Crypto Gloss

Proposer-Builder Separation

Proposer-Builder Separation (PBS) is an Ethereum architecture that divides block creation into two distinct roles: the block proposer (a validator that selects a block to propose, attests to it, and earns the bid payment) and the block builder (a specialized agent that constructs the full block content — selecting, ordering, and bundling transactions from searchers and the public mempool to maximize MEV and priority fees). Without PBS, validators must simultaneously perform both roles: they must run sophisticated MEV software, compete in searcher markets directly, and handle complex bundle ordering — resulting in centralization pressure toward large, technically sophisticated solo staking or validator pool operations. PBS solves this by creating a market: builders compete to offer validators the highest-paying block; validators simply select the best offer (highest bid) without needing to understand MEV strategies. The current PBS implementation (MEV-Boost, by Flashbots) works out-of-protocol — it relies on trusted relays to intermediate between builders and validators. A planned future upgrade, ePBS (Enshrined PBS), would implement separation natively at the Ethereum protocol level, removing relay trust and improving censorship resistance.

Key Facts

  • Concept origin: Propposed by Ethereum researchers (Vitalik Buterin, Justin Drake, et al.) as solution to MEV centralization pressure
  • Current implementation: MEV-Boost (out-of-protocol; relies on trusted relays)
  • MEV-Boost adoption: ~95% of Ethereum validators run MEV-Boost
  • Builder concentration: Top 3 builders: ~80-90% of winning blocks (2024)
  • Relay concentration: Flashbots relay: ~50-70% of all blocks (dominant)
  • ePBS: Planned Ethereum core protocol upgrade; timeline: post-Pectra (2025+)
  • Block proposer: Always a validator (selected by beacon chain RANDAO)
  • Block builder: Specialized off-chain agents (Beaver Build, Titan Builder, rsync-builder, etc.)

The Problem PBS Solves

The following sections cover this in detail.

Without PBS: MEV Centralizes Validators

In a world without PBS, validators both propose and build their blocks. To maximize revenue:

  1. Validators must run MEV searcher software (monitor mempool; execute strategies)
  2. Validators need large amounts of capital (for JIT liquidity, arb strategies)
  3. Sophisticated validator operations outcompete naive solo stakers by 10-100×

Result: Solo stakers earn far less than sophisticated validator pools → stakers migrate to large pools (Lido, Coinbase, etc.) → validator set centralizes → Ethereum: lower decentralization and censorship resistance.

With PBS: Decoupled Market

Builders have MEV expertise. Validators don’t need it. The market:

  1. Builders construct optimal blocks (aggregate searcher bundles + public txs)
  2. Builders bid for the right to have their block proposed
  3. Validators select: highest bid wins → validator earns MEV revenue without MEV expertise
  4. Solo stakers: earn same competitive MEV rate as Coinbase validators (both: just accept builder bids)

MEV-Boost: PBS Out-of-Protocol

MEV-Boost (Flashbots, 2022) is the current PBS implementation — software that validators run alongside their beacon client.

Architecture

“`

Searchers → Builders → Relays → Proposers (Validators)

“`

Participants:

Role Function Examples
Searcher Find MEV; construct bundles Private bot operators
Builder Assemble full blocks; bid for inclusion Beaver Build, Titan, rsync
Relay Trusted intermediary; verify blocks; pass bids Flashbots, Ultra Sound, BloXroute
Proposer Accept highest bid; propose block All validators running MEV-Boost

Relay Trust Model

The relay is the key trust point in current PBS:

  1. Builder: submits full block to relay (with bid amount)
  2. Relay: verifies block validity + bid payment; sends only block header + bid to proposer
  3. Proposer: selects highest bid; signs header (commit to proposing this block)
  4. Relay: only then reveals full block body to proposer for broadcasting

Why this is trustworthy: The relay holds both the full block and the proposer’s commitment; it cannot cheat without destroying its reputation as relay. Validators trust relays to: (1) not reveal blocks before proposer commits, and (2) only forward valid blocks with verified bids.

