Holder analysis
The chart tells you what price did. Holder analysis tells you who's holding the bag and whether they're all the same person.
Available as the 👥 Holder Analysis button in Telegram and the holder intelligence panel (with the bubble map) in the Web Analyzer. Both run the same analysis, so the numbers are identical by construction, not by luck.
What it reports
Supply concentration
🔴 High
The top 10 hold a majority of supply, or one wallet holds an outsized share of it
🟡 Mid
Meaningful concentration, but short of that
🟢 Low
Neither
Plus the raw numbers: holder count, top 10 %, top 20 %, and the largest single wallet.
Note this is owner-level, not account-level. On Solana one person can hold a token across many token accounts; aggregating by owner is the difference between "800 holders" and "800 accounts belonging to 300 people".
Wallet clusters (bubbles)
This is the interesting part.
Ten wallets holding 3% each looks like healthy distribution. Ten wallets holding 3% each that were all funded by the same address twenty minutes before launch is one wallet holding 30% wearing a hat.
The analysis traces how each significant holder was funded and when it first became active, then groups wallets that trace back to the same origin. Each cluster is reported with its combined percentage, its wallet count, and why it formed.
Evidence strength matters. Some links are strong evidence of one operator; others are circumstantial, because plenty of legitimate infrastructure touches many wallets. Weak links are held to a higher bar before they form a cluster at all, and they can't push the risk verdict to High on their own.
Known exchange wallets are excluded outright: everyone who bought on Coinbase shares a funding source, and that isn't a bubble. Any list of known exchange wallets is best-effort, so there's a behavioural backstop as well — an address that has funded thousands of wallets is infrastructure, not a deployer. (A chain that runs its own exchange is the exception that proves the rule: on Robinhood Chain, a wallet labelled "Robinhood" is the chain's own plumbing, so that name is not treated as an exchange signal there while still counting everywhere else.)
Sharing a funder is worth less on some chains than others. Where a chain is fronted by one app, nearly every wallet on it was funded through the same bridge or on-ramp — so "these ten holders share a funder" describes the chain, not a bubble. Those groups are dropped rather than shown as weak clusters.
What survives that is how the funding landed: wallets paid out of the same address in one block, or for byte-identical amounts, are coordinated no matter how busy the funder is. When a group carries one of those fingerprints, the cluster is built from exactly the wallets that share it — so you get the three wallets that were actually batched, not the thirty that used the same bridge.
Proximity in time is deliberately not treated as evidence: a busy on-ramp funds strangers continuously, so any window wide enough to catch a careful operator also fires on unrelated wallets.
Bubble risk
🔴 High / 🟡 Mid / 🟢 Low, derived from how much supply sits inside those clusters and how strong the evidence behind them is.
The card also shows total connected supply, and separately how much of it rests on strong evidence — so you can see when a scary-looking number is mostly weak inference.
One exception: our own token. The largest wallet group on $DON is the project treasury — supply held back for scheduled burns and for future partner and KOL allocations, not a whale cluster positioned to exit. It is drawn on the map exactly like any other group, at its real size, labelled project treasury, and it is the one group left out of the bubble-risk grade. The card says so in place: the grade never appears without the line telling you how much supply was excluded from it.
Launch signals
Who was around when the pool opened:
🌱 Fresh
Wallets whose first-ever activity is around this token's launch
🎯 Snipers
Bought in the earliest transactions after the pool opened
📦 Bundle
Buys landing together — coordinated, not coincidental
👤 Insiders
Wallets funded by the token's own creator
Where the size of the launch cohort is known, snipers and bundles are shown as launch% → now%. That arrow is the whole story: 40% → 4% means they dumped; 40% → 38% means they're still sitting on it.
Wallet quality
The top holders (LP excluded) bucketed by total portfolio value — whales, mids, and everything below. A holder base made entirely of tiny wallets gets flagged for paper-hand risk: nothing there is a conviction position.
Security
Where the data is available, contract security checks are layered on — mint authority, blacklist functions, rug ratio, and the big one:
⛔ HONEYPOT — sells are blocked. Do not buy.
If that's on the card, it's the first line, and nothing else on the card matters.
Chain coverage
Solana
Full — owner aggregation, funding traces, clusters, snipers, bundles, insiders
Ethereum / Base
Full — same output, resolved through a different data path
BSC
Partial — cluster evidence is thinner than on the chains above
Robinhood Chain
Partial — clusters, concentration, fresh wallets and insiders all work; snipers and bundles are not available
Arbitrum
Partial — same shape as Robinhood Chain. Concentration, wallet quality and shared-funder clusters all work; the behavioural tags (snipers, bundles) come from a source with no Arbitrum coverage, so those rows are absent
Mantle
Partial — same shape as Arbitrum. Concentration, wallet quality and shared-funder clusters all work; the behavioural tags (snipers, bundles) come from a source with no Mantle coverage, so those rows are absent
Arc
Partial — same shape as Arbitrum and Mantle. Concentration, wallet quality and shared-funder clusters all work; the behavioural tags (snipers, bundles) come from a source with no Arc coverage, so those rows are absent. The transfer graph runs on Circle's primary Arc endpoint out of the box; Arc's private mainnet phase makes that endpoint permissioned, so on a deployment it does not recognise those rows report as unmeasured rather than wrong. Wallet values are in USDC, Arc's native gas token
Every chain returns the same shape of result, which is why the Telegram card, the web panel and the bubble map all work unchanged across chains.
A signal that couldn't be measured is left out, not reported as zero. Where a chain has no source for snipers or bundles, those rows are absent from the card rather than showing 0% — an absent row means "not checked here", and a 0% row means "checked, and there were none". They are not the same claim, and on a launch the difference matters.
Why it's behind a button
A full holder scan is expensive — it means walking the entire holder set and then tracing the history of each significant wallet to find where its funding came from. That's orders of magnitude more work than reading a chart, and running it on every /don would exhaust the day's capacity by lunchtime.
So: on demand only, and results are cached per token for a while, so tapping the button twice in five minutes gives you the same answer without paying twice.
It's also budget-aware end to end. If a data source is close to its ceiling, the analysis degrades to a partial result rather than failing — and any hiccup anywhere in the pipeline degrades to "unknown" rather than crashing. The button always answers with something.
What it deliberately doesn't expose
You get the significant holders — address, percentage and role tags — and, for the full view, what's needed to draw the map. You don't get the complete holder set or the internals behind the verdicts.
That's enough to render the visualisation and check the work. It isn't enough to reconstruct the analysis, or to use the panel as a free data pipe.
Last updated