Ethereum’s MEV-Boost relay market just flipped. New data from relay-monitoring dashboards shows that neutral, non-filtering relays now handle roughly 70% of daily block production, while relays that once enforced U.S. sanctions screening have shrunk to a sliver of the network. The shift caps a four-year retreat from the censorship peak that followed the 2022 Merge, and it lands the same week Ethereum core developers held a dedicated call on encrypting the mempool itself. Together, the two stories describe a protocol quietly re-engineering how it resists censorship, block by block, EIP by EIP.
What Changed: Neutral Relays Take Over MEV-Boost
For a 24-hour window ending August 29, 2026, monitoring data tracked 11,656 blocks built through MEV-Boost relays on Ethereum. Ultra Sound Relay, which runs a strictly neutral, non-censoring policy, built 4,094 of them, good for 35.1% of the day’s total. Titan Relay, also neutral, followed with 3,471 blocks and a 29.8% share. Add Aestus at 7.2% and the neutral camp alone accounts for more than 70% of blocks proposed that day.
bloXroute’s regulated relay, the main remaining OFAC-screening option, built 2,771 blocks for a 23.8% share. Flashbots’ own relay, the service that dominated the market immediately after the Merge, produced just 278 blocks, or 2.4% of the day’s total. Smaller relays including EthGas and Agnostic Gnosis each held under 1%. Over the trailing 30 days, the share of censoring relays fell 28%, a trend that dashboard operators describe as a steady, not sudden, drift toward neutral block building.
The Relay Market in Numbers
The table below captures the snapshot that triggered the coverage: a single 24-hour period near the end of August 2026, broken down by relay, filtering posture, and block share.
| Relay | Filtering posture | Block share (24h) | Blocks built |
|---|---|---|---|
| Ultra Sound Relay | Neutral | 35.1% | 4,094 |
| Titan Relay | Neutral | 29.8% | 3,471 |
| bloXroute Regulated | OFAC-screening | 23.8% | 2,771 |
| Aestus | Neutral | 7.2% | 838 |
| Flashbots Relay | OFAC-screening | 2.4% | 278 |
| EthGas + Agnostic Gnosis | Mixed | <1.7% combined | <204 combined |
Total blocks measured in the window: 11,656. Source: MEV-Boost relay monitoring data compiled via relayscan.io and cross-referenced with mevwatch.info, both of which track relay-level block production in near real time.
From 80% Compliant to 2.4%: A Four-Year Reversal
To understand why a 2.4% share matters, rewind to October 2022. Weeks after the Merge, MEV Watch data cited by The Block put 63% of Ethereum blocks through OFAC-compliant relays. By mid-December that year, roughly 70% of proposed blocks routed through relays enforcing sanctions screening, and Flashbots alone had built more than 75% of MEV-Boost blocks since the transition to proof-of-stake. Some tallies during November 2022 put the compliant share as high as 80%, depending on whether the count covered all Ethereum blocks or only MEV-Boost traffic.
The retreat started fast. By May 10, 2023, The Block reported the daily share of OFAC-compliant blocks had dropped to 27%, down from that 80% November peak. A separate outlet put the same-day figure at 29%. Validator-side research from Metrika around the same period found seven of eleven active relays did not filter for OFAC compliance at all, even though MEV-Boost adoption among validators had climbed past 90%. Non-conforming transactions still landed on-chain 99.52% of the time, versus 95.32% for OFAC-screened flow, though they waited longer in the mempool: a median 22 seconds against 8 seconds for compliant routing.
Censorship Share, Then and Now
| Period | OFAC-compliant / censoring block share | Primary driver |
|---|---|---|
| October 2022 (post-Merge) | ~63% | Flashbots dominance, single dominant relay |
| November-December 2022 | ~70-80% | Validators defaulting to compliant relays |
| May 2023 | ~27-29% | Growth of neutral relay alternatives |
| 2023-2024 (validator survey) | 7 of 11 relays non-filtering | Multi-relay validator subscriptions |
| August 29, 2026 | ~26.2% (bloXroute + Flashbots combined) | Ultra Sound and Titan relay growth |
Read across five rows, the pattern is one direction only: down. What began as a near-universal compliance regime in late 2022 has settled into a market where roughly three out of every four blocks route through relays that make no attempt to screen transactions against U.S. Treasury sanctions lists.