Relay censorship risk: If Relay decides to censor specific transactions (e.g., Tornado Cash addresses under OFAC compliance), builders using that relay cannot include those transactions. During 2022-2023, Flashbots relay: briefly censored Tornado Cash-related txs; significant controversy → validators switched to non-censoring relays.

Block Builder Economics

The following sections cover this in detail.

Builder Revenue

  • Builder profit: total block MEV + priority fees – validator bid payment
  • Example: Block with $50,000 MEV → builder bids $45,000 to validator → builder profit: $5,000

Builder Competition

  • Validators: via MEV-Boost: simultaneously receive bids from multiple relays
  • Validators: select block with highest bid
  • Result: competitive market → most MEV: passed through to validators

Builder Centralization Concern

Reasons for concentration:

  1. Capital: large builders have capital for JIT liquidity, atomic arb
  2. Searcher relationships: exclusive order flow agreements (better bundles → better blocks)
  3. Latency: faster builders: submit bids later (wait for more searcher bundles); higher bids; more wins
  4. Private order flow: payment for exclusive access to certain tx flows (wallets, apps)

ePBS: Enshrined Proposer-Builder Separation

ePBS is the planned Ethereum upgrade that moves PBS from out-of-protocol (MEV-Boost: relies on Flashbots software + trusted relays) to in-protocol (Ethereum consensus: enforces builder-proposer separation natively).

ePBS Design Goals

  • Improve censorship resistance: protocol-level includes FOCIL (Fork-Choice Enforced Inclusion Lists): ensures basic transaction inclusion despite builder censorship
  • Same market structure: builders and searchers: continue to exist; their role: unchanged; the trust model: improved

ePBS Timeline

  • post-Pectra (2025+): earliest realistic ePBS inclusion
  • Current: MEV-Boost remains dominant; ePBS: still being designed

SUAVE: Decentralized Mempool

SUAVE (Single Unified Auction for Value Expression) is Flashbots’ experimental architecture for a decentralized, chain-agnostic block building and mempool system — essentially: a blockchain dedicated to MEV ordering that any chain (Ethereum, Optimism, etc.) can use. SUAVE: eliminates the need for centralized relays; replaces: trusted relay with: decentralized TEE (Trusted Execution Environment)-based executor network.

Related Terms

Sources

  1. “Proposer-Builder Separation Friendly Fee Market Designs” — Vitalik Buterin / Ethereum Research (2021). The foundational Ethereum research post introducing PBS as an architectural solution to MEV centralization — arguing that in a world where MEV is large and unavoidable, separating the proposer role (who selects blocks) from the builder role (who constructs them) allows validators to remain decentralized while builders specialize in MEV extraction, preventing MEV from becoming a centralizing force for validator operations.
  1. “MEV-Boost: Flashbots Architecture and Relay Trust” — Flashbots / MEV-Boost Documentation (2022). Technical documentation and analysis of MEV-Boost’s architecture — describing the relay trust model, the builder submission process, the proposer selection mechanism, and the security assumptions in the out-of-protocol PBS implementation.
  1. “Builder Centralization: Measuring Market Concentration in Ethereum Block Building” — EigenPhi / Blockworks Research (2024). Quantitative analysis of builder market concentration on Ethereum — measuring Herfindahl-Hirschman Index (HHI) of the builder market, tracking which builders win the most blocks, examining whether private order flow relationships drive concentration, and assessing whether builder centralization creates meaningful censorship risk.
  1. “ePBS: Removing Relay Trust from Ethereum Block Production” — Ethereum Foundation / Protocol Research (2024). Technical specification and motivation for ePBS (Enshrined Proposer-Builder Separation) — the planned Ethereum upgrade that implements PBS at the consensus layer, eliminating the need for trusted relays, and examining how ePBS interacts with FOCIL (Forward Choice Inclusion List) for censorship resistance.
  1. “Private Order Flow: How Builder-Wallet Deals Are Changing MEV” — Paradigm / MEV Analysis (2024). Analysis of how private order flow agreements — where wallets, aggregators, and dapps sell their users’ transaction flow exclusively to certain builders — are reshaping MEV distribution, builder concentration, and the privacy cost paid by users for “MEV protection.”