* Flashbots & MEV-Boost
Flashbots is the organization that built the infrastructure making Ethereum's post-Merge MEV market work. The Flashbots Relay connects professional block builders to validators via MEV-Boost, enabling PBS before Ethereum enshrines it in-protocol. The Flashbots RPC shields everyday users from sandwich attacks. And SUAVE is the cross-chain next generation that could make MEV nearly invisible to end users.
Flashbots Relay + Builder Architecture
The full Flashbots MEV-Boost pipeline - from user transaction to sealed block. Click any node to learn more.
Gas Price vs Flashbots Bundle Placement
Move the gas price slider to see how gas level affects where your bundle lands in the block - and how Flashbots affects your placement probability vs the vanilla mempool.
? Transaction Routing Comparison
How vanilla mempool, Flashbots, and Eden Network compare across key dimensions.
| Feature | Vanilla Mempool | Flashbots RPC | Eden Network | MEV Blocker |
|---|---|---|---|---|
| Sandwich Protection | None | Full | Full | Full |
| MEV Rebate to User | - | - | - | 25% rebate |
| OFAC Censorship | None | ? Filtered | None | None |
| Private to Searchers | Public | Private | Private | Private |
| Bundle Auction | - | Competitive | Competitive | OFA auction |
| Setup Required | None | RPC endpoint only | RPC endpoint only | None |
| Cross-Chain Support | - | SUAVE pending | Ethereum only | Ethereum only |
MEV-Boost Relay Market Share (2026)
SUAVE - The Next Generation
SUAVE (Single Unifying Action for Value Expression) is Flashbots' redesign of the entire MEV stack. Instead of submitting a raw transaction, users sign an intent - a structured preference statement - and competing solvers bid to satisfy it. The intent never hits the public mempool, the solver can execute across any chain, and MEV is minimized by design.
How Flashbots MEV-Boost works
MEV-Boost is the implementation of Proposer-Builder Separation that sits between validators and builders today. Before MEV-Boost, a validator running a vanilla execution client (Geth, Reth, Besu) would assemble its own block from the mempool - capturing the base block reward (~2-3% APR) but zero MEV. With MEV-Boost, the validator runs a sidecar process that queries a set of relays at slot proposal time, each of which returns the highest bid from connected builders. The validator signs the highest sealed bid's blinded header (seeing no block contents), and the relay reveals the body for block gossip.
The Flashbots Relay is the dominant implementation of this relay. When a builder like beaverbuild, rsync-builder, or EigenPhi's Penguin Build finds a profitable MEV opportunity - an arbitrage, a liquidation, or a sandwich - it constructs a block containing that bundle plus ordinary mempool transactions, and submits a bid to the Flashbots Relay. The relay validates the block (no invalid transactions, no MEV theft, no gas overuse), ranks it against competing builders, and forwards the highest bid to validators. The validator's MEV-Boost sidecar collects bids from all registered relays and picks the highest.
The trust model is deliberately constrained: the relay cannot steal the bundle because the validator signed the header before the relay revealed the body. However, the relay can equivocate - showing one header and then revealing a different block - which is why ePBS (enshrined PBS) is on the Ethereum roadmap: it moves the commit-and-reveal step into the consensus layer so the relay's role is minimized.
Flashbots and ePBS
Flashbots MEV-Boost is out-of-protocol PBS - it works because validators voluntarily run the MEV-Boost sidecar and relays voluntarily implement the sealed-bid auction. ePBS (enshrined PBS) is the Ethereum improvement proposal to make PBS a protocol-level requirement: the consensus layer would run the auction, builder attestations would replace relay attestation, and the trusted relay would be replaced by a cryptographic commit-and-reveal built into the beacon chain.
The connection is that Flashbots has been simultaneously the biggest proponent of MEV-Boost (demonstrating that PBS works) and the biggest advocate for ePBS (because MEV-Boost's trusted-relay assumption is the weakest link in the design). Flashbots has published extensive research on ePBS protoype implementations, including the builder-api spec that ePBS is expected to extend. In this sense, Flashbots is building the transitional infrastructure that makes ePBS adoption politically and technically feasible when it lands.
Key concepts
- Flashbots RPC / Protect
- The Flashbots RPC endpoint is a private transaction submission endpoint: your transaction is routed directly to Flashbots' network of professional searchers without ever entering the public mempool. Searchers compete to include your transaction in the highest-value bundle, and the resulting block is submitted to the Flashbots Relay. For DEX swaps, this eliminates 100% of sandwich attack surface while preserving the economic benefits of MEV extraction. The trade-off is that Flashbots (like most relays) implements OFAC transaction filtering.
- MEV-Boost and the relay auction
- MEV-Boost is the validator sidecar software, open-sourced by Flashbots, that implements the PBS auction. When a validator is selected to propose a block, MEV-Boost queries registered relays, collects sealed bids, and returns the highest-bid block payload to the execution client for block assembly and gossip. The sidecar communicates via the builder API spec that Flashbots pioneered and that ePBS will extend.
- Builder market and market share
- The block builder market is concentrated among a handful of professional firms: beaverbuild, rsync-builder, Titan Builder, EigenPhi's Penguin Build, and Flashbots' own builder service. These builders run continuous optimization: simulating thousands of candidate transaction orderings, computing the resulting block value, and submitting increasing bids to relays. Builder competition is what drives up the bids that validators receive - and what ultimately determines the MEV revenue split between searchers, builders, and validators.
- SUAVE and intent-based architecture
- SUAVE reimagines the user-facing MEV interface as intent-based rather than transaction-based. Instead of 'submit this swap transaction with this gas price', a user signs an intent like 'swap 5 ETH for at least 3,200 USDC by block 99,000,000'. Solvers see all open intents in a shared pool and compete to fill them at the best price. Because intents are private until matched and because solvers are competing across chains, MEV extraction is minimized: the solver captures only the legitimate execution value, not the information rents from front-running. SUAVE is in active development as of 2026.
- OFAC filtering and censorship debates
- The U.S. OFAC sanctions on Tornado Cash in 2022 forced MEV-Boost relays to choose between filtering sanctioned transactions or serving all valid ones. The market bifurcated: Flashbots and bloXroute implement filtering, while Ultra Sound (non-filtering mode) and Agnostic explicitly do not. This censorship debate is central to Ethereum governance, becausecensorship-resistant relays by definition cannot filter Tornado Cash or other sanctioned contracts - a legally and politically contested position that has attracted regulatory attention.
Why Flashbots matters
Flashbots is the most important infrastructure organization in Ethereum because it made MEV tractable as a problem. Before Flashbots, MEV was a dark forest where sophisticated actors extracted value from ordinary users with no recourse and no transparency. Flashbots' three contributions - the RPC for user protection, the MEV-Boost relay for validator revenue, and the research that quantified MEV - created a market structure where MEV can be measured, minimized, and redistributed. Understanding the Flashbots Relay, SUAVE, the relay market share numbers, and the difference between protected and unprotected transaction routing is essential infrastructure knowledge for any serious DeFi participant.