Some report sources did not pass validation.
Chapter 01 · Executive overview
The Phase 1 story, organized for exploration.
Start with the report-wide signals, then move into the evidence chapter that answers your question.
Report directory
Six focused chapters. One validated evidence bundle.
Choose the question you want to answer; every chapter keeps its own shareable URL.
01Overview
Headline findings and report-wide metrics.
3 surfaces 02Market flowsTrends, events, sensitivity, and routing.
7 surfaces 03OwnershipHoldings, concentration, supply, and staking.
8 surfaces 04WalletsPrioritization and address-level evidence.
3 surfaces 05RewardsWithdrawals, attribution, disposition, and restaking.
6 surfaces 06Validators & methodsValidator evidence, destinations, and coverage.
4 surfaces
Evidence coverage
See what Phase 1 can—and cannot—answer.
A fixed, validated Phase 1 evidence bundle with explicit boundaries.
—surfaces ready
—validated sources
—evidence groups
R5privacy-sanitized release
Phase 1 overview
Report-wide metrics
Validated aggregates across the fixed research periods
Primary evidence view
Weekly Weighted Sell Pressure vs ATOM Price
Loading validated bundle…
What this view shows
The chart summary will appear after the validated bundle loads.
View weekly chart data 22 rows
| Dates |
|---|
Executive readout
Four signals define the Phase 1 story.
The report points to concentrated, episodic market-facing flow—not a continuous uniform sell pattern.
01Concentration
50.6% from whale holders17.84M weighted ATOM across the tracked 22-week cohort study.
Lead signal02Burst detection
Two ≥2× baseline weeksFeb 26 and Apr 30 reached at least 2× their trailing four-week baseline.
Episodic, not uniform03Lead indicator
Unstaking did not lead sellingOne-week lead correlation was −0.31 across 21 week pairs.
No observed lead04Route concentration
96.7% routed direct to CEXThe tracked transaction-level flow strongly favored direct exchange routes.
Dominant route
Published error band
Sensitivity & Evidence Confidence
Gauntlet published a CEX-source sensitivity and a confidence-weighted split for every week. Both qualify the headline total.
Weighted pressure by confidenceRoute classification
share of weighted total
Exchange-sourced shareHighest weeks
CEX-sourced ÷ weekly weighted
View exact sensitivity data
| Week | Dates | Weighted | CEX-sourced | CEX share | Excluding CEX | High conf. | Medium conf. |
|---|
Route-weight what-ifRecomputed from published raw amounts
View recomputed route arithmetic
| Route | Confidence | Raw ATOM | Published weight | Applied weight | Recomputed ATOM |
|---|
Message-type mixWhat the routes are made of
View exact message-type data
| Message type | Events | Share of events | Raw ATOM | Weighted ATOM | Share of weighted |
|---|
This is Gauntlet’s own sensitivity, not a correction to their work.
The headline weighted total counts movement that originates at known exchange wallets. That flow may be internal exchange wallet management rather than holders selling, so the report published a figure excluding it. Treat the pair as a range. The confidence split describes certainty in route classification only — it does not claim that routed ATOM was sold.
Who holds ATOM
Cohort Holdings & Share of Supply
Weekly balances for the fixed Phase 1 cohort, split into liquid, delegated, and unbonding ATOM.
Share of chain supplyBalance held
actual share · 0–100%
Balance compositionDelegated share
vs cohort average
View exact holdings data
| Cohort | Addresses | Funded | Liquid | Delegated | Unbonding | Total ATOM | % supply |
|---|
View largest holders by balance
| Cohort | Rank | Address | Label | Liquid | Delegated | Total ATOM | % supply |
|---|
Balances are snapshot at each weekly window height for a cohort fixed at 2026-05-31. Holding ATOM is not sell pressure; these are balances, not flows.
Liquid staking providers
Liquid-Staking Protocol Balances
Stride, Quicksilver, and Drop pooled ATOM, matched to published protocol addresses.
Matched pool compositionShare of visible pooled ATOM
View exact liquid-staking data
| Protocol | Status | Matched addresses | Delegated | Unbonding | Total ATOM | % supply | 22-week change |
|---|
Where the pools delegateValidators receiving LST stake
View exact LST delegation data
| Protocol | Rank | Validator | Operator address | ATOM | % of pool |
|---|
Delegation figures are from the April 2026 on-chain review, so they predate the 2026-05-28 pooled balances above and the two should not be mixed.
Distribution analysis
ATOM Holder Concentration
How the tracked cohort’s balance is distributed across funded addresses at the final snapshot.
