Proposer-Builder Separation (PBS)

What Is Proposer-Builder Separation?

Proposer-Builder Separation (PBS) is a blockchain architecture that divides block production into two distinct roles: builders who assemble transactions into candidate blocks, and proposers who select and broadcast one of those blocks to the network. The proposer never sees the block's contents before committing to it: they simply pick the bid that pays the most.

The concept emerged from research into Maximal Extractable Value (MEV): the profit that block producers can capture by reordering, inserting, or censoring transactions. Before PBS, Ethereum validators both built and proposed blocks. Sophisticated operators could run MEV extraction strategies ( frontrunning, sandwich attacks, arbitrage) that gave them a structural advantage over home stakers. PBS decouples the two functions so that MEV revenue flows to any validator who wins a proposer slot, not just those running extraction software.

Flashbots, formed in 2020 to address MEV externalities, pioneered the first practical PBS implementation. Their MEV-Geth system for proof-of-work Ethereum was adopted by mining pools controlling over 84% of hashrate by April 2021. After Ethereum's transition to proof-of-stake in September 2022, MEV-Boost became the standard PBS middleware.

How It Works

PBS operates through a blind auction that runs every 12-second Ethereum slot. The mechanism ensures proposers cannot steal builders' MEV strategies while builders cannot withhold blocks after winning.

The MEV-Boost Auction Flow

1. Searchers identify MEV opportunities (arbitrage, liquidations) and submit transaction bundles to builders
2. Builders aggregate transactions from the public mempool and private order flows, then simulate thousands of orderings to construct the most profitable block possible
3. Builders submit sealed bids (block header plus bid amount) to one or more relays
4. Relays verify block validity and that the payment to the proposer's fee recipient matches the bid
5. Relays forward only the block header and bid to MEV-Boost: the proposer sees the payment amount but not the transactions
6. MEV-Boost aggregates bids from all configured relays and presents the highest to the proposer
7. The proposer signs the winning block header, irrevocably committing to that block
8. The relay reveals the full block body to the network for attestation

This commit-reveal scheme is the core of PBS security. The proposer cannot see the transactions (preventing MEV theft), and once they sign the header, they cannot substitute a different block.

Adoption and Economics

Over 90% of Ethereum validators run MEV-Boost. The additional MEV revenue lifts validator APY by roughly 1 to 2 percentage points. Since September 2022, cumulative MEV extracted on Ethereum has exceeded $1 billion, a figure that would otherwise accrue disproportionately to sophisticated operators rather than the broader validator set.

The Builder Market

A small number of builders dominate Ethereum block production. As of early 2026, Titan Builder produces approximately 52% of blocks, BuilderNet (a collaborative network launched by Flashbots, Beaverbuild, and Nethermind) accounts for roughly 25%, and Quasar handles about 15%. The top three builders collectively produce over 90% of blocks.

This concentration results from a flywheel effect: dominant builders attract exclusive order flow (private transactions routed directly to them), which produces more profitable blocks, which wins more auctions, which attracts more order flow. The Herfindahl-Hirschman Index (HHI) for the builder market sits around 3,892, well above the 1,800 threshold that regulators consider "highly concentrated."

Enshrined PBS

Current PBS depends on relays as trusted intermediaries between builders and proposers. If relays collude, censor, or go offline, the auction breaks down. Enshrined PBS (ePBS) moves the auction into the Ethereum consensus layer itself, eliminating external trust assumptions.

EIP-7732

EIP-7732 is the core proposal for enshrined PBS. It introduces several protocol changes:

• Proposers no longer include execution payloads in beacon blocks. Instead, they commit to a signed bid from a builder specifying the execution block hash and payment value.
• Builders become staked entities on the beacon chain, registered with a minimum deposit of 1 ETH. Their stake guarantees payment even if the proposer misbehaves.
• A Payload Timeliness Committee (PTC) of 512 validators per slot attests whether the builder revealed their payload on time and blob data remained available.
• The state transition splits into three phases: beacon block processing, execution payload envelope processing, and payment settlement.

