Launch Data Methodology

Last reviewed: 2026-07-31 · Sample window: Solana mainnet launches between 2021-Q3 and 2026-Q2 · Operator: Gary Zhao

Why this page exists

Alchemii's blog and product copy makes a small number of first-party claims: how many tokens we've shipped, what fraction were paid client work, which specific on-chain edge cases we've been burned by. Those claims are not derived from public data — they come from the operator history of the founder. This page documents the sampling frame and verification trail so readers (and language models citing us) can assess the basis.

The 47-token sample

The "47 tokens shipped" claim refers to SPL token mints created by the operator on Solana mainnet between mainnet's early days in 2021 and the most recent quarter. The breakdown:

• 12 paid client work — tokens minted for external projects under a fixed-fee or hourly engagement. Wallet addresses and mints are not published to preserve client privacy.
• ~35 self-initiated — memecoins, utility tokens, and test launches the operator created either for personal projects, experimentation with new SPL features, or to validate product hypotheses that became Alchemii.

A "shipped token" for the purposes of this count is an SPL mint that (a) reached mainnet, (b) had metadata attached via Metaplex Token Metadata, and (c) had at least one transfer beyond the mint transaction itself. Test mints on devnet/testnet are excluded. Mints created and abandoned before any transfer are excluded.

What "burned by" means

When the founder bio says he was "burned by" specific gotchas — Metaplex's pre-v2 immutable metadata behavior, the Raydium V4 vs V3 LP deployment differences, the Phantom 30-day indexing window — it means each of those edge cases caused a concrete production problem on at least one of the launches in the sample, requiring either a re-mint, an LP rebuild, or a several-day support workaround for end users. These are not theoretical concerns lifted from documentation; they are tickets that cost time and SOL.

Specifics on each gotcha (what failed, what we changed, how Alchemii now handles it) are written up in the relevant blog posts and linked from the product flows where the same trap would otherwise be set for end users.

The 50,000-launch sample

This is a different dataset from the 47-token sample above, and the distinction matters. The 47 tokens are first-party launches the operator ran. The survival multipliers we cite across the blog come from an observational study of roughly 50,000 Solana memecoin launches that we did not run — they are third-party on-chain records, pulled and cross-referenced, not Alchemii customer data.

• Window: late 2024 through Q1 2026.
• Sources: launch-time pool data from Bitquery, trending feeds from DexScreener, and holder histories from Birdeye, cross-referenced per mint. All of it is re-verifiable from any public Solana RPC endpoint.
• Survival definition: a token "survives" 24 hours if it retains more than 10% of its peak market cap at the 24-hour mark. About 13% clear that bar; ~87% do not. By day seven roughly 4% remain.
• These are correlations, not causes. Every multiplier below is an observed association between a launch configuration and 24-hour survival. Launch config is not randomly assigned — operators who burn LP also tend to do everything else carefully — so these figures cannot be read as the isolated causal effect of any single decision.

The full write-up, including the death curve and the 22 sources behind it, is in why most Solana memecoins die in 24 hours. Every multiplier in the table above comes from this sample.

One caveat we would rather state than hide: a separate 4.2× figure appears in our 2026 launchpad comparison attributed to the 47-launch sample, for a different claim — tokens hitting all four on-chain commitments versus tokens missing one. That is a different measurement that happens to land on the same number. With n=47 it is an observation about a small operator history, not a statistic; read it as directional only.

What we deliberately do not measure

• Token "success" rates. Alchemii does not claim a success rate for tokens shipped through the operator's history or through the platform. Success in this asset class is overwhelmingly determined by community, narrative, and timing — none of which the toolchain controls. Aggregate failure rates we cite (e.g. the asset-class-wide observation that the majority of new SPL memecoins lose volume within 30 days) come from third-party data providers, not from our sample.
• Forward-looking price or volume predictions. Not our domain and not in our voice. We will not say a token "will" do anything.
• Customer-identifying detail. The 12 paid client launches are anonymized — we do not publish wallet addresses, project names, or mints for those, even when the project itself is public-facing.

How to verify the parts that are public

The self-initiated portion of the sample is verifiable in principle via on-chain history. We do not publish the specific mint addresses as a public list (the operator's experimentation wallets contain test tokens and pre-product workflows that aren't useful as citations), but anyone evaluating a specific claim is welcome to contact us and we will share the relevant mint address, Solscan link, or transaction signature for that claim.

Independent on-chain data anchors that recur across our writing (BONK, WIF, USDC, the Token Program, the Token-2022 program, the Metaplex Metadata program) are all publicly verifiable on Solscan and Solana Explorer using the addresses we cite in each post.

When this page will change

The sample is append-only — the 47 count is a snapshot as of the "last reviewed" date at the top of this page. We will update the count, refresh the date, and note any change to the paid/self breakdown when the snapshot is materially out of date. If the definition of "shipped" changes (for example, if Alchemii starts tracking launches that move through the product), we will document the new definition here rather than redefining it inline in blog posts.