Why RPC Censorship Is an Existential Attack Vector

Centralized RPC providers like Infura and Alchemy are the de facto gatekeepers for ~80% of Ethereum traffic. They comply with OFAC sanctions, censoring transactions from sanctioned addresses. This creates a permissioned layer atop a permissionless protocol.

• Single Point of Failure: A handful of US-based entities control access.
• Compliance as a Weapon: Regulatory pressure turns infrastructure into an enforcement tool.
• ~$50B+ TVL at risk of filtered access.

The MEV supply chain—searchers, builders, and relays—has consolidated around a few dominant players like Flashbots. This creates centralized points where transaction ordering can be manipulated or censored.

• Builder/Relay Dominance: Top 3 builders control >80% of Ethereum blocks post-Merge.
• Censorship-by-Proxy: Relays can exclude OFAC-sanctioned transactions from blocks.
• Latency Arms Race: Favors well-capitalized, centralized actors, further entrenching power.

The counter-trend is the rise of decentralized RPC networks like Chainscore, Lava Network, and Ankr. These aggregate geographically distributed nodes, making censorship economically irrational.

• Incentive-Aligned Nodes: Providers are rewarded for uncensored, low-latency service.
• Fault Tolerance: No single legal jurisdiction or entity controls the endpoint.
• Intent-Based Future: Paves the way for UniswapX-style, censorship-resistant transaction routing.

The canonical case. In August 2022, OFAC sanctioned Tornado Cash smart contracts, forcing centralized RPC providers like Infura and Alchemy to censor related transactions. This created a two-tiered Ethereum, where compliant nodes filtered blocks, directly violating the network's credibly neutral base layer.

• Impact: Broke the core promise of permissionless access.
• Result: Spawned the MEV-Boost Relay censorship crisis, with relays filtering ~70% of blocks at its peak.

Post-Merge, validators outsourced block building to a handful of MEV-Boost Relays. When OFAC pressure hit, major relays like Flashbots began censoring transactions. This created a systemic risk where proposer-builder separation (PBS) became a censorship vector.

• Mechanism: Relays filtered OFAC-sanctioned addresses from blocks they built.
• Solution Push: Drove adoption of censorship-resistant relays (e.g., Ultra Sound, Agnostic) and in-protocol PBS research.

In 2022, Infura's misconfigured geo-blocking filter inadvertently blocked access to the entire Arbitrum network for users in certain regions. This wasn't malicious intent but demonstrated the single point of failure risk when dApps rely on a monolithic RPC provider.

• Revealed: Infrastructure fragility and the non-sovereignty of users.
• Catalyst: Accelerated demand for decentralized RPC networks like Pocket Network and Chainscore's own distributed gateway.

High-performance chains face unique risks. On Solana, the economic model led to RPC node centralization among a few entities. During peak congestion (e.g., NFT mints), these nodes could prioritize or deprioritize transactions based on fee bids or origin, creating a de facto censorship layer.

• Problem: RPCs act as the gatekeeper to the mempool.
• Architectural Flaw: Highlights the need for decentralized RPC layers and local client advocacy, even in high-throughput environments.

Centralized RPC providers like Infura and Alchemy control access for ~80% of Ethereum traffic. This creates a critical chokepoint for censorship, as seen when they blocked Iranian users or specific wallet addresses.\n- Network-Level Blacklisting: A single provider can deny access to entire protocols or regions.\n- Protocol Risk: DApps built on a single endpoint inherit its failure and compliance risks.

Protocols like POKT Network and Lava Network replace monolithic providers with a permissionless network of independent node operators. This creates a censorship-resistant mesh where requests are routed across a global pool of nodes.\n- Sybil-Resistant: Operators stake native tokens, aligning incentives for liveness and correctness.\n- Redundancy: No single operator can block or filter traffic, ensuring protocol neutrality.

Decentralized RPCs aren't just about censorship resistance; they create a competitive market for performance. Operators compete on latency, geographic coverage, and reliability to earn rewards.\n- Cost Efficiency: Market dynamics drive down costs versus centralized premium APIs.\n- Performance Optimization: Latency-sensitive applications can route to the fastest, nearest nodes automatically.

For protocols and DAOs, decentralized RPC is infrastructure sovereignty. It eliminates dependency on corporate entities whose interests may diverge from the chain's ethos, as seen with Tornado Cash sanctions.\n- Self-Hosted Gateways: DAOs can run their own gateway nodes, guaranteeing uncensorable access for their community.\n- Multi-Chain Abstraction: Networks like Lava provide unified access to Ethereum, Cosmos, Solana, reducing integration complexity.

Relying on a single provider like Infura or Alchemy creates a centralized kill switch. A government order or corporate policy change can censor entire wallets or dApps, breaking core Web3 guarantees.

• Risk: A single entity can blacklist addresses or filter transactions.
• Impact: $10B+ TVL in DeFi protocols is exposed to this vector.
• Example: The 2022 OFAC sanctions on Tornado Cash demonstrated this vulnerability.

Distribute RPC requests across a permissionless network of node operators. Projects like POKT Network and Lava Network use cryptoeconomic incentives to ensure liveness and neutrality.

• Key Benefit: Censorship resistance via geographic and political distribution.
• Key Benefit: Improved reliability and uptime through redundancy.
• Trade-off: Slightly higher latency (~50-100ms) vs. centralized giants.

The ultimate defense is reducing dependency on third-party RPCs. Teams should run their own nodes or use lightweight clients. Ethereum's Erigon and Reth are built for this.

• Action: Run a full node for critical infrastructure and fallback.
• Action: Implement RPC failover logic in your application layer.
• Reality Check: Requires ~2TB SSD and ongoing DevOps overhead.

Move beyond simple transaction broadcasting. Let users express what they want (e.g., "swap X for Y"), not how to do it. Systems like UniswapX, CowSwap, and Across use solvers that are RPC-agnostic.

• Key Benefit: User transactions become unstoppable; solvers find any path.
• Key Benefit: Better execution via MEV protection and optimized routing.
• Future: This pattern, championed by Anoma, makes the RPC layer less critical.