The Unfunded Mandate of Eternal Data Availability

Data permanence is a subsidy. Blockchains like Ethereum and Solana guarantee data availability for state execution, but the one-time transaction fee does not fund perpetual storage. This creates an unfunded mandate where future network participants bear the cost of historical data.

The cost compounds silently. Unlike compute, which is ephemeral, storage costs accumulate linearly with chain age. Protocols like Arbitrum and Optimism push this cost to layer-1 Ethereum via calldata, merely deferring the economic problem.

Rollups externalize the true cost. The dominant scaling narrative assumes cheap, permanent L1 data. This reliance makes the entire modular stack dependent on Ethereum's social consensus to maintain data availability, not a sustainable economic one.

Evidence: Ethereum's historical data now exceeds 15TB. Storing this via decentralized providers like Filecoin or Arweave would cost millions annually, a cost not captured by gas fees paid years ago.

The DA subsidy is unsustainable. Layer 2s like Arbitrum and Optimism pay for data availability (DA) on Ethereum, treating it as a variable operational cost. This creates a perpetual economic drain where sequencer profits are decoupled from the long-term cost of storing their state.

Rollups externalize their largest cost. The security model assumes Ethereum validators will forever store and serve this data, but the fee market only pays for short-term inclusion. This is a classic tragedy of the commons where the burden of historical data falls on a system not compensated for it.

Proof systems are not a panacea. Validity proofs from zkEVMs like zkSync verify state transitions, but they don't store the data needed to rebuild the chain. A prover with the only data copy becomes a centralized point of failure, negating decentralization guarantees.

Evidence: Ethereum's blob fee market is volatile. A single day of peak activity can cost an L2 like Base over $100k in DA fees, demonstrating the model's exposure to unpredictable, non-recoverable costs.

Decentralized identity requires permanence. Systems like Verifiable Credentials and Soulbound Tokens store attestations on-chain, creating an infinite data liability. This liability is a perpetual cost that no single entity or protocol can guarantee to pay.

Blockchains are not archival databases. Layer 1s like Ethereum and rollups like Arbitrum prune state to manage costs. Identity data, which is rarely accessed, becomes a prime candidate for deletion under state expiry or high gas fee pressure.

The economic model is inverted. Users pay once to mint an identity, but the network bears the recurring cost of storing it forever. This creates a tragedy of the commons where the system's utility directly increases its eventual insolvency.

Evidence: The Ethereum Foundation's Purge roadmap explicitly aims to reduce historical data burden. Arweave exists precisely because general-purpose blockchains are economically unfit for permanent storage, yet identity protocols rarely integrate it by default.

The 'Historical Data' Rebuttal is the primary defense: only new state needs DA, old data can be pruned. This is a cost-shifting fallacy. It assumes a permanent, altruistic archive like Arweave or Filecoin will always exist to serve historical proofs, creating a critical systemic dependency on an unpaid third party.

Pruning creates a time bomb. Protocols like Celestia and EigenDA explicitly design for pruning. A rollup that prunes its data outsources its liveness to external historians. This fragments security guarantees and reintroduces the very trust assumptions that modularity aimed to eliminate. The system's security decays over time.

Evidence from Ethereum's roadmap proves the point. Proto-Danksharding (EIP-4844) introduces blob storage with a ~18-day expiry. This is a deliberate economic policy, not a technical limitation. It forces rollups to implement their own long-term DA solution, passing the cost and complexity buck to individual application layers.

Ethereum's consensus is the bottleneck. The core mandate for data availability layers is permanence, not just temporary storage. L1s like Ethereum treat data as a consensus resource, creating a permanent cost that rollups must subsidize indefinitely.

Storage is not a service. Projects like Filecoin or Arweave provide raw storage, but they lack the cryptoeconomic guarantees of a live consensus layer. Their models work for archiving, not for the real-time state proofs required by validity or fraud proofs.

The solution is a service abstraction. Protocols must treat data availability as a verifiable compute service, not a storage product. This shifts the economic model from paying for bytes to paying for a cryptographic proof of data persistence and retrievability.

Evidence: Celestia's modular design separates execution from consensus and data availability, creating a dedicated market for DA. EigenLayer's restaking introduces cryptoeconomic security as a service, allowing new DA layers to bootstrap trust from Ethereum validators.