# 10 — Chain Selection: Prototype on L2, Decide at Phase III > *What we actually need from a chain is narrower than most crypto-native projects assume. Prototype on an existing L2; re-evaluate own-L1 only if empirical evidence or investor commitments require it.* ## The requirements — what the chain must provide 1. **Append-only commitment of signed events.** Attestations, cascades, validator credentials. Any chain provides this; the question is cost and tooling. 2. **Smart-contract execution for payment settlement + attestation registry.** EVM chains cover this amply. Non-EVM (Solana, Move chains, Cosmos SDK) are viable but narrower developer ecosystem. 3. **Per-transaction cost low enough that attestations + payment flows are economically viable.** Rules out Ethereum L1 mainnet (~$1–10 per tx in 2026). Targets L2s ($0.001–0.01) or alternative L1s (Solana, Near, Avalanche). 4. **Finality fast enough that cascade events propagate in useful time.** Sub-minute soft-finality acceptable. Seven-day optimistic-rollup fraud-proof window is tolerable for settlement but wants faster soft-confirmations for consumer UX. 5. **Data-availability guarantees for attestation hashes.** Every modern chain provides this. Celestia, EigenDA, Avail offer purpose-built DA if volume scales beyond EVM rollup comfort. 6. **Regulatory clarity of the chain itself.** A chain with unresolved regulatory questions adds exposure the token design cannot mitigate. Ethereum-ecosystem chains and Bitcoin have the clearest posture in 2026. 7. **Mature tooling: developer experience, RPCs, indexers, wallets.** EVM wins here by a large margin. Non-EVM chains have caught up but still lag in tooling density. 8. **Community and investor overlap with the Veritas mission.** Some chains have crypto-native investor communities more aligned with open-infrastructure plays than others. Optimism's Retroactive Public Goods Funding, Arbitrum's DAO grants, Base's creator economy — each has a specific shape. ## The requirements — what the chain does *not* need to provide - **Transaction throughput beyond ~100 tx/sec sustained.** Attestation volume at Phase II is under 1 tx/sec; at Phase III probably under 10 tx/sec. Peak-scenario calculations put sustained throughput under 100 tx/sec. Most L2s do 100+ easily. - **Zero cost.** Attestation is not free; real work is being done. The chain's cost just needs to be bounded relative to economic value. $0.001/tx on a modern L2 is fine. - **Novel cryptographic primitives.** Veritas uses standard primitives (Ed25519 / ECDSA signatures, SHA-256 hashes, optionally BLS aggregate signatures for validator quorums). Every modern chain supports these. - **MEV-resistance.** Veritas has no MEV because there is no arbitrage or ordering-sensitive transaction. - **Cross-chain composability.** Nice to have; not required. Veritas's primary interface is HTTP APIs, not on-chain composability with DeFi protocols. ## Candidate chains — evaluation ### Tier 1 — Strong fit for Veritas Phase II #### Base (OP Stack, Ethereum L2) - **Cost.** ~$0.0001–0.001/tx (2026). - **Finality.** ~1 second soft; ~7 days hard (optimistic rollup fraud proof window). - **Ecosystem.** Coinbase-backed; growing fast; regulatory posture clear (US-compliant). - **Developer tooling.** EVM-standard; Hardhat / Foundry / OpenZeppelin all work. - **Fit.** Very strong. Regulatory clarity + cheap + fast + mature tooling. - **Risk.** Coinbase has outsized influence on Base governance; concentration risk. #### Optimism - **Cost.** ~$0.0005/tx. - **Finality.** ~1 second soft; 7 days hard. - **Ecosystem.** Mature; governance via Optimism Collective; Retroactive Public Goods Funding is an actual grants programme Veritas could apply to. - **Developer tooling.** Strongest OP-Stack tooling ecosystem. - **Fit.** Strong. The public-goods-funding angle directly aligns with Veritas's non-profit + infrastructure framing. - **Risk.** OP Stack upgrades affect every OP-Stack L2; coordination risk across the superchain. #### Arbitrum - **Cost.** Slightly higher than Base / Optimism historically; similar in 2026. - **Finality.** ~1 second soft; 7 days hard. - **Ecosystem.** Arbitrum DAO grants. Largest L2 by TVL. - **Developer tooling.** Mature. - **Fit.** Strong; slightly less public-goods-aligned than Optimism but equally credible. - **Risk.