The Cost of Builder Dominance: The Rise of Censorship-Resistance Risks

Dominant builders like Flashbots and bloXroute can and do censor transactions from sanctioned addresses. This creates a two-tiered system where compliance is enforced at the protocol level, not just the interface.

• >50% of Ethereum blocks are built by OFAC-compliant entities.
• Tornado Cash sanctions demonstrated this is not theoretical.
• Risk extends to any protocol or user interacting with a blacklisted address.

Builders can suppress or reorder transactions for non-regulatory reasons, including competitive advantage or political motives. This attacks the core property of credible neutrality.

• A builder could front-run or censor a competing protocol's governance vote.
• No slashing mechanism exists for this behavior.
• Creates an information asymmetry where users cannot verify if their tx was suppressed.

PBS, a core design of Ethereum's roadmap, formally separates block building from block proposing. This limits a single entity's power.

• Proposers (validators) choose the most profitable header, creating a competitive market for uncensored blocks.
• Builders compete on inclusion, not exclusion.
• Enshrined PBS via EIP-4844 and Danksharding is the long-term fix.

Flashbots' SUAVE aims to decentralize the block building market itself. It proposes a chain-agnostic, decentralized mempool and block builder network.

• Decouples transaction flow from any single entity's control.
• Uses preferences and auctions to match users with builders.
• If successful, it commoditizes block building, reducing dominance risk.

Short-term, validators can commit to block diversity. Ethereum's Pledge and tools like Rated.Network allow validators to publicly commit to selecting the most profitable header, regardless of censorship.

• Creates social and reputational pressure.
• Lido, Coinbase, and Rocket Pool have signed the pledge.
• Provides measurable accountability for the ~$80B+ in staked ETH.

Protocols can architect around the risk. Systems like UniswapX, CowSwap, and Across use intents and solvers, abstracting the user from the underlying execution layer.

• User submits a desired outcome, not a specific transaction.
• A network of solvers competes to fulfill it, bypassing a censoring builder.
• Shifts censorship risk from user to solver network, which is harder to target.

The Problem: Dominant builders like Flashbots control ~90% of Ethereum blocks, creating a single point for OFAC compliance.\nThe Solution: SUAVE creates a decentralized, chain-agnostic mempool and executor network.\n- Separates preference (user intents) from execution (block building)\n- Enables permissionless, competitive block building to break cartel dominance

The Problem: Solo validators can't compete with centralized builders on MEV extraction, forcing them to outsource and cede control.\nThe Solution: EigenLayer's restaking pools capital to create decentralized, high-performance block building services.\n- Slashing conditions enforce censorship-resistant commitments\n- EigenDA provides high-throughput data availability, enabling cheap, private transaction bundles

The Problem: Frontrunning and censorship are trivial when transactions are public in the mempool.\nThe Solution: Shutter uses threshold cryptography to encrypt transactions until block inclusion.\n- Keyper network manages decryption keys in a decentralized manner\n- Neutralizes frontrunning and enables forced inclusion, bypassing builder censorship

The Problem: Proposer-Builder Separation (PBS) is incomplete; builders still hold ultimate power.\nThe Solution: Protocol-level MEV-burn and enforced PBS commit to blocks via crLists.\n- Burns excess MEV revenue, reducing economic incentive for centralization\n- crLists allow validators to force inclusion of censored transactions

Etherean PBS funnels ~90% of blocks through a handful of dominant relays. This creates a centralized chokepoint for OFAC compliance and transaction filtering.

• Top 3 relays control majority of block space.
• Censored blocks have exceeded 50% post-Merge.
• Protocols like Tornado Cash are de facto blacklisted.

Bake PBS directly into the protocol consensus layer to eliminate the trusted relay layer and decentralize block production.

• Removes relay cartel as a trusted third party.
• Enforces credibly neutral block inclusion via protocol rules.
• Long-term Ethereum roadmap item (post-Danksharding).

Dependence on Infura/Alchemy creates a single point of failure for dApp data and transaction submission, enabling infrastructure-level censorship.

• Two providers serve majority of mainnet traffic.
• Geopolitical pressure can filter access globally.
• dApps fail if primary RPC goes down or censors.

Architect for redundancy using networks like Pocket Network or incentivized P2P light clients (Helios, Succinct) to bypass centralized gateways.

• Pocket Network uses ~50k+ nodes for decentralized RPC.
• Light clients verify chain data locally, removing RPC dependency.
• Implement fallback strategies across multiple providers.

Rollups like Arbitrum and Optimism use a single, permissioned sequencer. This grants the operator unilateral power to reorder, delay, or censor transactions.

• No forced inclusion guarantees for users.
• Cross-domain MEV extraction is centralized.
• Upgrade keys often held by multi-sigs, creating governance risk.

Adopt shared sequencer networks (Espresso, Astria) or Based Rollups that outsource sequencing to Ethereum proposers, inheriting L1's decentralization and censorship-resistance.

• Shared Sequencers decentralize ordering across many nodes.
• Based Rollups use Ethereum builders for sequencing, enabling native cross-rollup atomic composability.
• Forced inclusion via L1 becomes viable.