Why Censorship-Resistance Became a Market Signal
The original scare, in 2022, centered on Tornado Cash. The U.S. Treasury’s Office of Foreign Assets Control sanctioned the Ethereum mixing service that August, and relays including Flashbots responded by excluding transactions tied to sanctioned addresses. Critics called it a stress test Ethereum was failing in real time: a chain marketed as censorship-resistant was routing most of its blocks through operators willing to drop specific transactions on request from a government agency.
What happened next is the part that gets less attention than the initial scare. Validators, who choose which relay to subscribe to, increasingly picked multiple relays rather than one. A validator running Ultra Sound, Titan, and Flashbots simultaneously takes whichever relay returns the highest-paying block, and a neutral relay competing purely on payout tends to win that race more often than a filtering one, since it can include transactions a compliant relay must exclude. That market mechanic, not a protocol upgrade, did most of the work in pushing censorship share down from 80% to under 30% within a year, and toward the current low-20s combined share for the two remaining screening relays.
The Parallel Story: Encrypting the Mempool
Relay-level neutrality solves one form of censorship. It does nothing for the separate problem of front-running, where bots watch Ethereum’s public mempool and insert their own trades ahead of a victim’s pending transaction. Ethereum protocol researchers and core developers convened a dedicated “Encrypt the Mempool” call around August 19, 2026, to discuss hiding transaction contents until their position in a block is already fixed. The pitch: if a bot cannot read a transaction before it lands, it cannot sandwich it, and MEV extraction that current estimates put in the billions of dollars annually loses its main data source.
The August call built on work that had been running for most of 2026. Two concrete Ethereum Improvement Proposals now anchor the effort: EIP-8184, known as LUCID, and EIP-8105, the Universal Enshrined Encrypted Mempool. Neither has a confirmed mainnet activation date. Both represent the clearest sign yet that Ethereum’s core dev community sees encrypted transaction flow, not just relay competition, as the next lever against MEV extraction.
Inside EIP-8184: LUCID
EIP-8184, dated March 4, 2026 with updates continuing into late August, specifies a public encrypted mempool built around commit-before-reveal inclusion. A user submits an encrypted transaction, the network commits to including it in a specific block position, and only after that commitment is locked does the transaction’s content become visible. The proposal is deliberately encryption-agnostic. It does not mandate a specific cryptographic scheme, which keeps the door open for threshold encryption, multi-party computation, or trusted execution environments to fill that role later. A March 2026 Ethereum Magicians working-group call, “Encrypt The Mempool #2,” reviewed the Lucid design alongside early implementation demonstrators and simulation tooling.
EIP-8105: The Universal Enshrined Encrypted Mempool
EIP-8105 takes a broader, protocol-enshrined approach. Rather than specifying one encryption method, it defines an interface that supports multiple decryption-key providers: threshold encryption schemes such as Shutter, MPC committees, trusted execution environments, delay encryption, and potentially fully homomorphic encryption further down the road. The design goal, according to the Ethereum Magicians thread tracking the proposal, is to stop real-time reordering and weak censorship while still allowing legitimate backrunning and arbitrage once a transaction has already been included. That distinction matters: the proposal targets predatory reordering, not all forms of MEV, since some arbitrage activity actually improves price discovery across DeFi markets.
SUAVE and Flashbots’ Long Game
Flashbots, the relay operator now down to a 2.4% daily block share, has spent the past two years building something bigger than a relay business. SUAVE, short for Single Unifying Auction for Value Expression, is Flashbots’ attempt at an MEV-aware, privacy-first encrypted mempool that functions as a shared sequencing layer across chains rather than a single-chain relay. Flashbots co-founder Phil Daian has described SUAVE as designed to hand MEV profit back to end users rather than concentrate it in the hands of specialized builders. The system has been in production testing through 2025 and 2026, positioning Flashbots to remain relevant in MEV infrastructure even as its legacy relay business shrinks toward irrelevance on the daily leaderboard.
The strategic logic is straightforward. If encrypted mempools and cross-chain block building become the standard architecture, owning that layer matters more than owning one relay among a half-dozen competitors. Flashbots effectively ceded the compliant-relay market rather than fight for share it was losing anyway, and redirected resources toward the infrastructure it expects to matter once EIP-8105 or a similar proposal actually ships.
