An MEV Perspective on Glamsterdam

An MEV Perspective on Glamsterdam

This post reflects the opinions of Data Always and Hasu and seeks to provide an MEV-informed perspective on the headlining proposals for Glamsterdam. Our aim is to highlight tradeoffs that are missing from the broader conversation, particularly as they relate to block building, auction design, censorship resistance, and geodecentralization.

Our views were shaped by broad internal discussion and input from external collaborators. We’re especially grateful to Alejo, Bruno, Burak, Chris, Christoph, Danning, dmarz, Mateusz, Phil, Quintus, Robert, and Shea.

This post is released alongside an extended analysis of the Free Option Problem in ePBS.

Summary

In our view, BALs are uncontroversially beneficial for the execution layer.

For the consensus layer, however, there is no standout candidate, as every option has clear pros and cons:

  • Enshrined PBS (EIP-7732) potentially improves MEV auction dynamics, but introduces a free option problem that may significantly increase the number of empty blocks on Ethereum.

  • Reduced slot times (EIP-7782) would improve UX and performance for DeFi users, but the second-order effects on geodecentralization remain understudied.

  • FOCIL (EIP-7805) evolves local block building and helps to constrain block builders as the L1 scales. However, its censorship benefits do not extend to 99% of transactions known to be excluded by block builders today.

While we share the community’s ambition to ship faster and more boldly, it’s important to be explicit about the tradeoffs that come with increased pace.

EIP-7732: Enshrined Proposer—Builder Separation (ePBS)

The current slot timeline is inefficient and a major barrier to scaling the L1. Any restructuring of slots would be a major win for the network, so we find ourselves largely aligned with the intentions behind ePBS.

The community currently prefers the version of EIP-7732 that retains the payload timeliness committee, allowing builders and proposers to bypass relays. Disintermediating relays would have a large, but likely net-positive, impact on MEV market structure.

However, before discussing these benefits, we need to talk about a major risk that ePBS introduces: the Free Option Problem. By shifting data availability from relays to builders, the winning builder gains the ability to cancel a payload several seconds after winning by withholding blob commitments. This opens the door for searchers to bribe builders into reversing unfavorable trades.

Unfortunately, there appears to be no clear mitigation. While proposers could reject builders who cancel payloads, the economic incentives cut against that. The free option holds real value for searchers, and builders can be expected to share a portion of that value with proposers. This makes payload cancellation a rational strategy across the block supply chain.

As a result, ePBS may significantly increase the rate of empty blocks on Ethereum.

In addition, if any non-trivial subset of block builders fails to deliver their blob commitments, the execution payload deadline becomes unreliable to infrastructure. The execution payload would resemble a preconfirmation, forcing wallets and block explorers that need validity guarantees to wait until blob data is available; essentially, this would add six extra seconds before they can confirm transaction inclusion.


ePBS also introduces meaningful benefits, primarily by changing the MEV market structure from an open to sealed-bid auction.

Open auctions incentivize backroom deals and heavy investment in latency-sensitive infrastructure. Bid transparency lets builders monitor competitors and capture most of the value from exclusive order flow. This turns block building into a business development exercise, rewarding those who preferentially treat order flow sources. Additionally, the uncertain end time of current auctions leads to PGAs, where low latency is the dominant edge—encouraging builder-relay colocation.

Sealed-bid auctions might reduce the value of exclusive order flow and the payoff from latency spending. While this would disadvantage incumbents, including BuilderNet, we believe it would be a positive step toward fair markets and would likely raise equilibrium payments to proposers.

Absent relay disintermediation in Glamsterdam, it’s unclear how market structure will evolve. Years of indecision around ePBS have prevented the relay market from stabilizing. New proposals such as preconfirmation gateways and relay block merging aim to restructure the MEV pipeline, effectively blending the builder and relay roles. Further delay could push the ecosystem toward a pseudo-allowlisted builder market.

In light of both its risks and potential, ePBS remains a high-upside, high-uncertainty proposal.

EIP-7782: Reduced Slot Times

Reduced slot times would have the largest impact on end users among the proposed changes. The idea has rightly generated excitement across the community and MEV ecosystem.

However, Flashbots places strong emphasis on geographic decentralization and it is unclear how compressing the timeline of a slot would distort incentives across the network.

While the network seems able to maintain liveness under the proposed slot time, current analysis overlooks how tighter deadlines increase pressure on proposers and attesters to colocate—with each other and with MEV chokepoints. If rewards for attesters in high-latency regions like South America and Asia decline, validator distribution and thus geographic decentralization will suffer.

The effects on out-of-protocol actors are similarly poorly understood, making the impact on market structure harder to predict.

Local block builders are already operating near their limits. Their fork rates appear acceptable only because they build smaller blocks without access to private order flow. Moving attestation deadlines forward may restrict future throughput increases unless the community pairs reduced slot times with a reevaluation of local builders’ roles—potentially via FOCIL.

We encourage further research to understand how these incentive structures might evolve.

EIP-7805: Fork-Choice Inclusion Lists (FOCIL)

As the L1 scales, block builders will wield increasing influence over the chain. Several roadmap candidates (e.g., preconfirmations, shorter slot times, slot auctions, and APS) risk reinforcing this trend and geographically centralizing the builder set.

Today, local block building serves as a decentralized counterweight to block builders. Although more than 90% of blocks are outsourced to MEV-Boost, the remaining proposers promise eventual inclusion for transactions excluded by the MEV supply chain. However, as network throughput increases, the share of local builders is expected to decline toward zero. This is where FOCIL comes in.

By introducing direct constraints on the content of blocks, FOCIL empowers includers to ensure censorship resistance. This represents a re-envisioning of local block building. The network would be less reliant on altruistic proposers who forgo MEV rewards to retain control over transaction ordering in their blocks.

FOCIL’s main shortcoming is that it provides no censorship resistance to the classes of transactions that would benefit the most from it today.

The primary target of censorship today is blob-carrying transactions. Some builders appear to exclude blobs intentionally to reduce latency to relays, affecting ~50,000 transactions per month. For instance, Titan includes blobs in only ~70% of blocks they are available.

Gattaca’s Optimistic V3 Relay proposal offers one mitigation path to this incentive problem, letting builders submit headers instead of full blocks when bidding. However, it weakens relay data availability and introduces an ePBS-like free option problem.

The second class of transactions that would benefit from guaranteed fast inclusion are those that operate on contentious state. Unfortunately, these transactions have left the public mempool and also won’t benefit from FOCIL. These users will not return to the public mempool for additional censorship resistance, as they would be exposed to frontrunning and other forms of extractive MEV.

By contrast, only a few Tornado Cash transactions are deliberately excluded each day, mostly by long-tail builders, and OFAC-sanctioned transactions are exceedingly rare. Therefore, FOCIL’s near-term impact would be small.

Still, FOCIL is a meaningful first step toward broader future coverage and synergizes well with both faster and bigger blocks. However, this proposal must be seen as just a step. If FOCIL were selected as the headliner and led to a decrease in local block building, any gaps in its coverage, especially blobs, would face sharply diminished censorship resistance guarantees.

11 Likes