Abstract
This proposal introduces a purely economic solution to Ethereum’s unbounded state growth. It decouples the cost of creating permanent state from the dynamic EIP-1559 gas market by requiring a fixed ETH “Storage Bond” for new state creation. By locking this ETH, the creator pays an “implicit rent” via the opportunity cost of missed staking yield, achieving continuous economic pressure to clean up state without risking smart contract insolvency. Furthermore, it introduces a dynamic, protocol-funded refund mechanism for state deletion that acts as a “State Thermostat,” incentivizing developers to prune temporary data.
Motivation
Ethereum’s active state grows by roughly 100 GB per year. While Data (blobs) and Execution (ZK-rollups) scale gracefully, the permanent storage of state objects remains the primary bottleneck to node decentralization.
Previous attempts to solve this have failed due to severe trade-offs:
State Expiry: Breaks backwards compatibility and introduces the cryptographic nightmare of proving “inexistence.”
Classic State Rent: Risks smart contract insolvency. If a contract must pay continuous rent from its balance, it can run out of funds and “die,” destroying DeFi composability.
Historical Gas Refunds: Created Gas Tokens (e.g., CHI). Because state creation was priced in fluctuating Gas, it became an arbitrage loop.
This EIP argues that the cost of permanent network burden should not fluctuate with temporary block congestion. By requiring a fixed, upfront ETH bond, we introduce a safe, continuous cost to state storage: Opportunity Cost. As the state grows, ETH is continuously locked out of circulation. The yield forfeited on this locked capital acts as a natural, ongoing rent that safely prices permanent storage while aligning perfectly with the Ultra Sound Money narrative.
Specification
1. New Transaction Field: maxStateFee
Transactions will be updated to include a new field: maxStateFee (denominated in wei).
This represents the maximum amount of ETH the sender is willing to pay strictly for the creation of net-new state slots during the transaction’s execution.
Wallets can estimate this by simulating the transaction. For example, a simple ERC-20 transfer to a new address will require a predictable state fee. A Uniswap swap, which typically only modifies existing state, would have a
maxStateFeeclose to zero.If a transaction attempts to create new state that exceeds the
maxStateFee, the transaction reverts.
2. The Fixed Storage Bond (STATE_CREATION_FEE)
The cost to initialize a new storage slot (SSTORE from zero to non-zero), create a new contract (CREATE/CREATE2), or initialize a new account balance will be decoupled from Gas.
A fixed fee (e.g.,
0.001 ETH) is deducted directly from the transaction sender’smaxStateFeepayload.This fee is burned or sent to a locked protocol contract.
3. The State Thermostat (Dynamic Refunds)
When a storage slot is set back to zero, or a contract is destroyed (SELFDESTRUCT or equivalent clearing opcode), the transaction sender receives a refund.
Decoupled from Gas: The refund is paid in fixed ETH, preventing the arbitrage loops of historical Gas Tokens.
The Thermostat Target: The protocol maintains a target size for the active state (e.g., 300 GB).
Dynamic Payout: If the network state exceeds the target, the protocol automatically increases the refund payout by minting new ETH. If the state drops below the target, the refund returns to a baseline.
This creates a lucrative, protocol-subsidized bounty for developers to design applications that clean up their temporary state.
Rationale
The Implicit Rent: Opportunity Cost and Missed Yield
The most significant innovation of the Storage Bond is how it re-introduces the concept of “state rent” without breaking smart contracts. In the modern Ethereum economy, ETH is a yield-bearing asset (via native protocol staking or LSTs). Therefore, locked ETH has a measurable opportunity cost.
The Mechanics of Implicit Rent: If a developer locks up a specific amount of ETH to create permanent state, they forfeit the native yield that ETH could have generated. If the prevailing staking yield is 2% APY, the developer is effectively paying a continuous, implicit “rent” equal to 2% of their bond every year.
Zero Insolvency Risk: Unlike traditional state rent, which explicitly deducts from a contract’s balance and risks breaking the application if funds run dry, implicit rent simply locks capital. If a contract is abandoned, it just continues to miss out on yield, but the application’s logic and balances remain fully intact and solvent.
Why a fixed ETH bond instead of Gas?
Gas measures transient computational effort. State measures permanent hardware burden. Pricing a permanent burden using a transient metric inherently misprices state depending on network traffic. A fixed ETH bond eliminates gas arbitrage.
Why mint new ETH for refunds?
To incentivize the cleanup of legacy state (created before this EIP), the protocol must offer a financial reward. Because historical state fees were burned, the protocol must mint ETH to pay these “cleaning bounties.” However, because new state creation requires a massive upfront ETH lockup, the overall system remains heavily deflationary and strengthens ETH’s scarcity.
Backwards Compatibility
This EIP does not break existing smart contracts. Legacy contracts do not need to be upgraded to pay “rent,” and their state will not spontaneously expire. The only change is to the transaction sender, who must ensure their wallet populates the new maxStateFee field if their interaction results in net-new state creation.
Security Considerations
Front-Running / Pre-Fork Spam
To prevent attackers from spamming the network with state before the EIP activates to harvest refunds later, the rollout must be multi-phased:
Phase 1: Introduces the high
STATE_CREATION_FEEwith minimal (or zero) refunds.Phase 2: A later hardfork activates the dynamic State Thermostat to increase refunds only after the arbitrage window has closed.
Griefing
Because the fee is deducted from the transaction sender (via maxStateFee) and not the smart contract’s internal balance, malicious actors cannot drain a contract’s ETH by forcing it to create state.
Permanence and Lost Keys
If a user or contract loses the ability to modify or delete their state (e.g., lost private keys, or a smart contract deployed without a clearing equivalent), the state remains permanently immutable, and the initial Storage Bond remains permanently locked. The protocol does not allow third parties to delete state they do not cryptographically own to claim refunds. This guarantees that abandoned state acts as a permanent, irrecoverable deflationary sink for the ETH supply.
