Chains With Most New Token Launches

Define a launch as the earliest confirmed block or slot in which a newly created fungible token has an enabled permissionless pool and can complete a nonprivileged swap. A pool creation transaction without active liquidity does not qualify. A transfer-only token does not qualify. The numerator for a chain is the number of unique canonical token contracts or mints first launched there during the window. The denominator for chain share is the sum of eligible unique launches observed on every chain included in the comparison.

Choose complete UTC calendar periods, such as Monday 00:00 through Sunday 23:59:59, and freeze results only after a declared indexing delay. Assign boundary events using block time, while storing provider discovery time separately. For each token, preserve contract creation, pool creation, first liquidity, and first successful swap when available. This prevents a provider's indexing speed from becoming the launch timestamp.

Build candidate sets from supported DEX factory events, pool program instructions, and provider new-token feeds. CoinGecko and Birdeye can broaden discovery and enrich identifiers, while Uniswap and Raydium documentation helps interpret factory, pool, and program events. Union the candidate sets before validation. Provider presence is not itself proof of launch, and absence from one provider is not proof that no launch occurred.

Deduplicate by canonical chain and contract or mint. Multiple pools, quote assets, or DEX venues count once. Proxy upgrades that retain the same address remain one token. Bridged representations are excluded from the primary new-launch count unless the research question explicitly measures new local representations, in which case they must be a separate category. Relaunches after abandoned liquidity do not become new tokens. Symbol and name are never sufficient identifiers.

Validate first tradability by locating a successful swap or by reconstructing pool state sufficient to show that a permissionless swap was possible. Save the transaction hash, pool address, block or slot, and parser version. When evidence is ambiguous, mark the candidate unresolved and include it in a published unresolved count, but exclude it from the ranked numerator until resolved.

Exclude NFTs, liquidity position tokens, wrapped native assets, stablecoin migrations, bridged copies under the primary definition, testnet activity, known test contracts, and pools restricted to allowlisted traders. Also exclude contracts created before the observation window that merely receive an additional pool during it. Launchpads may create intermediate bonding curve markets, so state whether permissionless curve trading qualifies or whether graduation to an external AMM is required. The same rule must apply across chains.

Report each chain's observed count and observed share, plus the supported DEX venues and launchpad programs. Avoid normalizing by chain transaction count unless that separate denominator is complete and consistently defined. Raw launch counts answer activity volume, while launches per active address or per transaction answer different questions and should not be blended into one ranking.

Every release needs the observation window, data freeze time, latest indexed block or slot by chain, retrieval timestamp by provider, supported venue list, and code or query version. Coinatio's live table supplies current examples and can move as late events arrive. A frozen research export should retain both initial and revised counts so backfills are visible.

Coverage uncertainty is structural. EVM event indexing and Solana program parsing expose different failure modes. Birdeye and CoinGecko do not cover every chain or venue equally, and RPC history can be pruned or delayed. Estimate completeness by comparing independent discovery channels and publish unmatched candidate counts, but do not convert that comparison into an unsupported global completeness percentage. Chainalysis research is useful context for cross-chain activity and illicit ecosystem patterns, not a substitute for launch-level event data. A chain ranked lower under partial venue coverage may simply be less observable.