Bundled pump.fun launches graduate ten times more often. Then they fall.
15.4 million launches sorted by how many wallets bought in the first ten seconds. The bundled ones graduate at 7.32%; the unbundled at 0.65%. A day later, the order flips.
Quick answer
A pump.fun launch where 31 or more wallets bought in the first ten seconds graduated 7.32% of the time; a launch with no buyer in its first ten seconds graduated 0.65% of the time. Graduation tracks the bundle almost perfectly, because the wallets that bundle a launch are the ones that buy it to the end of the curve. The day after tells the other half: of the 31+ bundles that were still alive at five minutes, 31.3% closed the next 24 hours at or above their five-minute market cap, against 76.7% for launches with a single early buyer. Bundles graduate coins and then sell them.
How many wallets buy in the first ten seconds
For every launch in the tables we count the distinct wallets that bought in the ten seconds after the create, and the share of the supply those buys took. The median launch has two or three.
| buyers in the first 10 s | launches | share of launches | graduated | graduation rate | avg. supply the bundle took |
|---|---|---|---|---|---|
| 0 | 1,048,008 | 6.8% | 6,762 | 0.65% | 0.0% |
| 1 | 2,855,563 | 18.6% | 26,620 | 0.93% | 2.9% |
| 2-3 | 4,438,107 | 28.8% | 36,885 | 0.83% | 6.2% |
| 4-10 | 4,432,086 | 28.8% | 53,376 | 1.20% | 14.6% |
| 11-30 | 2,024,770 | 13.2% | 37,494 | 1.85% | 34.7% |
| 31+ | 593,493 | 3.9% | 43,469 | 7.32% | 65.2% |
The 31+ group is 3.9% of launches and 21% of graduations. The record is 394 distinct wallets buying one coin inside ten seconds (GPNTYGsyCXt6xpabKR1SkqTFrNHf8WejjFgbyj25pump, which took 41.2% of the supply and did not graduate).
The supply share says it more plainly
Sort the same launches by how much of the supply the first-ten-second buyers took, and the graduation rate climbs a cliff at the halfway mark.
| supply taken in the first 10 s | launches | share | graduated | graduation rate |
|---|---|---|---|---|
| 0% | 1,307,184 | 8.5% | 6,800 | 0.52% |
| under 10% | 7,784,775 | 50.6% | 47,535 | 0.61% |
| 10-30% | 4,084,816 | 26.5% | 22,456 | 0.55% |
| 30-50% | 1,369,688 | 8.9% | 19,941 | 1.46% |
| 50-80% | 699,199 | 4.5% | 85,451 | 12.22% |
| 80% or more | 146,365 | 1.0% | 22,423 | 15.32% |
Half of all launches had under 10% of their supply bought in the first ten seconds; they graduate at 0.61%. The 5.5% of launches where the early buyers took more than half the supply produced 107,874 of the 204,606 graduations in the tables: 53% of all graduates came from the most bundled 5.5% of launches.
A bonding curve graduates when enough SOL has gone in. A group of wallets that buys 60% of the supply in ten seconds has put most of that SOL in already; the graduation is the same wallets finishing the job.
The day after
Graduating is not the same as holding. We measure every launch at five minutes and follow it for 24 hours: whether it traded at all (dead means no trade in those 24 hours), and where its market cap closed against the five-minute mark.
| buyers in the first 10 s | measured at 5 min | dead in the next 24 h | closed at or above the 5-min market cap |
|---|---|---|---|
| 0 | 69,745 | 42.5% | 61.9% |
| 1 | 2,807,846 | 66.7% | 76.7% |
| 2-3 | 3,808,795 | 48.6% | 65.1% |
| 4-10 | 4,081,831 | 37.5% | 56.5% |
| 11-30 | 2,017,428 | 30.9% | 46.4% |
| 31+ | 592,877 | 16.8% | 31.3% |
Two curves cross. The bigger the bundle, the less likely the coin is dead a day later (16.8% for 31+, 66.7% for one early buyer), and the less likely it closed at or above where it stood at five minutes (31.3% against 76.7%). The single-buyer coins that survive mostly do nothing, so they "hold" their five-minute price by default. The bundled coins trade, and trade down: the supply bought in the first ten seconds is for sale in the next 24 hours.
Read the two tables together: the bundle gets the coin to graduation more often than anything else in the data, and then the bundle gets sold.
Bundling is going out of fashion, slowly
The share of launches with four or more buyers in the first ten seconds has fallen by more than half since the start of the tables.
| month | launches | with 4+ buyers in 10 s | graduation rate |
|---|---|---|---|
| 2025-03 | 649,614 | 67.5% | 0.41% |
| 2025-06 | 771,771 | 62.9% | 0.87% |
| 2025-09 | 675,397 | 42.8% | 0.65% |
| 2025-12 | 620,682 | 37.3% | 0.71% |
| 2026-03 | 794,571 | 43.8% | 1.61% |
| 2026-06 | 852,083 | 35.8% | 0.85% |
| 2026-09 | 1,005,468 | 32.0% | 3.39% |
| 2026-10 | 745,481 | 28.5% | 2.43% |
October 2026 is through the 16th. The graduation rate went the other way over the same months, from 0.41% to over 2%, which is a story about who is launching rather than how, and is in the deployers report.
What we counted
- The tables as of 2026-10-16, built on our archive of every pump.fun trade: 15,392,027 launches from 2025-03-07 to 2026-10-16, 3,898,141 deployers, 204,606 graduations. The month table runs on calendar months (UTC); October is partial.
- Buyers in the first ten seconds: distinct wallets with a buy on the bonding curve within ten seconds of the create transaction, by block time. The deployer's own buy counts when it lands in that window.
- Supply taken: the tokens those buys received, over the total supply. A handful of launches show more than 100% because the same wallets sold and rebought inside the window; they sit in the 80% or more row.
- Graduated: the coin migrated to PumpSwap, at any later time.
- The day after: launches measured at five minutes of age (13,378,522 had a reading of $1,000 or more); dead means no priced trade, a buy or a sell on the curve or the coin's migration pool, in the 24 hours after that mark; closed at or above means the last market cap of that window was at or above the five-minute market cap. A dead coin closes where it stood.
- Graduations and bundle sizes are counted from the chain's ledger, not from an indexer. The tables are rebuilt weekly; this report is regenerated when they are.
Frequently asked questions
What is a bundle on pump.fun?
A set of wallets, usually controlled by one operator, that buy a coin in the same block or seconds after its launch, so that the operator holds a large share of the supply before anyone else can buy. We measure it as the distinct wallets buying in the first ten seconds and the share of supply they took.
Do bundled coins graduate more often?
Yes, by a wide margin: 7.32% for launches with 31 or more buyers in the first ten seconds, against 0.65% with none. Launches where the early buyers took more than half the supply graduated 12.22% to 15.32% of the time.
Does a bundle mean the coin will go up?
The data says the opposite over the first day. Of bundled launches alive at five minutes, 31.3% closed the next 24 hours at or above their five-minute market cap; for launches with one early buyer the figure is 76.7%, mostly because those coins did not trade at all.
How do I check a coin's bundle?
GET https://pumpdata.fun/launch/{mint} returns the buyers in the first ten seconds and the supply share they took, with the deployer's record and how launches that looked the same at the same age went over their next 24 hours. $0.005 a call, no account.
Every number here comes straight from the blockchain, the same data behind the paid calls: the calls ($0.005 to $0.01 each, USDC over x402, no account), yesterday's /stats and the daily wire are free.