Threshold Encryption: The Shutter Network Roadmap
Shutter Network has published the most detailed technical roadmap for getting Ethereum from today’s public mempool to a fully encrypted one, built around threshold encryption. In that scheme, a decryption key gets split across multiple independent parties, and only a threshold-sized coalition of them can reconstruct it, which prevents any single operator from decrypting transactions early. Shutter’s plan runs in three steps. First, an eth_sendEncryptedTransaction RPC method layered onto existing infrastructure like MEV Blocker. Second, integration into proposer-builder separation workflows so relays and proposers can commit to encrypted transactions while preserving inclusion guarantees. Third, what the roadmap calls in-protocol Shutterization, where the Beacon Chain itself handles threshold decryption natively.
None of the three steps has a fixed calendar date attached. But the roadmap gives EIP-8105 and EIP-8184 a concrete path to production rather than leaving them as pure research exercises, and it explains why the August 19 core-dev call treated encrypted mempools as a near-term engineering problem instead of a distant theoretical one.
Competitive Comparison: Three Approaches to the Same Problem
Relay neutrality, LUCID, and the Universal Enshrined Encrypted Mempool are not competing products in the traditional sense, but they do represent three distinct layers at which Ethereum can fight MEV and censorship, each with different tradeoffs.
| Approach | Layer | Solves censorship? | Solves front-running? | Status (Sept 2026) |
|---|---|---|---|---|
| Neutral relay competition | Off-protocol, market-driven | Partially, no enforcement guarantee | No | Live, ~70% share |
| EIP-8184 (LUCID) | Protocol-level, commit-before-reveal | Yes, by design | Yes | Active spec, no mainnet date |
| EIP-8105 (Universal EEM) | Protocol-enshrined, multi-scheme | Yes, by design | Yes | Under working-group review |
| SUAVE (Flashbots) | Cross-chain sequencing layer | Partially, depends on adoption | Yes, within SUAVE flow | Production testing 2025-2026 |
The market-driven relay shift is live today and cost nothing in protocol changes, but it depends entirely on validator preference continuing to favor neutral operators. Nothing stops a future regulatory push or a change in validator incentives from reversing it. The EIP-based approaches are slower and require broad developer and client-team consensus, but they would bake censorship resistance into the protocol rather than leave it to relay market dynamics.
Market Impact: What Shifts for Builders, Validators, and DeFi Users
For block builders, the relay shift changes where the profitable order flow lives. Builders that route primarily through Ultra Sound and Titan now compete for a much larger share of total blocks than those still routing through bloXroute’s regulated option or Flashbots. Builder concentration remains a separate concern: earlier 2026 research found MEV-focused builders account for close to 80% of Ethereum blocks overall, with individual operators like beaverbuild.org and Titan Builder responsible for outsized shares of monthly MEV revenue, so the relay-neutrality trend does not by itself fix the deeper builder-concentration problem.
For validators, the calculation is mostly about yield. A validator subscribed only to a compliant relay forgoes blocks that a neutral relay would have paid more for, since compliant relays must reject some profitable-but-sanctioned transaction bundles. Multi-relay subscription, now standard practice, lets a validator capture the higher payout regardless of which relay produces it. For everyday DeFi users and traders, the direct effect is smaller: encrypted-mempool proposals matter more to them long-term, since those target the sandwich attacks and front-running that hit ordinary swaps, not the sanctions-screening question that dominates the relay-share story.
Historical Context: How MEV Went From Dark Forest to Public Dashboard
MEV extraction predates the term itself. Early Ethereum researchers described the pre-2020 mempool as a dark forest where visible pending transactions got hunted by bots within seconds. Flashbots launched in 2020 to make that extraction transparent and auction-based rather than a free-for-all fought through gas-price spam. The Merge in September 2022 formalized proposer-builder separation across the validator set through MEV-Boost, and within weeks that same infrastructure became the vehicle for OFAC-driven censorship once Tornado Cash was sanctioned.
What has changed since is less about MEV disappearing and more about who controls it. Ethereum.org’s own documentation still defines MEV as the value a block producer can extract beyond the standard block reward and gas fees by choosing transaction inclusion and ordering. That has not changed. What has changed is the plumbing: from one dominant, compliant relay in 2022, to a fragmented market of neutral competitors today, to a protocol-level encryption push that could make relay-level filtering moot within a few years if EIP-8105 or a similar proposal reaches mainnet.
