What the pump.fun memes that worked have in common
Short answer: the memes that worked were, above all, a creator's first launch. They also had few bots among their first buyers and, more rarely, a complete set of real links. A website or an X account alone changes little. And even with every trait, most memes still fail: these are better odds, not a recipe.
Key numbers Β· 318,018 pump.fun launches Β· 11β18 October 2026
What "worked" means here
"Worked" can mean many things. We used three steps, each harder than the last, all measured on-chain:
- Graduated: the bonding curve filled and the meme moved to PumpSwap. 7,408 of 318,018 launches (2.3%).
- Held after the Boost: ten minutes after graduation, its price was still at least 90% of its price 30 seconds after graduation. 2,348 memes.
- Still standing a day later: in the Aporius village, a graduated meme becomes a stone house. 188 were still standing 24 hours or more after graduating.
The third step only covers memes whose creator did not launch another one (that is what the village follows), so it says nothing about serial creators by construction. The first two steps cover every launch.
1. A first launch, by far the strongest trait
Half of all launches (50%) come from wallets that already launched ten or more memes in the previous 24 hours. They almost never work.
- A first launch graduated 4.96% of the time, more than 4Γ the rate of a wallet with 10+ launches that day (1.16%).
- After graduating, 50% of first launches held their price ten minutes after the Boost, against 19% for those serial wallets.
- Of the 188 memes still standing a day after graduating, 181 came from a creator's first launch. Not one came from a creator with three or more launches.
2. Few bots among the first buyers
RugWatch measures how many of a meme's early buyers are bots: wallets that buy and resell within minutes, on dozens of tokens a day.
| Bots among early buyers | Launches | Graduated | Held 10 min after the Boost |
|---|---|---|---|
| under 10% | 9,951 | 5.4% | 50% |
| 10β30% | 5,800 | 4.5% | 36% |
| 30β50% | 8,015 | 5.8% | 34% |
| 50% or more | 33,317 | 3.3% | 15% |
Measured on the 57,083 launches where the share of bots could be computed. "Held" is among those that graduated.
Bots don't stop a meme from graduating as much as they stop it from holding: when bots made half the first buys, only 15% of graduated memes were still near their price ten minutes later. Bots buy to sell, and the Boost is their exit.
3. Real links, all of them
| Links declared at launch | Share of launches | Graduated | Held 10 min |
|---|---|---|---|
| no link at all | 75.8% | 2.1% | 40% |
| its own X account | 23.1% | 3.2% | 38% |
| a website that talks about the token | 6.5% | 5.0% | 37% |
| website + X + Telegram | 0.2% | 12.5% | 59% |
One link changes little: a website made in ten minutes from a template costs nothing, and many are exactly that (we are measuring how often, more soon). The complete set (its own website, X account and Telegram group) is rare, 687 launches in a week, but those graduated 12.5% of the time, five times the average, and 59% of them held after the Boost.
An honest note on our own verdicts: a first launch without verifiable links graduated slightly more often (7.9%) than a first launch with them (5.8%) in this week. The creator's record matters more than the links. That is why we are working on a stronger, harder-to-fake level than β .
4. What doesn't work
- Factories: wallets RugWatch puts in "jail" (5+ launches in a week, none graduated) were 40% of launches. They graduated 0.53% of the time, and none was still standing a day later.
- Rugs: about 1 in 10 graduations is rugged within ten minutes, usually right after the Boost stops buying. See what happens after a Boost.
- Speed alone: flash graduations (curve filled in under 90 s) held best in the first ten minutes (54%), but memes that filled their curve in 90 seconds to 10 minutes held worst (19%): many people bought them on the way up and sold into the Boost.
The profile, all traits together
A creator's first launch, with its own website and X account, and under 30% bots at launch: 1,765 launches in a week (0.6%). They graduated 10.1% of the time, more than four times the average.
That is the coherence of what worked: a creator who is not running a production line, a crowd that is not made of bots, and a project that shows its face. It also means that nine memes out of ten with that profile still did not graduate. Better odds are not a guarantee, and a good creator guarantees no price.
How we measured
- Every pump.fun launch seen live by the RugWatch tracker from 11 to 18 October 2026: 318,018, each followed for at least 24 hours.
- Creator's launches: counted in the 24 hours before each launch, from the same wallet.
- Bots: wallets that buy then resell within 15 minutes on 20 or more tokens a day (how RugWatch verdicts work).
- Held after the Boost: price ten minutes after graduation at least 90% of the price at 30 seconds, on PumpSwap prices read every 15 seconds; broken readings removed, real crashes and rugs kept.
- Observed frequencies, not predictions. Not financial advice.
See every birth live on RugWatch. The creator's record, their links and the share of bots, for every new pump.fun meme, within seconds.
Open RugWatch The villageFAQ
What do successful pump.fun memes have in common?
Measured on 318,018 launches (11 to 18 October 2026): the creator had launched nothing else in the previous 24 hours, few of the first buyers were bots, and the meme had real links (its own website, X account and Telegram). A creator's first launch graduated 4.96% of the time, against 1.16% for creators with 10 or more launches that day.
Does a website make a pump.fun meme more likely to succeed?
A little, not on its own. Memes with a website about the token graduated 4.95% of the time, against 2.07% with no link at all. Memes with a website, an X account and a Telegram group together graduated 12.5% of the time, but they were only 0.2% of launches. Many websites are templates made to look serious.
Do bots kill a pump.fun meme?
They hurt its chances to hold. Among graduated memes, 50% of those with under 10% bot buyers at launch were still at or near their price ten minutes after the Boost, against 15% when bots made half or more of the first buys.
Found this useful? Share it with someone who trades pump.fun.
Data: public Solana on-chain activity (pump.fun, PumpSwap) collected by the Aporius RugWatch tracker, 11β18 October 2026. Medians and observed frequencies, not predictions. Not financial advice.
Aporius