Ethereum Guild 2025-04 (June 3, 2025): Local block building vs. builder blocks, min-bid, protocol alignment

[philknows]

When: June 3 @ 1300 UTC
Meeting Link: https://meet.google.com/akk-zzid-hws

Topics for Discussion

Research about local block building vs. builder blocks

• Reading: https://ethresear.ch/t/on-solo-staking-local-block-building-and-blobs/20540
  • Reorg rates/share
  • Block Propagation
• Reading: https://ethresear.ch/t/blobs-reorgs-and-the-role-of-mev-boost/19783
  • Blob propagation/inclusions of local vs. builders
  • Censorship
• Reading: https://ethresear.ch/t/big-blocks-blobs-and-reorgs/19674
  • Lodestar's increased reorg rate

MEV-Boost & Mempool

Reading: https://ethresear.ch/t/is-it-worth-using-mev-boost/19753
Reading: https://ethresear.ch/t/expanding-mempool-perspectives/22022

• Amount of MEV buildup post-local block
• Tx inclusion for network participants
• Artificial network congestion
• Censorship resistance
• Builder diversity
• min-bid effectiveness

Node Operators

• Is min-bid on Lido node operators harming the network?

[philknows]

Ethereum Guild 2025-04: Local Block Building vs. Builder Blocks Discussion Summary

Recording: https://www.youtube.com/watch?v=rhWVLJ9EvDg

This summary details the Ethereum Guild discussion from June 3, 2025, featuring Toni Wahrstaetter from the Ethereum Foundation and the Lodestar team, focusing on block building infrastructure, MEV-boost dynamics, and the future of Ethereum's block production pipeline.

The Current State of Local Block Building vs. Builder Blocks

The conversation began with Philip Ngo asking about ChainSafe's Lido contribution and whether it helps or harms the network. Toni acknowledged that local block building aligns with Ethereum's values by avoiding censorship and centralization risks when validators build their own blocks.

However, Toni highlighted significant challenges facing local block builders:

• Local block builders miss out on most ETH rewards, though this is "still very negligible" as most rewards come from the Consensus Layer rather than execution layer
• Higher reorg rates and missed head votes make locally built blocks less stable
• Local block building is "slowly fading away" due to PeerDAS and proposals like inclusion lists, and gas limit scaling making it increasingly difficult for non-sophisticated infrastructure to push blocks out fast enough
• Local builders typically use only 5-10 million gas compared to builder blocks that use 15-16 million per block with MEV-Boost
  • This utilization difference reflects both the lack of access to private order flow and the technical limitations of local building infrastructure in optimizing block space usage.

Toni noted: "There are people that are building blocks locally, but what we definitely see is higher reorg rates for locally built blocks. More missed head votes. So your blocks are in general kind of less stable

Local block building represents the original Ethereum vision where individual validators construct their own blocks using transactions from the public mempool, maintaining direct control over transaction inclusion and ordering. This approach aligns with Ethereum's decentralization principles but faces mounting practical challenges.

Min-Bid Settings and Relay Timing Games

A significant technical discussion centered on min-bid settings and timing challenges:

• Relays engage in "timing games" by delaying responses to validators using min-bid flags
• When relays respond with values below min-bid setting after delays of 1-2 seconds, validators lose precious time to start building local blocks
• This interaction between relays and validators using min-bid significantly increases reorg probability

Nico Flaig noted that ChainSafe and other top performers like Attestant and Sigma Prime produce around 90% local blocks due to high min-bid settings, prompting discussion about whether this should be reconsidered given market changes.

Censorship Resistance and Liveness Concerns

Matthew shifted the conversation to Ethereum's ethos, expressing concern that researchers seem resigned to external builders being inevitable. The discussion emphasized that MEV-Boost builders and relays represent "sophisticated parties" with "well connected and specialized" infrastructure for maintaining "low-latency connections to peers." This infrastructure advantage extends beyond simple network connectivity to include access to multiple exchanges, private order flow sources, and optimized execution environments that individual validators cannot realistically replicate.

Toni identified two primary risks of abandoning local block building:

1. Censorship resistance: External builders could potentially censor transactions
2. Liveness: If MEV-boost experiences bugs, it could threaten network liveness since it's not part of the core protocol

Toni cited past incidents where:

• Relays gave bad headers/blocks
• Relay circuit breakers had bugs requiring manual intervention
• Builders censored certain transactions during regulatory events

The Economics of Block Building

The discussion revealed complex economic dynamics in the block building ecosystem:

• Private order flow ("dark pools") gives external builders significant advantages, as transactions are submitted directly to builders bypassing the public mempool
• Toni called competing with builders for profitability "a lost fight" since arbitrage opportunities require "two-digit million dollars on multiple big exchanges"
• Relays use "bid adjustment" strategies that reduce validator rewards by adjusting bids downward when they have timing advantages
  • Some relays get bids earlier than others due to lower latencies from co-location and relays will look at the delta between the two top bids. Using the builder's key, it can sign the readjusted bid to only slightly outbid the second highest bid and arbitrage the delta for the relay and builder, instead of passing it onto validators
  • Example: Top bid is 100 ETH and second top bid is 1 ETH. By adjusting the bid downward to 1.1 ETH, the block wins and the remaining 98.9 ETH is split between the builder and the relay.
  • This is potentially also why we are seeing validators get less rewards.
• External builders typically fill blocks to capacity (15-16M gas) while local blocks use only 5-10M gas

Interestingly, during the OFAC sanctions period, MEV-Boost builders actually included "more tornado cash transactions than local builders" on a per-block basis, challenging assumptions about which building method better serves censorship resistance. Toni explained: "What you saw is almost like tornado cash transactions stacking up and at some point their priority fees helped some censoring builder to win the auction."

