What Is a Bundled Token? How to Spot One Before You Buy

If you've looked at a new token and seen a huge green candle in the very first second of its life, you've seen a bundle.

The term gets thrown around as a synonym for scam, which isn't quite right and misses what actually matters. Here's the real version.

What it originally meant

Bundling is putting several transactions into one atomic package so they execute together, in the same block, in an order the sender controls.

For a token launch that means a deployer can create the token and buy it from multiple wallets in the same transaction. Nobody else gets a look in, because there's no gap between the token existing and the supply being taken.

The original use for this was defensive, which surprises people. Teams didn't want sniper bots grabbing a destructive share of supply in the first block, so they took that supply themselves before the bots could. Not admirable exactly, but the logic held.

It's also worth knowing that a small bundle is built into pump.fun by design. When you create a token there, you're prompted for how much SOL you want to buy — and that dev buy is bundled with the creation transaction. So "this token has a bundle" is technically true of nearly every pump.fun launch. Size is what matters, not existence.

What it means now

The current version is more deliberate. A script generates twenty to thirty fresh wallets, funds each with a small amount of SOL, and fires simultaneous buys spread along the bonding curve. By the time the token is halfway to graduating, one person can control most of the supply across wallets that look entirely unrelated.

The tooling has got sophisticated about looking organic, too. Bundler software now generates plausible-looking wallet profiles — each holding a scattering of unrelated tokens so they don't read as fresh — and is explicitly marketed as defeating cluster-visualisation tools. That's the state of play: the detection tools and the evasion tools are in an active arms race, and the evasion side is well funded because it's directly profitable.

Why it matters to you

Two reasons, and the second is the one people underweight.

Cost basis. Because of how a bonding curve prices tokens, buying in the first block means buying at the lowest price the token will ever have. Bundled wallets aren't slightly ahead of you, they're structurally ahead — often by an order of magnitude.

Supply overhang. Whoever holds that supply needs somewhere to sell it. If one entity controls 40% of a token across thirty wallets, every buyer after them is potential exit liquidity, and the chart can look completely healthy right up until it isn't.

How to check

Several tools do this now. Trench Radar pulls the flagged launch transactions into a readable format and — usefully — checks whether the bundled wallets are still holding. pump.fun has its own bundle detector built in. Birdeye and various terminals surface it too, and Bubblemaps is good for seeing wallet clusters and funding relationships visually.

What you want from any of them is three numbers: how many wallets bought at launch, what share of supply they took, and whether they've sold.

Why two tools give you different answers

This is the part that catches people out, and it's not a bug in either tool.

Run the same token through two bundle checkers and you'll often get wildly different percentages. The disagreement is almost always about scope, not arithmetic.

Some tools examine only the transaction that created the token. But on several launchpads, part of the bundle settles in separate transactions within the same block — and those get missed completely, so the reported figure comes back far too low.

Others count every account whose balance increased, which includes the bonding curve itself. The curve holds nearly the entire supply at launch, so this produces figures above 100% and makes ordinary launches look catastrophic.

The correct approach is reading the complete block and filtering out program-owned accounts. Most tools don't tell you which method they're using.

Practically: if a result matters, check it in two places and look at which wallets each tool is counting rather than trusting the headline percentage. A checker that shows you the wallet table is more useful than one that just gives you a verdict. This is the same lesson as the pre-bond data problem in wallet PnL — the tools are reading a subset of reality and rarely say which subset.

Held versus sold — the number that actually matters

Most coverage stops at "is it bundled." The more useful question is what the bundle did next.

Still holding. The supply is sitting there. It's a pending dump with no timetable, and it will eventually be sold to someone. The larger the share and the longer it's held, the worse the overhang.

Already sold. The bundle distributed into early buyers. Those people made their money — from the early buyers — but the overhang is gone and the supply is now spread across real holders. Counterintuitively that can be a healthier structure than a held bundle, even though someone was clearly farmed to get there.

Partially sold. Most common. Worth looking at whether they're distributing steadily or waiting for a bigger exit.

A 30% bundle that has fully exited is a different situation from a 30% bundle that hasn't moved. Treating them the same is the most common mistake in reading this data.

What to do with it

Don't treat any bundle as disqualifying. A small dev buy is normal. Some defensive bundling is reasonable. Blanket-avoiding anything with a bundle means avoiding nearly every launch.

Do treat large held bundles as a hard skip. If a handful of connected wallets hold a substantial share and haven't sold, you're volunteering to be their exit.

Check funding sources, not just holder counts. Thirty wallets that all trace back to the same funder are one wallet wearing a costume. This is where cluster visualisation earns its place — though bear in mind the bundler tools are specifically built to defeat it.

Combine it with the rest of the checks. Bundling is one input among several. Supply concentration, liquidity status and whether the sell tax can be changed all matter alongside it — the full set is in the rug pull checklist.

And accept the limits. Detection is imperfect and getting harder as evasion improves. A clean bundle check means the obvious version wasn't present. It doesn't mean nobody controls the supply.

Trading with it in mind

If you're sniping launches, you're operating in the same block as the bundlers, and no amount of checking helps at that speed — position sizing does the work instead.

Axxel supports sniper, market, limit and trailing orders across Ethereum, Base, BSC, Solana and Robinhood Chain, with configurable slippage and maximum buy and sell tax limits so orders won't fire on tokens that breach your thresholds. Flat 0.9%, non-custodial, no subscription.

The checks are still yours to run. Nothing automates judgement about whether a bundle is defensive or predatory — that's a read, and it's usually the difference between a good entry and being the reason someone else had one.

Crypto trading carries risk. Most memecoins lose value. Nothing here is financial advice. Axxel is not available in all regions.