** Governance by Arbitrum DAO has had contentious episodes; smaller-holder-voice concerns. ### Tier 2 — Worth evaluation but not Phase II default #### Polygon PoS / zkEVM - **Cost.** Very cheap. - **Finality.** Fast. - **Ecosystem.** Large but reputationally mixed; Polygon team pivots have been frequent. - **Fit.** Technical fit is fine; ecosystem alignment is weaker for a public-infrastructure project. #### zkSync Era, Starknet, Linea - **Cost.** Low (ZK rollup). - **Finality.** Faster hard-finality than optimistic (hours vs days). - **Tooling.** Less mature than OP-Stack; differentiated VMs (Cairo on Starknet) have adoption friction. - **Fit.** Promising for Phase III+; not optimal for Phase II. ### Tier 3 — Not Phase II candidates - **Solana.** High performance; non-EVM tooling; ecosystem less aligned with institutional / public-interest projects; 2022–2023 outages left operational-reliability concerns. - **Avalanche, Near, Aptos.** Technical fit varies; ecosystem alignment weaker for this use case. - **Ethereum mainnet.** Too expensive at Phase II volumes; worth considering for Phase III settlement anchoring. - **Bitcoin + L2s (Stacks, Rootstock).** Tooling significantly narrower; not recommended. ## Own-L1 — when and why The argument for own-L1: 1. **Permissionlessness neutrality.** An own-L1 is not subject to another chain's upgrade decisions, governance capture, or regulatory decisions. For Veritas's hostile-validator design, this may matter — a state actor less willing to trust an Ethereum L2 (because of US regulatory connection) may be more willing to trust a neutrally-governed own-chain. 2. **Protocol-native governance.** An own-L1 can encode Veritas governance directly at the consensus layer, not just at the smart-contract layer. 3. **Economic alignment with validators.** An own-L1 can align validator economics with chain-security economics (Veritas validators also validate the chain). This is the Cosmos AppChain thesis. 4. **Investor preference.** Crypto-native investors often prefer own-L1 because it concentrates economic upside in the protocol's token rather than distributed across the underlying chain's ecosystem. The argument against own-L1: 1. **Security budget.** A Cosmos-SDK chain requires ~$500K/year in validator-set security budget (realistic minimum). An OP-Stack L2 requires ~$100K/year sequencer + DA budget. Ethereum L1 gives security "for free" (subsidised by L1 staking). Own-L1 cost scales with threat model. 2. **Tooling reinvention.** Wallets, indexers, block explorers, RPC infrastructure — all need dedicated builds or integrations. L2 inherits all of Ethereum's tooling. 3. **Time to ship.** Own-L1 adds 6–12 months to the Phase II timeline. 4. **Regulatory complexity.** Operating a novel L1 introduces questions a ride-existing L2 inherits cleanly from the parent chain's precedents. 5. **Smaller community.** Fewer developers know how to operate, audit, and build on a novel L1. ### Decision framework Phase II prototypes on **Base or Optimism** (exact choice depends on investor commitments and grants programme alignment at Phase II funding phase). At **Phase III gate** (month 18), decide own-L1 vs continue-on-L2 based on: 1. **Empirical necessity.** Has any chain property not fit during Phase II operation? If everything worked on L2, there is no empirical case for own-L1. 2. **Investor commitments.** Have crypto-native investors (Protocol Labs, Filecoin Foundation, specialised infrastructure DAOs) committed Phase III funding contingent on own-L1? If so, weigh. 3. **Hostile-frame participation.** Have state-actor or ideologically-opposed validator cohorts refused participation specifically citing the underlying chain's jurisdictional baggage? If yes, own-L1 neutrality argument applies. 4. **Governance / upgrade conflicts.** Has the parent chain's governance made decisions Veritas would have preferred otherwise, and does the friction justify migration? Absent at least one of the four, continue on L2 indefinitely. The base case is: **L2 forever**. ## Cosmos / Polkadot / OP Stack as own-L1 candidates If the Phase III decision favours own-L1, the candidates are: ### Cosmos SDK - **Pros.** Mature. IBC connectivity to other Cosmos chains. Tendermint BFT consensus well-understood. Validator-set model flexible. - **Cons.