On-chain measurement · Solana
The median pump.fun creator has earned at least —
Reconstructed from — creator fee vaults on Solana mainnet.
The average is real. It is also nobody's income.
Both numbers below are correct, computed over the same creators at the same moment. The gap between them is not a rounding error or a disagreement about method — it is the shape of the distribution. A mean is a total divided by a headcount, and it describes the typical participant only when the total is spread evenly. Here it is not.
Two populations, not one
Most creators in the sample hold nothing they could claim. Their coin launched, attracted almost no trading, and produced no fee worth the transaction to collect. Averaging them together with the coins that worked produces a number that describes neither group.
Where the money actually comes from
Creator fees accrue in two separate places, and they behave nothing alike. A coin that never graduates earns only on the bonding curve, where the sums are small. The money is on the AMM, and reaching it requires graduating — which most coins never do.
Measuring only the bonding curve would have undercounted every large earner while leaving the long tail untouched — inflating the inequality figure and making this site's own argument look stronger than it is. Both venues are measured. Method.
The top of the curve is real too
Nothing here says large creator payouts are fictional. They are on-chain and anyone can verify them. The claim under test is not whether the top exists — it is whether the top describes the typical creator.
Drawn from a separate census of the highest-market-cap coins, and deliberately not mixed into the sample statistics above — folding winners into a random sample is exactly how a median gets quietly inflated.