Section IV: The Centralization Crisis

Despite Flashbots' innovations, the MEV ecosystem faces a fundamental challenge: concentration of power among a small number of operators.

The extent of this concentration becomes clear when examining recent data. In October 2024, just two builders produced 90% of blocks over a two-week period. From October 2023 through March 2024, three builders controlled approximately 80% of MEV-Boost blocks. During this same timeframe, a significant share of blocks, often around 60%, were relayed via OFAC-compliant infrastructure (adhering to U.S. Office of Foreign Assets Control sanctions). The pattern is unmistakable: these high barriers to entry have consolidated power among a handful of operators, directly undermining blockchain's decentralized principles.

The relay layer introduces additional centralization concerns. Because only a handful of trusted relays dominate the market, their compliance decisions (such as filtering transactions to adhere to OFAC sanctions) can have network-wide effects. These supposedly neutral intermediaries become powerful chokepoints that shape which transactions actually make it into blocks regardless of individual validator preferences. The choice of which relays to trust can determine transaction inclusion, making censorship resistance vulnerable to a small set of gatekeepers.

Responses to Centralization

The concentration revealed by these metrics made clear that MEV-Boost alone couldn't solve the centralization problem. The relay layer remained a chokepoint, and builder concentration continued unabated. The industry needed more fundamental restructuring.

In November 2024, major players launched BuilderNet, a decentralized block-building network jointly operated by Flashbots, Beaverbuild, and Nethermind. BuilderNet uses specialized hardware that creates secure enclaves where code runs in isolation, even from the machine's owner. This technology enables multiple operators to share transaction order flow and coordinate block building while keeping contents private until finalization, since no single party can see or manipulate the data being processed.

The goal is to create a more transparent and permissionless system for MEV distribution, replacing the opaque, custom deals that currently define the market. Beaverbuild has already begun transitioning its centralized builder to this network, with additional permissionless features planned for future releases.

Beyond BuilderNet, the ecosystem has developed several complementary approaches to combat centralization and return value to users.

One approach focuses on returning value directly to users. Order flow auctions let users auction off their transaction flow to the highest bidder, potentially earning rebates from the MEV their trades create. The private routing solutions discussed earlier represent one implementation, while encrypted mempools hide transaction details until execution.

At the protocol level, researchers are exploring ways to distribute MEV rewards more evenly across validators rather than letting them concentrate among the fastest operators. Enshrined PBS would build the proposer-builder separation directly into Ethereum's protocol rather than relying on external infrastructure like MEV-Boost.

More advanced attacks also require attention. Time-bandit attacks, where validators attempt to reorganize recent blocks to capture MEV, are constrained by stronger finality guarantees under proof-of-stake, though related attack vectors remain an active research concern.

While these solutions show promise, results in practice remain mixed, and the arms race between MEV extraction and protection continues to evolve.