Ethereum Storage and the Future of State

Layer 2s like Arbitrum and Optimism become the primary state managers. They already batch and compress data, making them natural candidates for advanced state management. Their business model shifts from simple sequencing to offering state-as-a-service to their app ecosystems, locking in moats. Expect them to integrate Verkle proofs and EIP-4444 execution clients first, forcing other chains to rent their infrastructure.

Companies like RISC Zero and Succinct become critical infrastructure. Stateless verification requires constant generation of state proofs. These firms, already building general-purpose ZK provers, will pivot to become the AWS for state attestations. Their market expands from L2 validity proofs to servicing every light client and bridge. The bottleneck shifts from block space to proof generation latency and cost.

Traditional full node operators face existential cost pressure. EIP-4444's historical data expiry destroys the archive node business model. Operators must either:

• Specialize as state providers, serving data to light clients for fees.
• Shut down, consolidating the network into fewer, larger professional entities. This increases resilience risk but is the inevitable cost of scalability.

Dapps that naively store everything on-chain get wrecked. Protocols must adopt a state-minimal design philosophy, leveraging storage proofs (like those from Axiom or Herodotus) and stateless clients. This unlocks new architectures: trustless historical data queries, cross-chain statesync, and light-client-based bridges. Winners will be apps built from day one for this paradigm.

Proposer-Builder Separation (PBS) centralizes state access. Builders, who construct blocks, will be the only entities with low-latency, full-state access. This creates a powerful cartel. Solutions like suave aim to democratize this, but the default path is a builder oligopoly that also controls the most efficient state management, extracting maximal value from the chain.

Celestia, Avail, and EigenDA win as state growth decouples from execution. Their value proposition skyrockets when Ethereum actively prunes history. They become the canonical, cost-effective settlement layer for state commitments. Rollups will post data and state roots here, creating a multi-chain hub for state verification. This solidifies the modular stack and drains value from monolithic chains.

Every new account and smart contract storage slot adds ~100 bytes to Ethereum's state, which must be stored by every full node. This creates prohibitive hardware requirements, centralizing consensus power.

• Current state size exceeds 1 TB and grows by ~50 GB/year.
• Running an archive node requires ~12 TB+ of fast SSD storage.
• This directly undermines Ethereum's decentralization and censorship resistance.

Ethereum's core upgrade path replaces Merkle Patricia Tries with Verkle Trees, enabling stateless clients. Validators no longer need to store the full state to verify blocks.

• Enables light clients with full security guarantees.
• Reduces witness sizes from ~1 MB to ~150 bytes per block.
• Paves the way for The Purge (EIP-4444), allowing historical data expiration.

Execution-layer rollups like Arbitrum, Optimism, and zkSync externalize state growth. They post compressed proofs or data back to Ethereum L1, which acts as a secure data availability layer.

• L1 state growth slows; activity shifts to L2s.
• Enables experimental state models (e.g., Optimism's Plasma-style fault proofs).
• Creates a multi-layered state hierarchy, similar to CPU cache vs. RAM.

Proposals like EIP-4444 mandate clients to stop serving historical data older than one year. This forces the ecosystem to build decentralized history networks like Ethereum Portal Network or Theta.

• Forces innovation in decentralized storage (Portal Network, BitTorrent).
• Cuts node storage requirements by ~90% for non-archive nodes.
• Transforms Ethereum into a live state machine, not a permanent archive.

Proposer-Builder Separation (PBS) and rising hardware costs formalize a divide between low-resource validators and professional block builders who must manage full state.

• Centralizes state holding to a few professional entities.
• Builders require high-performance hardware and MEV strategies.
• Validators can remain lightweight, preserving consensus decentralization.

Ultimately, zk-EVMs and validity proofs (e.g., from Scroll, Taiko) can compress state changes into a single cryptographic proof. The L1 only needs to verify the proof, not store the intermediate state.

• Ethereum becomes a verification layer, not a computation layer.
• Enables infinite state scaling with constant L1 footprint.
• Final convergence of modular and monolithic scaling theories.