A quarter of pump.fun graduates are bought whole by their own launcher, in the second they launch

What happens in that second

One coin, read from the chain: 12c3QbnTGMWph6wbafsYrpkhXEiWDxRKQsTMjtakpump, created on 2026-10-21 in slot 459641324.

1. The create transaction (35AssbvJ3CAzju7Ak8NXAoQL9EohjSMty182pBedZTR5hZGL5zucuPmvf9inxdvNMABkQFhvB6KCazNRb7h7bdu2) runs CreateV2 and then Buy. The wallet that signed it, 3Dpbh8h33A3hHBwJ5FtnUpxBHLwXEPXXd5QFbKH2Hy7M, pays 86.08 SOL and receives 793.1M tokens: 79.31% of the supply, the whole bonding curve. The curve takes in 85.0 SOL.
2. In the same slot, pump.fun's migration (5jeAs4J7Dt3YzwYdEjsCQ7X1bmDRdybYnJSgMjWSqtg71Xt8uB9mQXneY3p5dTi6UyYoKLvWSUC3ABazmofcVZMh) runs MigrateV2, CreatePool and InitBoost. The pool opens with the other 206.9M tokens (20.69% of the supply) and 67.4 SOL; 17.6 SOL goes to the coin's boost vault, which pump.fun spends over the next minutes buying and burning the coin.
3. The pool's opening price, times the whole supply, is a market cap of 325.8 SOL. The launcher's 793.1M tokens are worth 258.4 SOL at that price. It paid 86.08 SOL for them.

Nobody else has bought yet. Every buyer from here on buys from the pool's 20.69%, and the price they pay is the price the launcher's bag is marked at.

It began in June, and BOOST made it the norm

The share of each month's graduates that were bought whole at launch and migrated within two seconds:

Two things changed. The first is June: the count went from a couple a day to 15, which is tooling, someone shipping a launcher that buys the curve in the create transaction. The second is BOOST: pump.fun's mode, on every new coin from July 20, that moves about a fifth of the migration's SOL into a vault and spends it buying the coin back over its first minutes. A launcher that buys the whole curve now gets a guaranteed buyer of its own coin in the pool it just opened. Within a month the pattern was a fifth of graduates, and it has held near a third since.

Who does it

22,249 distinct wallets launched these 24,232 coins since BOOST, and 21,298 of them did it exactly once. The busiest wallet did it 78 times; 27 wallets did it ten or more times, 597 coins between them; the ten busiest account for 1.5% of the total.

This is not a ring of operators. It is a product. Someone sells, or gives away, a launcher that creates a coin, buys its curve and lets BOOST open the pool, and thousands of people have used it once each, a fresh wallet every time. The deployer's record, the first question our check asks about any coin, is nearly always one launch, one graduation.

What the market cap means on these coins

A market cap is the last price times the whole supply. On an ordinary graduate the supply was bought along the curve by hundreds of wallets at rising prices; the market cap is a rough sum of what people paid. On a coin bought whole at launch, 79.31% of the supply was bought by one wallet at the curve's prices, 86.08 SOL in the example, and is then marked at whatever the pool's 20.69% trades at.

The pool can be real. Doroni and Kimchi, two of the week's most-asked-about coins, are both this kind, and both have thousands of SOL of real liquidity from real buyers. What the market cap cannot tell you is what the launcher's bag would fetch. Selling 79.31% of a supply into a pool that holds 20.69% of it cannot return anything like the mark; it returns the SOL the buyers put in, less whatever is left behind. The launcher's profit is capped by the pool's SOL, and so is every other holder's exit.

pump.fun coins carry no freeze and no sell tax: the launcher can sell into the pool at any moment, and so can you. A small test buy and sell on Doroni went through both ways.

A buyer an hour in

For every graduate since BOOST that had migrated within its first hour, we take the price at one hour and follow it 24 hours. The column that matters is the last one.

