms as a default, meaning users increasingly get paid for the extraction that used to simply cost them money.

Winners and Losers in the New Structure

Sophisticated searchers running pure front-running strategies against visible mempools are the clearest losers — encrypted mempools and intent architectures are specifically designed to eliminate the information edge their business model depended on. Validators see modest gains from PBS auction competition but remain largely price-takers, since builder competition — not validator leverage — determines how much of the extracted value flows back through the relay auction.

The clearest winners are retail and institutional traders whose order flow is now, for the first time, treated as an asset with negotiable value rather than a resource to be silently extracted from. Solvers and builders who invest in genuine execution-quality infrastructure rather than pure latency arbitrage are also positioned to capture a durable share of the market, since intent-based competition rewards routing sophistication over speed alone.

What This Means for Traders and Builders

Traders executing meaningful size on-chain should evaluate whether their wallet or DEX aggregator routes through an OFA or intent-based solver network rather than submitting directly to a public mempool — the difference in realized execution price on a volatile pair can be substantial, particularly during periods of high network congestion when sandwich attacks are most profitable. Developers building on-chain applications should treat MEV exposure as a design parameter from day one, not a problem to patch later: batch auctions, commit-reveal schemes, and intent-based order submission are now mature enough to integrate directly rather than bolted on as an afterthought.

Takeaways

• Check whether your wallet or DEX routes through an order flow auction or intent-based solver network — this alone can materially improve realized trade prices.
• Encrypted mempools eliminate the information asymmetry that sandwich attacks depend on; prioritize protocols and chains that have adopted them for high-value transactions.
• If you're building on-chain, design for MEV resistance from the start using commit-reveal or intent-based patterns rather than retrofitting protection later.
• Builder and relay concentration remains a real neutrality risk — track which relays your validator or staking provider uses and whether they've implemented censorship-resistance guarantees.
• Watch the solver market, not just the builder market — intent-based competition is where the next round of MEV value redistribution to users is happening.