TL;DR:
- Vitalik Buterin says Glamsterdam brings ePBS so builder centralization does not leak into validator centralization via permissionless builders.
- FOCIL would have 16 random attesters require transaction inclusion, with Big FOCIL narrowing builders to MEV ordering and execution.
- Encrypted mempools, Tor and mixnets target toxic MEV and ingress surveillance, while a BitTorrent-style scaling vision clashes with synchronous shared state and cites Flashnet as a privacy option to further reduce manipulation.
Ethereum’s next upgrade cycle is being framed as a governance and market-structure reset. In a detailed post, Vitalik Buterin said the ‘Glamsterdam’ upgrade would introduce enshrined Proposer-Builder Separation, or ePBS, to stop block builder centralization from spilling into validator centralization. The idea keeps proposers free to outsource block construction to a permissionless builder market, while preserving staking decentralization. But Buterin argued that builder concentration still needs explicit countermeasures to prevent hidden control from accumulating over time. He described the effort as changes to block-building control.
Finally, the block building pipeline.
In Glamsterdam, Ethereum is getting ePBS, which lets proposers outsource to a free permissionless market of block builders.
This ensures that block builder centralization does not creep into staking centralization, but it leaves the…
— vitalik.eth (@VitalikButerin) March 2, 2026
How ePBS could reshape Ethereum block building
To address concentration, Buterin highlighted FOCIL, a censorship backstop embedded in the protocol. Under this design, 16 randomly selected attesters each nominate transactions that must be included, and a block can be rejected if those transactions are missing. That forces inclusion power to be partially distributed even if one builder dominates ordering. He also pointed to a ‘Big FOCIL’ direction that could scale the model further, shrinking builders toward MEV-relevant ordering and state execution while limiting duplicate inclusions. Duplication is mitigated by partitioning senders and retrying transactions next slot.
Buterin also flagged ‘toxic MEV’ as a practical risk surface, including sandwich attacks and frontrunning. The team is exploring encrypted mempools to blunt pre-inclusion exploitation, keeping transactions hidden until they are included so attackers cannot react early. He noted the engineering tradeoff: transactions must remain valid and reliably decrypt at the right moment. Beyond the mempool, he stressed the network-layer ingress gap, where users are exposed between submission and inclusion. Ideas cited include Tor routing, Ethereum-specific mixnets, and latency-optimized privacy networks like Flashnet, as defenses mature.
Zooming out, Buterin described a BitTorrent-like scaling vision that confronts synchronous shared state. He floated transaction categories that would not require full global synchronization, allowing Ethereum to process far more activity than any single server needs to coalesce locally. The constraint is Ethereum’s core value proposition: synchronous shared state means any transaction could depend on any other, which centralizes ordering and block building. The direction is to reduce hidden centralization, curb MEV abuse, and strengthen censorship resistance. The post follows the Foundation’s ‘strawmap’ roadmap discussion stretching toward decade end.