What Comes Next: Five Predictions
- Neutral relay share keeps climbing past 75% within the next two quarters as validator tooling makes multi-relay subscription the default rather than a manual configuration choice.
- Flashbots’ legacy relay share falls further, potentially below 1%, while the company’s SUAVE effort absorbs the engineering resources once dedicated to the compliant relay product.
- Neither EIP-8184 nor EIP-8105 reaches mainnet activation before the end of 2027, given the absence of a confirmed hard-fork slot for either proposal as of September 2026.
- Regulatory attention returns to relay-level censorship if a future high-profile sanctioned-address transaction clears through a neutral relay and draws public scrutiny, similar to the 2022 Tornado Cash episode.
- Shutter Network’s RPC-based encrypted transaction option sees early production pilots on layer-2 networks before any layer-1 proposal reaches mainnet, since L2 sequencers can iterate faster than Ethereum’s base-layer consensus process.
The Regulatory Wildcard
Nothing in the current relay-share trend is protocol-enforced, which is exactly what makes it fragile. OFAC’s Tornado Cash sanctions triggered the original compliance wave through informal pressure on relay operators rather than a law requiring specific technical behavior from Ethereum infrastructure. A future enforcement action, whether against a specific address, a mixing protocol, or a DeFi front-end, could just as easily push validators back toward compliant relays if the legal exposure for running a neutral one changes. Chain-analytics firms including Chainalysis continue to track sanctioned-address flows across chains regardless of which relay processed the underlying block, which keeps that enforcement lever available to regulators at any point.
That is the core argument for the EIP-8105 camp: a market-driven neutral-relay majority can reverse itself in weeks if incentives change, while a protocol-level encrypted mempool would make relay-level filtering technically impossible regardless of the regulatory climate. Whether Ethereum’s core developers move fast enough to lock that in before the next enforcement cycle is the open question hanging over both stories.
Frequently Asked Questions
What is MEV-Boost and why does relay choice matter?
MEV-Boost is the software most Ethereum validators run to outsource block construction to specialized builders through relays. The relay a validator subscribes to determines which blocks it sees and, in the case of filtering relays, which transactions get excluded before the validator ever sees them.
What does neutral mean for an MEV relay?
A neutral relay builds and forwards blocks without screening transactions against sanctions lists such as OFAC’s Specially Designated Nationals list. Ultra Sound Relay, Titan Relay, and Aestus all currently operate on a neutral basis, according to relay-monitoring data.
Did Flashbots stop operating its relay?
No. Flashbots’ relay is still active and OFAC-screening, but its daily block share has fallen to roughly 2.4% as validators route more traffic to neutral competitors. Flashbots has shifted strategic focus toward SUAVE, its cross-chain encrypted sequencing project.
What is an encrypted mempool and how would it stop MEV?
An encrypted mempool hides a pending transaction’s contents until its position in a block is already committed. Since MEV bots rely on reading pending transactions to front-run or sandwich them, encryption removes the information those bots currently exploit.
What are EIP-8184 and EIP-8105?
EIP-8184, known as LUCID, defines a commit-before-reveal encrypted mempool without mandating a specific encryption scheme. EIP-8105, the Universal Enshrined Encrypted Mempool, proposes a broader protocol-level interface supporting threshold encryption, MPC, and trusted execution environments. Neither has a confirmed mainnet activation date as of September 2026.
Is the drop in censoring relays permanent?
Not necessarily. The shift is driven by validator preference for higher-paying neutral relays, not by a protocol rule, so it could reverse if regulatory pressure on relay operators increases. That fragility is the main argument proponents make for moving censorship resistance into the protocol itself through proposals like EIP-8105.
How can I check current MEV-Boost relay market share myself?
Dashboards including relayscan.io and mevwatch.info publish near-real-time relay block counts and filtering postures, updated daily.
Does this affect Bitcoin or only Ethereum?
This story is specific to Ethereum’s proposer-builder separation architecture. MEV exists in other forms across various blockchains, but MEV-Boost relays and the associated censorship debate are Ethereum-specific concepts tied to its post-Merge validator design.
