Proposer-builder separation - PBS - was formalised into Ethereum’s design as a way to keep validators honest. The idea: split the job of choosing transactions from the job of ordering them. Validators propose blocks; specialised builders construct them. Conflict of interest removed, MEV democratised, problem solved.
Except that’s not what happened.
What PBS Actually Produces
Block building is not a role that distributes naturally. It rewards scale, low latency, and access to private order flow. The firms that can co-locate infrastructure near major DEXes, maintain relationships with large wallet providers and CEXes, and run highly optimised simulation engines will consistently produce more profitable blocks than anyone else. Validators - which can number in the thousands and include solo stakers with modest hardware - have almost no ability to compete at that layer.
The result is a builder market that, by mid-2026, remains concentrated among a very small number of entities. Flashbots, beaverbuild, and Titan Builder have historically captured the majority of Ethereum blocks routed through MEV-Boost relays. The relay layer itself carries a similar concentration profile. None of this is secret; it’s visible on-chain.
The original argument for PBS was that if searchers extract MEV anyway, it’s better to make the process transparent and push the value toward validators rather than allowing a single dominant miner to capture it silently. That argument holds. But the downstream effect - a professional builder oligopoly controlling block construction - was somewhat anticipated and somewhat dismissed as a temporary market condition that competition would erode.
Competition hasn’t eroded it.
Why the Incentives Don’t Self-Correct
Builders win by having better order flow. Better order flow comes from being trusted by large volume sources - market makers, CEXes, wallets using proprietary transaction routing. Those relationships take time and capital to build, and once established, they compound. A new entrant offering marginally better execution has to overcome not just a technical gap but a trust and integration gap.
This creates a structural moat that market forces alone are unlikely to breach. Attesting validators, meanwhile, are economically rational: they submit to whichever relay offers the highest bid. They have little reason to care who built the block.
The Enshrined PBS Question
Ethereum’s roadmap includes enshrined PBS (ePBS), which would embed builder-proposer separation directly into the consensus layer rather than relying on the off-chain MEV-Boost infrastructure that dominates today. The theoretical benefit is reducing reliance on trusted relays and giving the protocol more control over the auction mechanism.
Whether ePBS materially changes builder concentration is less clear. The economic logic that rewards scale and private order flow doesn’t disappear because the auction moves on-chain. It’s possible ePBS improves censorship resistance and relay trust assumptions without doing much at all about who ends up building most blocks. That’s a meaningful improvement in some ways - and a fairly hollow one in others.
The centralisation question in block building is structural, not procedural. Ethereum’s architecture can constrain the worst outcomes, but it cannot mandate competitive markets into existence.