{ "version": "1.0", "asOf": "2026-08-02", "methodology": "https://riadefi.com/atlas/#method", "license": "Copyright; quotation and citation permitted with attribution", "grades": { "sovereign": "No single issuer or administrator has an address-level freeze function at this layer.", "crypto-backed": "On-chain governance or collateral mechanisms matter, but no conventional issuer blocklist defines the asset.", "hybrid": "On-chain mechanisms coexist with custodians, real-world assets, centralized ordering, or other compellable parties.", "issuer-controlled": "A named organization or administrator controls issuance, transfers, reserves, redemption, or address restrictions." }, "chains": [ { "id": "Ethereum", "displayName": "Ethereum", "reviewed": "2026-07-31", "verdict": "approved", "sovereignty": "sovereign", "stage": "n/a", "sequencer": "None — permissionless proposer set", "escapeHatch": "n/a — settlement layer, no exit dependency", "upgradeControl": "Social consensus + hard fork. No key can change state or seize funds.", "liveness": "No chain halt since the Merge. Finality ~13 minutes. The standing concern is client concentration: Geth held ~50.1% execution-layer share as of Q4 2025, and a supermajority client bug is the one scenario that could finalise a wrong chain. Institutional operators moved toward multi-client stacks through 2026 in response.", "thesis": "The baseline against which every other chain is measured, and the only venue where a sovereignty-first mandate is fully honoured. No sequencer, no upgrade key, no operator who can be compelled. The cost is gas — a client with a small position pays materially more here than on an L2, and that trade-off has to be stated plainly rather than solved by quietly routing them to Base.", "killCriteria": [ "Any single execution client exceeding 66% share", "Any single consensus client exceeding 66% share", "Finality failure lasting more than 1 hour" ], "url": "https://riadefi.com/atlas/chains/ethereum/" }, { "id": "Arbitrum", "displayName": "Arbitrum One", "reviewed": "2026-07-31", "verdict": "approved-with-limits", "sovereignty": "hybrid", "stage": "stage-1", "sequencer": "Centralised — >99% of transactions through a single sequencer", "escapeHatch": "Forced inclusion via the L1 delayed inbox. forceInclusion threshold is 24 hours; in practice a full self-service exit is slow and costly.", "upgradeControl": "DAO plus a security council. Admin keys can upgrade bridge contracts on roughly a 7-day timeline — as at every other top-5 L2.", "liveness": "Sequencer outages have occurred; user funds were never at risk, but access was.", "thesis": "Stage 1 with live permissionless fraud proofs, which is the current state of the art — as of mid-2026 there are still NO Stage 2 rollups. So this is as good as an L2 gets, and it is still not Ethereum. Appropriate for yield-first and tokenised-income mandates where cheap transactions matter more than censorship resistance. NOT appropriate for a sovereignty-first mandate: a client who is paying us to make them un-freezable should not sit behind a single sequencer and a 7-day upgrade key.", "killCriteria": [ "Fraud proof system disabled or made permissioned", "Force-inclusion delay extended beyond 24 hours", "Security council threshold lowered, or upgrade timelock shortened below 7 days", "Sequencer outage exceeding 6 hours" ], "url": "https://riadefi.com/atlas/chains/arbitrum/" }, { "id": "Base", "displayName": "Base", "reviewed": "2026-07-31", "verdict": "approved-with-limits", "sovereignty": "hybrid", "stage": "stage-1", "sequencer": "Centralised — operated by Coinbase, a single regulated US entity", "escapeHatch": "OP-stack forced inclusion via L1. Mechanically available; economically unattractive for small positions.", "upgradeControl": "Improved governance with longer timelocks and a larger security council, but admin keys can still upgrade bridge contracts within ~7 days.", "liveness": "Sequencer outages have occurred. Funds safe, access interrupted.", "thesis": "Stage 1 with live fraud proofs and the best consumer UX of any L2 — cheap, fast, and the natural default for an embedded-wallet onboarding flow. That is exactly why it needs saying clearly: the sequencer is Coinbase. One identifiable, regulated, compellable US company orders every transaction. For a client whose thesis is \"no single party can stop me,\" routing them to Base because it is cheap quietly reintroduces the counterparty they were trying to leave — and it is the same failure as putting a sovereignty client into USDC. Fine for yield-first and tokenised-income. Not for sovereignty-first.", "killCriteria": [ "Coinbase publicly censors or filters transactions at the sequencer", "Fraud proof system disabled or made permissioned", "Upgrade timelock shortened below 7 days", "Sequencer outage exceeding 6 hours" ], "url": "https://riadefi.com/atlas/chains/base/" }, { "id": "Solana", "displayName": "Solana", "reviewed": "2026-07-31", "verdict": "approved-with-limits", "sovereignty": "crypto-backed", "stage": "n/a", "sequencer": "Leader schedule across the validator set — no single operator", "escapeHatch": "n/a — L1, no exit dependency", "upgradeControl": "No admin key over user funds. Client software upgrades are social.", "liveness": "Last