Distribution curve
Tracked balance by funded address
Observed distributionPerfect equality
View exact curve data
| Point | Population share | Cumulative balance share |
|---|
Addresses are not people: one holder may control many addresses, and a single exchange address may custody balances for many holders. Concentration describes the address distribution only.
Issuance context
Weighted Pressure vs New Issuance
Weekly ATOM emitted, derived from chain-supply deltas between snapshot heights, against weighted sell pressure.
Weekly evidence comparison
Emitted ATOM vs weighted pressure
Shared ATOM scale · Jan–May 2026
Emitted ATOM
Weighted pressure
1.0× parity
View exact emission data
| Week | Dates | Emitted ATOM | Weighted pressure | Pressure ÷ emission |
|---|
Emission is the chain-supply delta between consecutive weekly snapshot heights, so it is net of any burns or community-pool movement in the same window. Weighted pressure is a routing estimate, not confirmed sales, so this ratio is directional rather than a realized sell-through rate.
Queued exits
Unbonding Queue
ATOM sitting in the 21-day unbonding queue at each weekly snapshot height.
Unbonding stock by week
Jan–May 2026 · hover or focus a bar for exact values
View exact unbonding data
| Week | Dates | Unbonding ATOM | Delegated ATOM | % of delegated |
|---|
This is a stock, not a flow: it is the balance in the queue at one moment, so values cannot be added across weeks. Unbonding carries a route weight of zero and contributes nothing to the weighted sell-pressure total. A rising queue signals ATOM becoming liquid, not ATOM being sold.
Holdings versus flows
Routing Propensity by Balance Tier
How much of its own balance each size of address actually routed, next to how much ATOM the tier moved in total.
Median share of balance routedTypical address
median, per tier
Weighted ATOM routedTier total
share of routed volume
View exact propensity data
| Balance tier | Addresses | Median % routed | Mean % routed | Routed nothing | Weighted ATOM | Share of routed |
|---|
Balance is the largest observed snapshot balance for each sending address, so the ratio is indicative rather than a precise portfolio percentage. Routing is not selling.
Pre-window baseline
Where ATOM Stood Before This Study
The network snapshot Cosmos Labs published in its tokenomics RFP, quoted so the Phase 1 findings have a starting point.
Inflation mechanicsWhere the rate sits in its band
floor, current, ceiling
Staking participationBond ratio against target
share of circulating ATOM
View exact baseline figures and sources
| Metric | Value | Unit | Source |
|---|
A dated third-party snapshot, not a live reading and not a Phase 1 finding.
These figures come from Cosmos Labs’ own Request for Proposals: ATOM Tokenomics Update, which sourced supply from CoinGecko and staking from Mintscan. The RFP predates this study’s January–May window, so treat it as the starting position rather than a current or end state. Nothing on this panel was produced by Gauntlet, and none of it moves with the live ATOM price shown in the header.
Network context
Weekly Inflation Claim Benchmark
Cosmos compared with NEAR and Ethereum using the report’s fixed peer snapshot.
Claimed share of weekly emissionsReward realization
Claimed ÷ emitted
Claimed share of token supplyRelative supply impact
Claimed ÷ supply
View exact peer-benchmark data3 networks
| Network | Emitted/week | Claimed/week | Claimed/emitted | Claimed/supply |
|---|
Peer definitions, issuance mechanics, and supply structures differ. This is descriptive context—not a performance ranking.
Evidence boundary
A fixed Phase 1 research snapshot—not a live market-monitoring product.
The report does not include ongoing monitoring or full address coverage beyond its fixed cohort.
Cohorts + routesJan 1 – May 3122 weekly windows
Rewards + validatorsJan 1 – Mar 2211 weekly windows
Top-wallet net flowJan 1 – Jun 144-wallet study
Historical events7 windowstracked-cohort shares
Stakeholder cohorts this report does not resolve.
Of the ten cohorts Cosmos proposed, the exports identify exchanges, validators, ICO participants, treasury, an aggregate employee cohort, and liquid-staking providers. Employee addresses are excluded from public address-level evidence. The exports do not separately identify DeFi/DEX participants (only IBC routing intent stands in for them), community-pool recipients, or passive long-term holders (no pre-January holding history exists in the snapshot). Institutional and fund holders appear only as the size-based “HNW / small fund” tier, not as an identified entity list. Absent does not mean small.
Evidence boundary
Phase 1 Coverage Matrix
What each validated surface covers—and what it does not.
Supported
Evidence the validated surface directly provides.