** Smaller developer community than EVM. Cosmos-specific tooling narrower. Requires real security budget. ### OP Stack - **Pros.** Forks of Optimism / Base. Inherit OP-Stack tooling. Relatively cheap to launch (sequencer + DA). "OP chain" is a first-class concept. - **Cons.** Still depends on Ethereum L1 for settlement; not fully sovereign. For the "hostile-validator neutrality" argument, OP-Stack does not solve the problem — the chain still has Ethereum-ecosystem connection. ### Polkadot parachain - **Pros.** Shared-security model; validator infrastructure provided. - **Cons.** Polkadot ecosystem has cooled; parachain auctions expensive; DOT-holder economics add complexity. ### Sovereign rollup (Celestia-based) - **Pros.** Own execution + consensus, Celestia for DA. Maximum sovereignty. - **Cons.** Newer ecosystem; tooling still maturing; smaller track record. ## Phase II recommendation — concrete **Chain: Optimism (OP Mainnet).** **Rationale.** Public-goods funding alignment (RetroPGF); mature tooling; cost profile fits; regulatory clarity via Ethereum ecosystem; long-term governance by Optimism Collective aligns with non-profit spirit; sequencer decentralisation roadmap credible. **Alternative. Base.** If a partnership with Coinbase emerges that provides distribution / financial infrastructure advantages and if Coinbase's concentration risk is mitigated by specific commitments, Base is a credible alternative. **Contracts deployed.** - `veritas-attestation-registry` — on-chain event log of attestations. - `veritas-cascade-log` — on-chain retraction events. - `veritas-validator-credentials` — self-issued + foundation-issued VCs; revocation list. - `veritas-token` — ERC-20 utility token. - `veritas-treasury` — multisig + parametric burn/buyback. - `veritas-investigation-market` — escrow + commissioning + payment. - `veritas-refusal-registry` — signed foundation document on-chain. ## Critical analysis **1 — Betting on one L2 creates lock-in.** Response: the core schema is chain-agnostic; migration to another L2 or to own-L1 is mechanically possible (though operationally painful). Lock-in is real but not catastrophic. **2 — The hostile-validator argument for own-L1 is real and was made seriously.** Response: acknowledged. The Phase III gate specifically evaluates whether hostile validator cohorts refuse L2 participation. If yes, migrate. **3 — Chain costs scale with attestation volume and may surprise.** Response: pricing curves are published by L2 operators; Veritas can model worst-case. At 100 tx/sec sustained, Optimism costs are $3K/day (~$1M/year). At 10 tx/sec (Phase III realistic), $300/day (~$110K/year). Manageable. **4 — Cross-chain migration is expensive.** Response: yes. Commitment to Optimism for Phase II is real. Migration to Phase III destination (own-L1 or different L2) requires formal decision + user communication + state migration scripts. **5 — What if Optimism has a security incident?** Response: the chain has its own security posture and history; Veritas inherits both. Diligence during Phase I includes Optimism security review. Contingency: Veritas state is reconstructible from event logs; migration is possible even under incident conditions. ## Open questions - Is the Optimism Superchain interop (OP-Stack chains sharing sequencer) material to Veritas? Possibly; affects whether to deploy on Optimism directly or a derivative. - What's the latency of OP Mainnet sequencer under peak load, and does it affect cascade propagation? - Does Veritas benefit from an operator-run sequencer or can it rely on chain sequencer indefinitely? - What's the Phase III migration checklist if own-L1 is chosen? ## What we'd build - **`veritas-l2-deploy`** — deployment scripts for Optimism / Base / Arbitrum. - **`veritas-contract-suite`** — Solidity contracts for all on-chain Veritas operations. - **`veritas-rpc-proxy`** — RPC intermediate with rate-limiting + caching for consumer-facing reads. - **`veritas-subgraph`** — The Graph-style indexer for fast chain-state queries. - **`veritas-l2-monitoring`** — ops dashboards for chain health, gas costs, transaction volumes. - **Chain-migration runbook** — phase-III documentation for own-L1 transition if chosen. - **Phase-III decision artefact** — at month 18, a formal decision document evaluating the four own-L1 criteria.