Running Ethereum Nodes Without Full State

State growth is exponential. Each new smart contract and user account permanently bloats the global state, a dataset every full node must store and sync. This creates a centralizing force, pricing out individual operators.

Statelessness is the solution. The core roadmap shifts the burden of state storage from nodes to clients. Nodes verify blocks using cryptographic proofs (like Verkle Trees) instead of holding the entire state, radically reducing hardware requirements.

Erigon and Reth lead. Client teams are pioneering archive node pruning and flat storage models to mitigate the issue today. Their work proves state size, not chain throughput, is the real bottleneck for node decentralization.

Evidence: A full Ethereum archive node requires over 12 TB of SSD storage. Without statelessness, this cost will make solo staking and lightweight clients economically impossible within 5 years.

State growth is terminal. Ethereum's execution state grows ~40GB/year, making solo node operation a luxury. This centralizes the network and creates a hard scalability ceiling for all L2s like Arbitrum and Optimism.

Stateless clients verify, they don't store. A node validates blocks using a Verkle proof for each transaction, not a local database. This decouples validation cost from total state size, enabling light clients with full security.

The bottleneck shifts to proof generation. Provers like Risc Zero or Succinct Labs must generate these proofs, creating a new market for specialized hardware. This is the computational cost of eliminating state bloat.

Witness data is the new bandwidth tax. Nodes receive ~1MB of proof data per block instead of storing 1TB. This trade-off is mandatory; the network's security depends on widely distributable validation.

Statelessness redefines node hardware. Full nodes no longer need terabytes of fast SSD storage for the world state, shifting the bottleneck to bandwidth and CPU for verifying proofs.

Verkle Trees enable this shift. They replace Merkle Patricia Tries with cryptographic proofs that are constant-sized, making witness data portable and verification efficient.

Light clients become first-class citizens. Protocols like the Portal Network allow resource-constrained devices to securely query and verify chain data without trusting centralized RPC providers.

The economic model inverts. Operational costs shift from capital expenditure on storage to operational expenditure on compute, lowering barriers for home operators and decentralizing the network.