confirmed major halt: 2024-02-06, a ~5 hour outage caused by a program-execution bug. Since then: 22+ months of uptime and 100% year-to-date as of March 2026. Frankendancer (Firedancer networking + Agave execution) has run on mainnet since 2025; once Firedancer passes ~25-30% of stake, a single client bug can no longer halt the network. Concentration is moving the other way: validators fell ~2,560 (Mar 2023) to ~795 (Jan 2026), a 68% decline, and the Nakamoto coefficient fell 31 to 20. Validators remain spread across 45+ countries and multiple hosting providers.", "thesis": "UPGRADED from under-review. The earlier memo leaned on \"has halted and required coordinated restart\" without pinning the date — and the date turns out to matter a great deal. The last major outage was February 2024, nearly two and a half years ago, and the structural fix is not a promise but shipped: multi-client execution is live on mainnet. Twenty-two-plus months of uptime is a real record, not a lucky streak, and holding the old verdict would have been stale caution rather than judgement. Liveness and decentralisation are moving in OPPOSITE directions here, and they are different axes. Liveness has improved structurally; validator concentration has worsened materially. Approved-with-limits reflects exactly that split — the chain is usable, the trend in operator count is the thing to watch, and no admin key can seize client funds either way. This unblocks Kamino, Jito and the Jupiter venues to be judged on their OWN merits rather than deferred behind the chain.", "killCriteria": [ "Any full-network halt requiring coordinated restart", "Nakamoto coefficient falling below 15", "Validator count falling below 600", "Firedancer adoption stalling below 25% of stake" ], "url": "https://riadefi.com/atlas/chains/solana/" }, { "id": "Tron", "displayName": "Tron", "reviewed": "2026-07-31", "verdict": "rejected", "sovereignty": "issuer-controlled", "stage": "n/a", "sequencer": "27 elected Super Representatives (dPoS), rotating every 3 seconds", "escapeHatch": "n/a", "upgradeControl": "Nominally a DAO. In practice voting power concentrates among a few large holders including exchanges and entities aligned with the founder.", "liveness": "No notable halts. Throughput is not the concern here.", "thesis": "REJECTED, and the deciding fact is not the validator count. 27 Super Representatives is centralised but arguable. What is not arguable: in 2024, $732M of Bitcoin was removed from the USDD stablecoin reserve WITHOUT a DAO vote. A governance structure that can be bypassed at that scale is decoration. Tron's own former CTO described the network as \"pseudo-decentralised\". Add sustained regulatory scrutiny of the founder and associated entities, and this fails every property the sleeve exists to provide. Material because JustLend holds ~$3.3B here — the largest single protocol left without a memo, and it is disposed of entirely by this chain verdict.", "killCriteria": [ "n/a — rejected" ], "url": null }, { "id": "Hyperliquid L1", "displayName": "Hyperliquid / HyperEVM", "reviewed": "2026-07-31", "verdict": "rejected", "sovereignty": "issuer-controlled", "stage": "n/a", "sequencer": "Permissioned validator set — 21 validators as of March 2026 (16 at launch)", "escapeHatch": "None independent of the validator set. The native bridge is operated BY those validators, so bridge security and chain security are the same assumption.", "upgradeControl": "Centralised governance. Core code is not open source.", "liveness": "Fast finality by design; that is the tradeoff being made.", "thesis": "REJECTED for this client. Genuinely impressive engineering and the fastest-growing venue in the survey, but the properties are wrong for a sovereignty or diversification sleeve. Twenty-one validators, closed-source core, and — the decisive part — a native bridge operated by that same permissioned validator set. Bridge risk and chain risk are not independent, so there is no diversification benefit to be had between them. If a threshold of validators were compromised or coerced, assets bridged in are at risk with no separate recourse. Given the April 2026 lesson that the bridge IS the security model for any wrapped asset, a chain whose bridge is its validator set concentrates exactly the risk we now know to look for. Disposes of kinetiq-khype (~$0.78B) and hyperlend (~$0.43B).", "killCriteria": [ "n/a — rejected" ], "url": null }, { "id": "Avalanche", "displayName": "Avalanche", "reviewed": "2026-08-01", "verdict": "under-review", "sovereignty": "crypto-backed", "stage": "n/a", "sequencer": "Permissionless validator set, subnet architecture", "escapeHatch": "n/a — L1", "upgradeControl": "No admin key over user funds on the primary network.", "liveness": "Ranks in the more decentralised tier of proof-of-stake networks on Nakamoto coefficient. NOTE: the comparative rankings available were sourced from social posts and aggregator dashboards using inconsistent methodologies — one set scores Ethereum at 2 and Base at 1, which is measuring sequencer or client concentration rather than validator stake. Treated as directional only, not as a number to act on.", "thesis": "Plausibly approvable and not yet properly examined. Blocks benqi-staked-avax. Approval