Centralization Concerns

The conversation highlighted several centralization vectors:

• Two builders (Beaverbuild and Titan) dominate block production, though Toni believes they are controlled by separate entities based on their different responses to OFAC sanctions
• Relays operate "fast lanes" that prioritize known builders, disadvantaging smaller builders
• Matthew expressed skepticism about trusting any of these entities (whether they are even controlled by different stakeholders) given the financial incentives: "There's too much money to be made and too many greedy people in the world."

The conversation explored whether maintaining censorship resistance through local building justifies the reduction in network efficiency. As Faith suggested, the focus might need to shift toward "making locally produced blocks more competitive" rather than relying solely on punitive measures for external builders. However, it is very difficult, if even possible, to make local blocks more economically valuable.

Circuit Breakers and Liveness Concerns

Current circuit breaker implementations in consensus clients received attention as both a solution and a source of additional complexity. These mechanisms are designed to fall back to local building when MEV-Boost experiences failures, but their effectiveness remains questionable. The discussion noted instances where circuit breakers themselves contained bugs, failing to properly respond to relay failures and requiring manual intervention.

The conversation highlighted the fundamental difference between core protocol failures and MEV-Boost failures. When core protocol software has bugs, the community accepts this as "a core protocol bug that destroys not only the chain" but maintains consistency in treatment. However, MEV-Boost failures are treated differently despite having significant network impact, creating an asymmetry in how different types of failures are addressed.

Future Directions and Potential Solutions

Several potential paths forward were discussed:

• Enshrining MEV-boost: Integrating it into the core protocol could address liveness concerns, though Toni expressed reservations about the complexity
• Encrypted mempools: While promising, Toni noted these would still have MEV issues through transaction spamming which leads to probabilistic MEV strategies rather than eliminating/minimizing MEV.
• Programmable encrypted mempool: This concept could provide algorithmic ordering guarantees. See "Building with TEEs."
• Application-layer solutions: Matthew suggested moving MEV handling to application layers like UniChain does. Toni thinks that this would not be as effective as we want it to be (see further below).
• Building with TEEs: While potentially useful, Toni expressed concerns about trusting hardware security modules that "break every few years" and needing to trust the hardware vendors.

Alternative approaches like COW Swap and application-layer MEV mitigation received attention as potential models for moving MEV extraction closer to users. However, these solutions face adoption challenges and may not achieve the efficiency levels of current builder infrastructure.

The conversation highlighted the fundamental challenge of maintaining competitive block building while ensuring adequate fallback mechanisms. As MEV opportunities decline and building costs increase, maintaining multiple competitive builders becomes increasingly difficult, potentially leading to market concentration that undermines the benefits of outsourced block building.

Enshrined Proposer-Builder Separation (EPBS)

EPBS emerged as a potential solution to many current infrastructure challenges, though the discussion revealed significant reservations about its complexity and implementation requirements. Toni expressed concerns about EPBS due to "fork choice complexity," describing it as "way too complex" due to the integration of external inputs into the fork choice function through payload timeliness committees.

Despite complexity concerns, EPBS offers several advantages including better propagation, improved pipelining, and the enshrining of relay functionality within the core protocol. This would address current liveness concerns by making relay functionality part of the tested, validated core protocol rather than external software that could fail without triggering standard protocol bug responses.

The conversation explored how EPBS might address current centralization concerns while introducing new complexities. By enshrining relay functionality, EPBS would ensure that local builders could "step up and say, 'Okay, I submit a block to myself'" when external builders experience failures, providing a more robust fallback mechanism.

MEV Extraction and Value Flow

The conversation revealed that most MEV value derives from sophisticated arbitrage operations requiring substantial capital and exchange access rather than simple transaction reordering. Builders need "two-digit million dollars on multiple big exchanges" to compete effectively in arbitrage opportunities, creating natural barriers to entry that favor large, well-capitalized operations.

Current MEV flows primarily benefit validators through builder payments, but the discussion acknowledged this represents an extraction from users rather than value creation. As Wahrstaetter noted, the European Central Bank has described the current system as various entities "basically rugging users," highlighting the need for mechanisms that return value to transaction originators.

The conversation explored potential solutions like payment for order flow models where builders pay users for transaction access, then compete to extract MEV while compensating both users and validators. This would create a more sustainable economic model that "goes back to the user" rather than concentrating value extraction among infrastructure providers.

Conclusions on Min-Bid and Local Building

The conversation concluded with practical considerations for validator operations:

• With censorship less prevalent now, the ChainSafe team questioned their high min-bid settings that result in 90+% local blocks
• Phil noted their priority is "making sure that we're not hindering the network in any way"
• Toni suggested that absent censorship concerns, there's little reason to prefer local building except for idealistic reasons: "Local builders really being like, hey, this is now my time to contribute some data to this thing, what we call the blockchain."

This guild meeting provided valuable insights into the complex trade-offs between decentralization, efficiency, economics, and Ethereum's core values in the evolving block building landscape. ChainSafe has removed min-bid from its validators and have started connecting to additional non-censoring relayers, even if regulated/censored relays pose little to no difference in censorship.