The centralization risk of MEV-Boost relays

Four entities control 95% of Ethereum blocks sent to validators. This concentration of control over the relay market creates a massive single point of failure for the network. In the last 14 days, MEV-Boost produced 91.5% of total Ethereum blocks. bloXroute Labs holds 45.6% of this share. I view this concentration as a direct threat to network stability. If a few relays fail or act maliciously, the proposer-builder separation system breaks.

Relay trust and technical vulnerabilities

Relays act as intermediaries. They facilitate auctions, ensure valid payloads, and manage payments between builders and proposers. Builders can gain latency advantages by colocating with relays rather than following the geographic distribution of proposers. This undermines geographic decentralization.

Relays still present trust issues. Builders must trust that relays accurately assemble and reveal the payload of the block. Proposers must trust that the relay receives the full signed block from the builder. Simon Brown notes that no trustless way exists for proposers to be confident that the builder’s block will be released.

The market fluctuates.

In April 2023, an attacker named "sandwich the ripper" exploited a bug in the Flashbots relay code to drain $20 million from MEV bots. The attacker disguised themselves as a validator. This exploit revealed that a rogue relay could control a massive percentage of Ethereum block space. MEV-Boost also has no way to verify that the value in a bid is actually what the relay claims. Malicious relays can submit fraudulent bids, which affects profit switching logic. Node operators also face incentives to hide MEV rewards earned in a block.

You should monitor how these failures impact validator rewards.

Censorship and emerging solutions

Censorship remains a persistent risk. While neutral relays now handle 70% of blocks, regulated relays still hold influence. bloXroute’s regulated relay held 2,771 blocks in the 24-hour window ending August 29, 2026, representing 23.8% of the 11,656 blocks measured. Proposers must trust that builders provide the correct payload and that relays do not steal the MEV, yet no trustless way exists for proposers to be confident that the builder’s block will be released or that they will receive payment.

The market shifts.

In late 2022, OFAC-compliant relays handled up to 80% of MEV-Boost traffic. That share fell to 2.4% by August 2026. This reversal happened because validators began picking multiple relays to maximize payouts. A neutral relay competing on payout wins more often than a filtering one because it can include transactions a compliant relay must exclude. In the August 29 window, Ultra Sound Relay built 4,094 blocks, while Titan Relay built 3,471 blocks.

Developers are working on EIP-8184 and EIP-8105 to address these issues. EIP-8184, known as LUCID, specifies a public encrypted mempool. EIP-8105 defines an interface that supports multiple decryption-key providers, such as threshold encryption or trusted execution environments. These proposals target predatory reordering while allowing legitimate backrunning and arbitrage once a transaction has already been included. Flashbots is also building SUAVE, a shared sequencing layer.

Shutter Network provides a roadmap for threshold encryption. They propose using a decryption key split across independent parties. Only a threshold-sized coalition can reconstruct the key. This prevents a single operator from decrypting transactions early.

It stays risky.

Does threshold encryption eliminate all relay needs?