Why Most pump.fun Tokens Fail (and How to Spot a Rug)

MemeSwap

pump.fun changed meme coins. Anyone can create a token on a bonding curve in about a minute, and when the curve fills, the token graduates to an open pool. That openness is the point, and it is also why buyers have to do the filtering themselves. This guide covers what the public data says about how pump.fun tokens end up, the difference between a token that fades and one that is pulled out from under you, and the specific checks that catch most rugs before you sign.

This article does not claim pump.fun itself is a scam. pump.fun is a launchpad; the problems below come from what some creators and early buyers do with it, and many apply to every permissionless launchpad. They are also exactly the problems MemeSwap's launchpad was built to stop.

What the public data says

The most cited number comes from the Solidus Labs 2025 Rug Pull Report. Solidus looked at tokens launched on pump.fun from January 2024 to March 2025 and found that of more than 7 million tokens with at least five trades, only about 97,000 still had more than $1,000 of liquidity. It labelled the remaining 98.6% as showing pump-and-dump or rug characteristics, using that liquidity threshold as its measure (CoinDesk coverage). pump.fun publicly disputed the report.

It is worth being precise about what that figure measures. "Fell below $1,000 of liquidity" captures both tokens that were deliberately drained and tokens that simply never found buyers. Both outcomes cost the people holding them, but they are different problems with different warning signs.

A wider view comes from Chainalysis, which looked at all tokens launched across chains in 2024, not just pump.fun. It flagged 3.59% of launched tokens as showing patterns consistent with pump-and-dump schemes, and found that about 94% of the pools involved were rugged by the address that created the pool. In other words: when a deliberate rug happens, it usually comes from the people who set the token up.

MemeSwap's own recorder adds a third angle. On the venues it covers, the graduation study found 5.3% of a 4,376-launch cohort graduated their bonding curve. Most launches never reach an open pool at all.

Fading versus rugging

Fading is the common case. A token launches, a few people buy, interest dries up, and the price slides until nobody trades it. No one stole anything; the market just moved on. The risk is real, but it is the basic risk of buying something new.

Rugging is deliberate. Someone with control over the contract, the liquidity or a large share of the supply uses that control to exit at the buyers' expense. The warning signs for a rug are mostly visible on-chain before you buy, which is why checks work.

Five ways tokens rug

### 1. The liquidity pull

The address that created the pool removes the liquidity. Price collapses because there is nothing left to sell into. This is the classic rug, and per Chainalysis, the pool creator is the usual culprit. Check: is the liquidity locked or burned, and how deep is it?

### 2. The contract trap

The deployer keeps powers that let them change the rules after you buy: mint more supply, pause trading, blacklist wallets, upgrade the contract, or claw tokens back. On Solana, mint and freeze authority are the first things to look at.

### 3. The honeypot

You can buy, but you cannot sell, or the sell tax is set so high that selling is pointless. A changeable tax is a softer version of the same trap. MemeSwap has a full explainer on honeypot tokens.

### 4. The insider dump

A few wallets, often the creator's or wallets funded by the creator, hold most of the supply and sell into the first wave of buyers. Bundled launches, where many wallets buy in the same block as creation, can make this look like organic demand.

### 5. The fast dev exit

The creator simply sells their whole position minutes after launch. Technically this is a holder selling, not a rug, but for buyers the effect is similar. MemeSwap's recorder flagged creator sells a median 141 seconds after launch, with a quarter inside the first minute. Where the size was recorded, full exits outnumbered partial sells roughly twenty to one.

The checks that catch most rugs

Run these before you buy any new token, from any launchpad:

Sell route. Is there a working route to sell, not just to buy?
Contract powers. Mint, freeze, pause, blacklist, upgrade, take-back. Any of these left with the deployer is a question you need answered.
Taxes. A sell tax higher than the buy tax, or one the owner can change, is a classic trap.
Top holders. If the top five wallets hold most of the supply, one of them can end the market.
Liquidity. Under a few thousand dollars, you may not get out near the chart price. MemeSwap marks pages under $5,000 as thin.
Dev wallet. Has the creator already sold? All of it?
Creator history. A rug on record is the clearest red flag. An address with hundreds of prior launches is usually a minting script, according to MemeSwap's repeat creator research.

None of these makes a token safe. Together they filter out most of the traps people get caught by.

Graduation is not a safety badge

When a pump.fun token completes its curve, it graduates to an open pool. That tells you one thing: enough money went in to fill the curve. It is not an audit, not a liquidity lock, and not a promise that you can sell later. Graduated tokens still need every check above.

How MemeSwap's launchpad is built to stop this

Every rug pattern above depends on a lever someone keeps after launch. MemeSwap's launchpad takes those levers away at the source:

Liquidity pull: Liquidity locked or burned automatically
Insider dump / bundled supply: Fair launch at open; bundle and sniper wallet detection
Fast dev exit: Dev token vesting or sell limits
Anonymous scam launch: Verified creators

Launches run on Solana and Base. After launch, tokens can earn MemeSwap Verified, a live badge backed by a screen that re-checks every few minutes:

Not a honeypot, no blacklist, no take-back function
No unchecked mint, pause or upgrade powers
Top holders under the concentration ceiling
Enough liquidity and a pool old enough to judge
A deployer with no rug on record, and not a copycat
A live market and a working swap route

Buying tokens launched elsewhere? Graduated pump.fun tokens that pass the screen are on the pump.fun hub, and you can paste any Solana contract into @ava_mae_bot on Telegram.

None of this can stop a token from losing value. It is built to stop the traps.

FAQ

Are 98% of pump.fun tokens scams?

A Solidus Labs report claimed 98.6% of pump.fun tokens fell below $1,000 liquidity, which it treated as pump-and-dump or rug behaviour. pump.fun disputed it. That measure includes tokens that simply faded, so it is better read as "most tokens do not keep liquidity" than "every token is a scam."

What is the most common way a meme coin rugs?

In Chainalysis's 2024 data across chains, about 94% of pools in suspected pump-and-dump schemes were rugged by the address that created the pool.

Can a graduated pump.fun token still rug?

Yes. Graduation means the curve filled. It does not lock liquidity or remove contract powers.

How does MemeSwap stop rugs at launch?

On MemeSwap's coming-soon launchpad, liquidity is locked or burned automatically, dev tokens vest or are sell-limited, launches open fair with bundle and sniper wallet detection, and creators are verified.

How do I check a token quickly?

Open it in the MemeSwap app (https://apps.apple.com/us/app/memeswap-wallet/id6808042688) to see contract checks next to the chart, or paste the contract into @ava_mae_bot on Telegram.

Coming soon — MemeSwap launchpad. Get first access on Telegram: https://t.me/memeswapdotfun

Buying? Get the MemeSwap iPhone app: https://apps.apple.com/us/app/memeswap-wallet/id6808042688 · Telegram: https://t.me/memeswapdotfun

Meme coins are high risk and can lose all value. MemeSwap is not affiliated with pump.fun. Protections and screens are built to stop common traps, not guarantees. NFA.