Proposer-builder separation - PBS - started as a response to a specific, uncomfortable fact: validators on Ethereum were increasingly running MEV-boost software to outsource block construction to specialist builders, and the network had no formal way to handle that relationship. The solution was to bake the separation in at the protocol level. What’s emerged instead is something messier and more politically charged than the original problem.
The Actual Mechanic
Under Ethereum’s current setup, validators (proposers) choose which block to include from a marketplace of blocks built by specialist builders. MEV-boost acts as the middleware. Builders compete to offer the most profitable block, proposers take the highest bid, and the profit gets split via relay infrastructure. The Ethereum protocol itself doesn’t natively enforce this - it’s still running on a voluntary, off-protocol system.
Enshrined PBS (ePBS), which would move this into the consensus layer, has been under active research for several years. The core design being discussed involves an explicit protocol-level auction where builders commit to a block and proposers commit to publishing it. It removes the relay as a trusted intermediary.
Why the Relay Problem Matters
Relays are currently the weakest architectural point in the system. They’re trusted to not reveal the block contents to the proposer before payment, and trusted to not lie about block validity. There’s no cryptographic enforcement of either. A small number of relays - operated by entities like Flashbots and BloXroute - handle the majority of MEV-boost traffic, which means Ethereum block production runs through a de facto oligopoly that exists entirely outside the protocol.
This isn’t a hypothetical risk. Relay censorship during the OFAC sanctions period in 2022 demonstrated that the off-protocol relay layer could filter transactions based on external legal pressure.
The Concentration Problem
What’s less discussed is what happens if ePBS succeeds. Specialist builders would still exist - the economics that created them don’t disappear when the protocol absorbs the auction mechanism. Sophisticated block construction, especially anything touching cross-domain MEV or private order flow, requires infrastructure and relationships that small operators can’t replicate. Builder concentration may actually increase once the relay buffer is removed, because the competitive pressure shifts directly onto who can build the most profitable block, not who can win a relay relationship.
Whether that’s a worse outcome than the current relay dependency is genuinely unclear. The architectural problem gets solved; the economic one might just relocate.