requires the subnet architecture to be understood specifically — assets on a subnet do not necessarily inherit primary-network security, and that distinction is the whole question for a lending or staking venue.", "killCriteria": [ "n/a — pending review" ], "url": null }, { "id": "Sui", "displayName": "Sui", "reviewed": "2026-08-01", "verdict": "under-review", "sovereignty": "crypto-backed", "stage": "n/a", "sequencer": "Permissionless validator set", "escapeHatch": "n/a — L1", "upgradeControl": "No admin key over user funds.", "liveness": "Ranks mid-tier on decentralisation metrics; same methodology caveat as Avalanche.", "thesis": "Not yet researched to our standard. Blocks navi-lending. A young chain with a novel object model and its own consensus — the time-in-market argument that holds Jupiter Lend applies here with more force, since a chain failure has no venue-level mitigation.", "killCriteria": [ "n/a — pending review" ], "url": null }, { "id": "Stellar", "displayName": "Stellar", "reviewed": "2026-08-01", "verdict": "under-review", "sovereignty": "crypto-backed", "stage": "n/a", "sequencer": "Federated Byzantine Agreement — quorum slices, not stake-weighted", "escapeHatch": "n/a — L1", "upgradeControl": "Protocol upgrades via validator consensus.", "liveness": "Long-running network. Not researched to our standard.", "thesis": "Not yet researched. Blocks blend-pools-v2. Stellar uses Federated Byzantine Agreement rather than stake-weighted consensus, so the Nakamoto-coefficient frame we use elsewhere does not transfer and a different analysis is required. Also note Stellar assets commonly carry issuer-level freeze and clawback flags at the protocol level, which is directly material to the sovereignty question.", "killCriteria": [ "n/a — pending review" ], "url": null }, { "id": "Cronos", "displayName": "Cronos", "reviewed": "2026-08-01", "verdict": "rejected", "sovereignty": "issuer-controlled", "stage": "n/a", "sequencer": "Small validator set closely associated with Crypto.com", "escapeHatch": "n/a", "upgradeControl": "Effectively controlled by a single corporate ecosystem.", "liveness": "Not the binding concern.", "thesis": "REJECTED on the same reasoning as BSC. A chain whose validator set and direction are governed by one exchange company offers none of the properties this sleeve exists to provide. Disposes of tectonic. Rejecting on structure rather than incident: there is no exploit to point at, and none is needed — a single compellable corporate operator is disqualifying on its own.", "killCriteria": [ "n/a — rejected" ], "url": null }, { "id": "Mantle", "displayName": "Mantle", "reviewed": "2026-07-31", "verdict": "rejected", "sovereignty": "issuer-controlled", "stage": "stage-0", "sequencer": "Centralised, as at every major L2.", "escapeHatch": "Undermined by instant upgrade authority — see below.", "upgradeControl": "The Mantle team can push INSTANT upgrades to the chain. This is the specific reason L2BEAT withholds Stage 1.", "liveness": "Not the binding concern.", "thesis": "REJECTED, and it is the only L2 here that fails on a bright line rather than a judgement call. L2BEAT rates Mantle STAGE 0 — the most cautious tier — because the team can push instant upgrades. Base and Arbitrum are Stage 1 with roughly 7-day upgrade timelocks, which at least gives a client a window to exit a rule change. Instant upgrade authority removes that window entirely, and an escape hatch that can be upgraded away is not an escape hatch. Second, compounding concern: Mantle uses EigenDA for data availability, a network secured by Ethereum RESTAKERS. We rejected restaking exposure directly at ether.fi; inheriting it underneath the data availability layer of a chain is the same risk arriving somewhere harder to see.", "killCriteria": [ "n/a — rejected; revisit only on Stage 1 promotion" ], "url": null }, { "id": "Opbnb", "displayName": "opBNB", "reviewed": "2026-07-31", "verdict": "rejected", "sovereignty": "issuer-controlled", "stage": "stage-0", "sequencer": "Centralised, Binance-operated", "escapeHatch": "Inherits BSC settlement, which we reject.", "upgradeControl": "Binance-controlled.", "liveness": "n/a for our purposes.", "thesis": "REJECTED. A rollup settling to BSC inherits every property that caused us to reject BSC, and adds a centralised sequencer on top. Listed explicitly so it cannot slip through as an unrecognised chain.", "killCriteria": [ "n/a — rejected" ], "url": null }, { "id": "BSC", "displayName": "BNB Smart Chain", "reviewed": "2026-07-31", "verdict": "rejected", "sovereignty": "issuer-controlled", "stage": "n/a", "sequencer": "Small permissioned-in-practice validator set aligned with Binance", "escapeHatch": "n/a", "upgradeControl": "Effectively controlled by a single corporate ecosystem.", "liveness": "Has halted the chain deliberately in response to an exploit.", "thesis": "REJECTED as a settlement venue. A chain whose validator set is concentrated around one company, and which has halted itself by decision, offers none of the properties this sleeve exists to provide. Relevant because ~$2.9B of USYC sits here — a tokenised-income client could be routed onto BSC by following the liquidity, and should not be.", "killCriteria": [ "n/a — rejected" ], "url": null } ] }