RUGSHIELD
Token Safety · Solana

Bundled launches on Solana: how to spot wallet clusters before you buy

RUGSHIELD Blog · 2026-11-12 · ~7 min read

A token launches on pump.fun, the LP is burned, mint authority revoked, top holder shows just 2.4%. Looks clean. Ninety minutes later the chart is a cliff. What happened? The deployer didn't hold the supply in one wallet — they bundled it: buying most of the supply at launch across dozens of fresh wallets, each small enough to stay under every concentration alert. Bundling is the standard way modern Solana rugs hide insider supply, and it's the single most common reason "all checks green" tokens still dump.

1What a bundle actually is

In the same transaction — or the same block — that creates the token and its liquidity pool, the deployer uses a bundler (Jito bundles, launchpad "bundle buy" features, or custom scripts) to fund N fresh wallets and have each of them buy a slice of the supply. Typical setups run 20–80 wallets. Individually, each wallet holds 0.5–2%: invisible to a naive top-10 check. Collectively, one entity controls 60–90% of the supply and can exit into any pump the community creates.

Why it matters: holder-concentration checks that read only the top-10 list are structurally blind to bundles. You must check relationships between wallets, not just balances.

2The fingerprints bundles leave on-chain

Bundles are hidden from casual inspection but not from the ledger. Four fingerprints give them away:

Common funder. Every bundle wallet needs SOL for fees and its initial buy. Tracing each holder's first inbound transfer usually collapses the cluster: 40 "independent" wallets funded by one address, often within minutes of each other, and often from an intermediary wallet funded by the deployer.

Same-block or same-slot funding. Organic holders arrive over days. Bundle wallets are born together. If 30 of the top 100 holders received their first SOL in the same block as the token launch, that's not a community — that's a script.

Identical buy sizes and timing. Bundlers slice the buy evenly. When dozens of wallets buy near-identical amounts (e.g. 1.85–1.92 SOL each) in the launch block, you're looking at one actor's allocation table.

Synchronized selling. The exit is the tell. Cluster wallets sell in tight sequences — the same wallet ordering logic that bought them in — often routing through the same intermediate wallet back to the deployer or a CEX deposit address.

3The 5-step cluster check (manual workflow)

You can do a credible bundle check with just an explorer and fifteen minutes:

StepWhat to doWhat a bundle looks like
1. Pull holdersList the top 50–100 holders, excluding LP and burn addresses.Many wallets with nearly identical balances and no history before launch.
2. Trace fundersFor each top holder, open the wallet and find its first inbound SOL transfer.10+ holders share one funder, or funders themselves share one funder (a tree).
3. Check timingCompare the funding timestamps and first-buy slots across holders.Wallets funded and bought within the same block or a tight slot range.
4. Sum the clusterAdd the balances of every wallet linked to a common funder.Cluster controls >20% of supply — one actor, by any reasonable definition.
5. Watch the exitsCheck whether any cluster wallets have already sold, and where proceeds went.Sequential sells consolidating into one wallet or a known exchange deposit.

4Reading the verdict

Not every shared funder is malicious — friends fund friends, and some trading tools route through shared fee wallets. Context matters:

SignalBundled rugBenign explanation
Cluster size20%+ of supply across linked wallets2–3 wallets, small combined share
Wallet ageAll created at launch, no other activityWallets have months of unrelated history
Buy patternUniform sizes in the launch blockVaried sizes spread over hours/days
Sell patternSequential sells, proceeds consolidateIndependent timing, varied destinations
Rule of thumb: any single linked cluster above ~15–20% of supply is a hard pass. The LP can be burned and the mint revoked — none of it matters if one actor holds a fifth of the float.

5Why this is hard to do by hand at speed

The workflow above works, but it doesn't scale: funding trees go two or three levels deep, bundlers rotate through fresh funder wallets, and a hyped launch gives you minutes, not hours. This is exactly the class of check an automated scanner does better than a human — graph the funders, cluster the wallets, sum the linked supply, and flag the launch-block buys in one pass.

Whatever tooling you use, make cluster analysis a non-negotiable step alongside the authority and LP checks. A token with revoked mint, burned LP and a 70% bundled supply is not a safe token — it's a rug with good paperwork.

RUGSHIELD runs cluster detection automatically: it traces funder trees, groups linked wallets, and reports bundled supply as a single number — alongside authority, LP and sell-simulation checks. Free tier credits, no card required.

Get free scan credits →