Boundary
6 evidence groups · original CSV row order preserved
Questions the source cannot answer.
| Evidence surface | Frozen window | Observed population | Included evidence | Not included | Status |
|---|
Interpret these results as behavioral indicators—not allegations.
Wallet addresses are public blockchain records. Labels and classifications reproduce the cited Phase 1 sources and may be incomplete or change over time.
Published cohort leaderboard
Cohort Sell Pressure
Likelihood-weighted market-facing flow · 22 weekly windows · Jan–May 2026
| Rank | i | Weekly share range |
|---|
Selected event brief
Event Detail
7-day scan from event date
Event percentages use each event’s period-correct tracked-cohort balance, not total ATOM supply.
Participation lens
Cohort Participation & Flow Intensity
Active-wallet observations distinguish broad participation from high flow per observed wallet.
Weekly activity
Jan–May 2026 · hover or focus a bar for exact values
View exact participation data22 weekly windows
| Week | Dates | Active-wallet observations | Weighted ATOM | ATOM / observation | Leading cohort |
|---|
Counts are summed cohort-week activity observations from the export. They are not a complete holder census and should not be added across weeks as unique wallets.
Seven peer windows
Historical Event Ranking
Complete peer comparison using each event’s own period-correct tracked-cohort denominator.
| Pressure rank | Event | Date | Type | Weighted ATOM | Tracked-cohort share | Daily pace / baseline | Baseline multiple | Whale share | Details |
|---|
Ranks compare seven report-defined 7-day windows. They do not establish causation between an event and market-facing routing.
Selected-event concentration
Event Flow Contributors
Complete selected-event wallet leaderboard.
Largest contributorsSelected event
Share of event weighted pressure
View complete contributor table—
| # | Address | Known label | Weighted ATOM | Event share | Cumulative share | Transactions |
|---|
Contribution ranks describe observed market-facing routing in one report-defined window. They do not establish ownership, intent, or confirmed sale execution.
Wallet Prioritization Heuristic
A transparent review queue—not a risk rating or prediction
| # | Address | Cohort | Flow pts | Event pts | Behavior pts | Total | Evidence confidence | Sensitivity | Details |
|---|
How the prioritization score works
Flow component: log₁₀(1 + Jan–May weighted ATOM).
Event component: 2 points per historical event leaderboard appearance.
Behavior component: +3 for a top net-seller tag, +1.5 for router-sensitive inflows, and −2 for a top accumulator tag.
Evidence confidence: reflects how many independent Phase 1 surfaces observed the wallet; it is not confidence that the wallet is risky.
Address-level evidence
Top Sell-Pressure Wallets Per Cohort
Tracked subset from the final report
| # | Address | Label | Holder tier | Staker tier | Weighted ATOM | Raw ATOM | Weeks | Cohort share | Details |
|---|
Cohort-to-cohort routing
Transfer Clusters & IBC Destinations
Where tracked routed ATOM originated and where it landed, from the transaction-level weekly exports.
Cohort to cohortWeighted routing
actual share of 35.22M weighted ATOM
Exchange-bound
Unlabeled destination
0–100% scale
IBC destination chainsWhere ATOM left to
actual share of 1.96M IBC weighted ATOM
Destination share
0–100% scale
View exact cluster data
| Source cohort | Destination | Events | Raw ATOM | Weighted ATOM | % weighted | Weeks seen |
|---|
View exact IBC destination data
| Destination chain | Events | Swap-memo events | Plain events | Raw ATOM | Weighted ATOM | % weighted |
|---|
Named swap destinationsContracts resolved from transfer memos
View exact swap-destination data
| Chain | Contract | Events | Raw ATOM | Weighted ATOM | Share |
|---|
Destination labels are the Phase 1 label set; unlabeled means no entity label was resolved, not that the counterparty is unknown to the chain. A swap memo indicates routing intent, not a confirmed swap or sale. Named contracts are read directly from transfer memos, which resolves slightly more swap intent than the route label does: 194 transfers carry a router contract on transfers the route classifier labelled plain IBC, while a small amount of swap-labelled flow carries no resolvable contract, leaving the named-contract total about 4.2K ATOM above the swap-memo subtotal.
Reward Disposition
Withdrawn staking rewards + commission · Jan–Mar 2026
Hover for weekly details · click to pin
Across 11 weeks, 42.6% of 8.18M withdrawn ATOM was routed to same-week selling, 27.7% was re-staked, and 29.7% remained liquid.
| Reward-sourced route | Weighted ATOM | Share |
|---|
Fixed wallet study
Net Flow — Top 44 Wallets
Jan 1–Jun 1, 2026 · raw ATOM · counted-flow baseline
Withdrawal origin
Delegator Rewards vs Validator Commission
Exact weekly withdrawal origin from the 11 validated inflation exports.