The doubles are the same. Everything else is split. A coin bought whole at launch either sits where it was (42.9% closed at or above the 1 h price, and 17.3 of those points are coins with no trade at all in the day, five times the share of other graduates) or it is crushed: 34.6% closed below a tenth of the 1 h price, which is the launcher's 79.31% going into the pool. Almost none of them lose a little. Other graduates spread their losses out; these go to zero or nowhere.

Where they stand a few days later

From the live recorder, for the graduates of 2026-10-20 to 2026-10-23, as of 2026-10-23 23:59 UTC: 985 were bought whole at launch, 26.9% of the graduates in those days.

Two readings from one table. Most of these pools are empty within days (87.1% hold under 1 SOL), and in 71.4% of them the launcher still holds its 79.31%: a market cap on a pool with no SOL in it. But the ones that found a crowd found a bigger one: 6.7% hold 100 SOL or more, five times the share of other graduates. When this launch style works, it is the biggest coin of the day; when it does not, the launcher is still sitting on four fifths of a coin nobody bought.

In 28.0% of them the launching wallet holds nothing: it sold. That is the same event as the 34.6% in the table above, seen from the launcher's side.

What we counted

• Graduates: launches in our tables (every pump.fun trade, read from the chain) that migrated to a pool, by migration day, 2026-01-01 to 2026-10-19; the tables are built to 2026-10-20.
• Bought whole at launch: the wallets that bought in the first ten seconds (one or two of them) took 79.31% or more of the supply, and the migration is within two seconds of the create. The curve holds 793.1M of the billion tokens; the pool gets the rest.
• A buyer an hour in: graduates that migrated within an hour of the create and had a price at one hour; the peak and close over the next 24 hours as multiples of that price, from every trade on the curve or in the coin's own SOL pool. A coin with no trade in those 24 hours is counted as closing where it stood.
• A few days later: the live recorder's fill, graduates created from 2026-10-20 11:37 UTC (where the tables end) to 2026-10-23 23:59 UTC; the pool's real SOL after its last trade; the launching wallets' share of the supply net of their sells.
• The example coin was read from Solana mainnet on 2026-10-23: the two oldest signatures on the mint and their parsed transactions.
• The month, deployer and one-hour figures are from scripts/research/boost_bundles.py in our repository; every number on this page is checked against its output before the page is published.

Frequently asked questions

What does "bought the whole curve" mean?

A pump.fun coin has a billion tokens. 793.1M of them sit on the bonding curve; when they have all been bought the coin graduates and the remaining 206.9M open a pool with the curve's SOL. On these coins one wallet buys all 793.1M in the create transaction itself, so the coin graduates before anyone else can see it.

Is the market cap fake?

It is computed the same way as any other coin's: the pool's last price times the supply. What differs is who holds the supply. 79.31% of it was bought by the launcher at curve prices and is marked at the pool's price. The pool's SOL, which our check page now shows beside the market cap, is what can actually be taken out.

Can the launcher dump on me?

Yes, into the pool, at any time; pump.fun coins have no freeze and no sell tax. In 34.6% of these coins the price an hour in had fallen below a tenth of itself a day later, and in 28.0% of the recent ones the launching wallet already holds nothing.

Does this mean the coin is a scam?

No. It means the coin's supply is concentrated and its market cap says less than usual. Some of the biggest coins of the week were launched this way and have deep pools. Read the pool's SOL, and whether the launcher still holds its bag, before reading the market cap.

How do I check a coin?

GET https://pumpdata.fun/free/launch/{mint}: bundle_at_creation gives the first ten seconds' buyers and their share; migrated_at against created gives the gap. The check page, https://pumpdata.fun/check/{mint}, says it in a sentence when it applies, with the pool's SOL beside the market cap and the biggest holders from the chain.

Related: The pump.fun API · How to backtest a pump.fun strategy · The pre-buy check for a pump.fun trading agent · The pump.fun dataset · Free pump.fun coin check · all guides