Vitalik Buterin Warns Ethereum Upgrade Could Shift Centralization Risk

Vitalik Buterin outlines Ethereum’s block building future, warning that upgrades like ePBS shift centralization risks rather than eliminate them.

Ethereum co-founder Vitalik Buterin has raised fresh concerns about the network’s block building pipeline.

In a detailed post on X, Buterin outlined several proposed upgrades coming to Ethereum. He stressed that while these changes improve decentralization in some areas, they shift the risk elsewhere.

The upgrades span from in-protocol changes to network-layer anonymization tools.

ePBS and FOCIL: Ethereum’s First Line of Defense

Buterin explained that Glamsterdam, Ethereum’s upcoming upgrade, will introduce enshrined Proposer-Builder Separation (ePBS).

This mechanism lets block proposers outsource work to a free and open market of block builders. The goal is to prevent block builder centralization from bleeding into staking centralization.

However, Buterin was clear that ePBS does not solve block builder centralization on its own. It simply separates the two concerns. That distinction matters a great deal for how Ethereum’s security model evolves over time.

To address this gap, Buterin pointed to FOCIL, which stands for Fork-Choice Enforced Inclusion Lists.

FOCIL picks 16 random attesters and gives each of them the power to select a few transactions. Those transactions must appear in the block or the block gets rejected. Even if one hostile actor controls all block building, FOCIL ensures that censored transactions still get through.

Buterin also floated the idea of “Big FOCIL,” a more ambitious version. In this design, the FOCIL set is large enough to include all block transactions. Each participant covers a slice of sender addresses to avoid overlap.

The result is a system where block builders may only need to handle MEV-heavy transactions like decentralized exchange arbitrage.

_Related Reading: _****Vitalik Buterin Reveals Ethereum’s Plan to Fight Censorship

Encrypted Mempools and the Transaction Ingress Layer

Buterin also addressed what he called “toxic MEV,” a term for exploitative behaviors like sandwiching and front-running. These tactics harm regular users by wrapping their transactions in hostile ones.

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

Encrypted mempools are one proposed fix. If a transaction stays encrypted until it is included in a block, no one can exploit it beforehand.

The technical challenge, Buterin noted, is ensuring that transactions can be validated and decrypted at the right time. The method used to decrypt them must be reliable, efficient, and not allow early exposure. This remains an open problem with active research ongoing.

Buterin also highlighted a layer that rarely gets attention in MEV discussions: the transaction ingress layer. This is the path a transaction travels from a user’s wallet to a block.

A hostile actor watching this path can still front-run or grief users, even without touching the blockchain itself.

To fix this, Buterin pointed to several emerging solutions. These include using Tor for routing, building a custom Ethereum-focused mixnet, and low-latency designs like Flashnet.

He noted that the kohaku initiative is expected to explore pluggable support for such tools. Both the network layer and on-chain privacy protocols are part of this growing design space.

Long-Term Vision: Distributed Block Building Like BitTorrent

Buterin shared a longer-term vision for Ethereum’s block building architecture.

He compared it to BitTorrent, a system that processes data without any single server holding everything. The dream is for Ethereum to handle far more transactions than any single machine needs to see at once.

The challenge, Buterin acknowledged, is that Ethereum’s core value is synchronous shared state. Any transaction can depend on any other. That dependency forces block building to remain centralized to some degree.

His proposed workaround involves creating new categories of transactions with limited global scope. He noted that over 95% of Ethereum activity does not truly need full global state access.

By creating cheaper transaction types for that majority, and leaving existing types in place but more expensive, the network could enable more distributed building over time.

Buterin framed this as an open and exciting long-term design space.

No firm timeline or specification exists yet. But the direction points toward a more modular, distributed Ethereum that reduces reliance on any single actor in the block production chain.