// Simplified ePBS beacon block structure (EIP-7732)
BeaconBlockBody {
  // Consensus fields remain
  attestations: List[Attestation]
  sync_aggregate: SyncAggregate

  // Execution payload removed; replaced by builder bid
  signed_execution_payload_bid: SignedExecutionPayloadBid {
    builder_index: ValidatorIndex
    block_hash: Hash32
    value: Gwei       // Payment to proposer
  }

  // PTC attestations added
  payload_attestations: List[PayloadAttestation]
}

EIP-7732 is the consensus-layer headliner for the Glamsterdam hard fork. As of April 2026, ePBS devnets are being stabilized ahead of the generalized Glamsterdam testnet.

Inclusion Lists (FOCIL)

A companion proposal, EIP-7805 (Fork-Choice Enforced Inclusion Lists, or FOCIL), addresses censorship concerns. Validator committees generate mandatory inclusion lists that builders must incorporate, preventing indefinite transaction censorship. Together with ePBS and future encrypted mempools, FOCIL forms what researchers call the "holy trinity" of censorship resistance.

Bitcoin and Block Template Construction

Bitcoin faces an analogous problem: traditionally, mining pool operators control which transactions go into blocks. Individual miners contribute hashrate but have no say in transaction selection. This concentrates censorship power in a handful of pool operators.

Unlike Ethereum, Bitcoin does not need PBS in the same way because its proof-of-work consensus and simpler scripting language create far less MEV. There is no equivalent of the sophisticated frontrunning and sandwich attacks that plague Ethereum's DeFi ecosystem. The primary concern is censorship resistance, not MEV distribution.

Stratum V2

The Stratum V2 protocol addresses Bitcoin's block template concern by letting individual miners construct their own block templates. Miners run a local Bitcoin node, select transactions from their own mempool, and negotiate the template with the pool. The pool validates basic requirements but cannot dictate transaction selection.

Stratum V2 also adds encrypted communication (preventing ISP-level snooping on miners) and a binary protocol that reduces bandwidth usage compared to V1's JSON format.

As of May 2026, seven major mining pools representing roughly 75% of global Bitcoin hashrate have committed to Stratum V2 adoption: Foundry, AntPool, F2Pool, SpiderPool, MARA Pool, Block Inc, and DMND. This represents a significant shift in mining decentralization. For more on the economics driving these changes, see the Bitcoin mining economics research.

Key Differences from Ethereum PBS

Why It Matters

PBS shapes how value flows through a blockchain. Without it, block producers who run sophisticated MEV strategies earn more than those who do not, creating a centralizing pressure where only well-resourced operators remain competitive. Over time, this undermines the decentralization that makes blockchains valuable.

For users, PBS matters because MEV extraction often comes at their expense: sandwich attacks inflate swap costs, frontrunning steals arbitrage from ordinary traders. Solutions like Flashbots Protect route user transactions through private channels that bypass the public mempool, reducing exposure to these attacks.

For Bitcoin-based systems like Spark, PBS is less directly relevant because Bitcoin's UTXO model and simpler scripting create minimal MEV surface. However, the broader principle of decentralizing block construction applies: Stratum V2's miner-driven template selection strengthens the censorship resistance that Bitcoin Layer 2 protocols depend on for their security guarantees.

Risks and Considerations

Builder Centralization

PBS was designed to decentralize MEV revenue, but it has concentrated block construction instead. A single builder producing over half of all blocks creates systemic risk: if that builder goes offline, experiences a bug, or decides to censor transactions, the impact is network-wide. The exclusive order flow flywheel makes it difficult for new builders to compete.

Relay Trust Assumptions

Current MEV-Boost PBS relies on relays to honestly mediate between builders and proposers. Relays can censor blocks, misreport bids, or leak block contents. While multiple relays exist (Ultra Sound, Titan, bloXroute, Aestus, Flashbots, Agnostic), the relay layer remains a point of centralization and trust. Enshrined PBS (EIP-7732) is specifically designed to eliminate this dependency.

Censorship Resistance

Builders may choose to exclude certain transactions for regulatory or commercial reasons. If dominant builders censor transactions, users have limited recourse under the current system. FOCIL inclusion lists (EIP-7805) aim to address this by requiring builders to include committee-selected transactions, but this proposal is still in development.

Latency and Timing Games

Builders and relays in geographic proximity to validators gain an advantage: submitting bids milliseconds later (with more recent MEV data) can win auctions. This "timing game" incentivizes infrastructure colocation and creates another axis of centralization that protocol design alone cannot fully address.

This glossary entry is for informational purposes only and does not constitute financial or investment advice. Always do your own research before using any protocol or technology.