View exact weekly origin data11 weeks
| Week | Delegator rewards | Validator commission | Total withdrawn | Commission share | Commission events |
|---|
This separates where withdrawals originated. Phase 1 does not attribute validator-commission destination or subsequent disposition at the individual-validator level.
Reward-lot attribution
How Much Routed ATOM Came From Inflation
Gauntlet's reward-lot tracking follows withdrawn rewards through transfers and counts the share that reaches a market-facing route.
Inflation-attributed shareBy route
inflation ÷ overall weighted
Weekly share11 reward weeks
Jan 1 – Mar 22, 2026
View exact attribution data
| Week | Dates | Withdrawn | Re-staked | Inflation-attributed route | Overall weighted route | Share |
|---|
Reward credit created each weekLeft liquid vs re-staked
View exact reward-credit data
| Week | Withdrawn | Left liquid | Re-staked | Credit tracked | Re-staked share |
|---|
Denominator boundaryRead this alongside the 35.22M figure, not against it.
This attribution runCalendar reward weeksChain-wide route populationHeadline study35.22M weighted ATOM62,868-address fixed cohortRequired interpretationDo not multiplyDifferent populations and week boundaries
This attribution run measures chain-wide route volume over calendar-anchored reward weeks. The headline 35.22M weighted total measures the fixed 62,868-address cohort over strict seven-day windows from January 1. The two use different populations and different week boundaries, so the percentage here is the inflation share of this run's own denominator and cannot be multiplied against the cohort total. Reconciling the two onto a common denominator is open Phase 2 work.
Where rewards went
Commission vs Delegator Reward Disposition
Reward-lot attribution splits withdrawn rewards by origin kind and by the route they reached, including validator commission.
Share of withdrawals reaching a routeBy origin kind
routed ÷ withdrawn
Route mixOrigin kind × route
weighted ATOM
View exact disposition data
| Origin kind | Route | Attributed ATOM | Weighted ATOM | Events |
|---|
View per-validator delegator-reward routing
| # | Validator | Direct CEX | IBC swap memo | IBC no memo | Weighted ATOM | Share | Weeks |
|---|
Commission is attributed in aggregate only: the export carries no validator identity on commission rows, so per-validator commission destination remains unavailable. Per-validator figures below are delegator rewards attributed to the originating validator — they describe where that validator’s delegators routed, not the validator’s own conduct.
Withdrawal step
Who Received the Rewards
Every reward withdrawal, split by whether it went to the delegator or to a different address.
Recipient of the withdrawalShare of withdrawn ATOM
delegator vs redirected
Delegator concentrationShare of withdrawn ATOM
by delegator group
View exact recipient data
| Recipient | Transactions | ATOM | Share |
|---|
View largest withdrawing delegators
| # | Delegator | Withdrawn ATOM | Share | Transactions | Validators | Weeks |
|---|
Redirected does not mean sold. A delegator can set any withdrawal address, and custodial or multi-wallet setups routinely use a separate one. Where Phase 1 could not identify the recipient, that is a limit of the label set, not evidence about the recipient. Delegator counts are addresses, not people.
Where rewards compound
Restake Destinations
Withdrawn rewards that were re-staked, attributed to the validator that received them.
View exact restake data
| # | Validator | Re-staked ATOM | Share | Delegators | Events | Weeks |
|---|
This is where re-staked rewards landed, not a judgement about any validator. Delegators choose where to compound. Counts are per-week distinct delegators summed across the window, so they are not unique across weeks.
Distribution analysis
Reward Withdrawals By Originating Validator
Jan 1 – Mar 22 2026
Cumulative distributionOriginating-validator concentration
Attributed withdrawalsEqual distribution
Ranked validator evidence
Attributed delegator withdrawals
Showing the leading 60 of 239 originating validators
| # | Validator | Tier | Stake share | AUM | Withdrawn ATOM | Share | Withdrawal events | Peak recipients | Weeks |
|---|
Coverage boundary
Validator tiers are shown for part of the set only.
Read tier-mapping methodology
Venue concentration
Exchange Destination Concentration
Two validated routing surfaces with separate populations and denominators.
A Reward-sourced routesJan–Mar 2026
B Top-wallet routed pressureJan–Jun 2026
Destination shares reproduce the report aggregates. The HHI figures are lower bounds because unlisted residual destinations are not assumed to be one venue.