The coin passes every check. Sell simulation fine, 0% tax, ownership renounced, LP locked. The holder list looks healthy, no wallet above 3%. Then the price grinds down for an hour while every buy gets sold into. That's what a bundle looks like from the buyer's side, and none of the usual safety checks are built to see it.
It's one operator buying a large chunk of a new token at the very first price, split across many wallets, in the same block the pool opens. The point is to own a lot of supply cheaply without it showing up as one big holder.
How the buys get in depends on the chain:
A bundle is not the same as sniper bots. Snipers are strangers racing each other. A bundle is the team, and it knows when the selling starts.
Because holder lists count addresses, not people. A bundle that owns 43% of supply across 15 wallets shows up as 15 holders with about 3% each, which looks better than a coin where one wallet holds 10%.
It also dodges the limits meant to stop this. A max-wallet cap of 3% stops one address from buying more, not one person with 15 addresses. Holder-count stats get the same boost: 15 bundle wallets plus 15 more they forward tokens to adds 30 "holders" before anyone real has bought.
The tell people miss: bundle wallets often send their tokens on to brand-new addresses after the buy. Those new wallets never touched the pool, so a check of "who bought in the launch block" misses them unless you follow the transfers.
A lot, for very little. Here is a made-up BNB launch in a standard constant-product pool, ignoring fees.
Compare that with the sniper example on our sniper page: three bots took 37.5% for 3 BNB. A bundle takes a similar share, but one person controls all of it.
Usually slowly, into buyers. Dumping 429M tokens in one go would cut the price by two-thirds or more in one red candle and scare everyone off. Selling in pieces keeps the chart looking alive while the bundle cashes out.
Patterns that show up on bundled coins:
Back to the example. Buyers put in 5 BNB over the next hour and get about 238M tokens. Then the bundle sells all 429M. It takes out about 6.75 BNB for its 3 BNB, a profit of about 3.75 BNB. If the buyers then sell their 238M tokens, they get back about 1.25 BNB for their 5. That's a 75% loss, and it equals the bundle's profit.
The LP stayed locked the whole time and nobody touched the contract. A bundle takes its profit through the pool like any other seller, which is why locked liquidity doesn't protect you from it.
Four steps on the block explorer. The aim is one number: how much of the pool's token balance the launch wallets still control.
My line: if a linked group still holds more than a fifth of what's in the pool, I treat it as one whale who knows the plan, and I pass. Our token safety checklist puts this under the holder-spread step.
No, but it's rarely in your favour. Some teams buy their own supply at launch to fund marketing or to stop snipers getting it first, and a few say so publicly with the wallet addresses. A disclosed team allocation you can watch is a known risk.
The problem is the hidden kind. Splitting a buy across 15 fresh wallets has one purpose: making the team's holdings look smaller than they are. If the bundle has already sold out, the risk has mostly passed. If it's still holding, you're buying the supply it plans to sell to you.
On coins it calls, MoonRadar clusters the launch cohort. If one operator controls a large share of supply at launch, the coin is hidden from the radar instead of called. Same-block and same-size buys are grouped and labelled, so fifteen wallets from one bundle count as one signal, not fifteen.
Paste any contract into the free scanner and it also covers the checks a bundle usually sits next to: a sell check (its own on-chain buy-then-sell test on Robinhood Chain and ARC, GoPlus's simulation on BNB), buy and sell tax, whether ownership is really renounced, and a one-line Safety Grade. It ranks every wallet that traded the token by profit, which gives you a short list of wallets to run through the funding check above.
No clustering catches every bundle, and a coin you paste yourself may never have been through the radar, so do the four explorer steps on anything you're about to size into. Memecoins are high risk and most fail. These checks improve your odds, they don't remove the risk.
Scan a token on MoonRadar