At the pinned 25 September 2026 snapshot, the estimated float for the identified ownership map is 5,626,068.46–5,796,756.45 FP/IMD, worth $49,631,233.20–$51,136,983.70 at the main Ethereum pool’s on-chain spot price of 0.0032532542 ETH, or $8.8217, per token. This is a conditional, reproducible range, not a claim that every beneficial owner has been identified. The wider deduction removes funded associates and all inventoried LP balances; the narrower deduction removes verified controllers and their LP principal. Unidentified insiders could reduce float further. See the deduction table and coverage limits below.

The strongest new finding is that 0x34325…1266 is a protocol administrator, not merely an early buyer with friendly funding. FP’s owner() is 0x54c4…206c (FpFeeDistributor), whose owner() is 0x34325…1266. The distributor can change the team fee recipient, grant fee exemptions, designate AMM pairs and transfer FP ownership to its team wallet. Its 463,878.29 FP therefore enters the verified-controller deduction. This does not prove a legal employment relationship or establish whether its other activity is treasury management or market making. FP contract, fee distributor source; pinned calls in snapshot.json, account_checks.json, weekly_admin.json.

The complete launch-month scan finds 3,460 pool recipients and 5,691 buys. All 2,000 free seat mints and 781 current holder addresses reconcile from transfer events. Current seats comprise 127 uninterrupted free-mint holdings, 1,365 acquisitions with matched paid Seaport orders, and 508 subsequent transfers whose payment status remains unresolved. Funding during a bounded mint window identifies 23 shared-funder groups, which are farm candidates, not proven common owners. The requested exhaustive later cheap-buy population, lifetime exits, first-ever funding histories and bridge-timing clusters remain incomplete. They are not presented as completed research.

Snapshot and evidence convention

VERIFIED means supported by the captured contract read, receipt, event or explorer transaction record, at the stated block. INFERRED means an attribution or economic interpretation. UNCLASSIFIED, blanks and explicit coverage notes mean unknown, never zero. ENS labels are explorer labels unless independently stated otherwise; control conclusions use addresses.

Chain	Block	Block time UTC	Block hash
Base	51770376	2026-09-25T09:54:59+00:00	0xd3c5a97f7ace4bd106c8535d0719351791d1b79b21326834cf99eda0aab9121d
Ethereum	26053702	2026-09-25T09:54:47+00:00	0xeb7d424367af5b7f7f2c07d90a398a4064da897f8e4a06e4bdae303d1cb18729
Arbitrum	508727084	2026-09-25T09:55:00+00:00	0x5c9a505cac7a0f20fb5a5ff1a209af073a50e2322f018b48f4024dba6fc088ca
Robinhood	72135616	2026-09-25T09:55:00+00:00	0x1ade376b92c96aef3f27353af62e1241819b22e3283d9576c2a801eac74005df
HyperEVM	46841552	2026-09-25T09:55:00+00:00	0xee30fa05a337c89abcb22161db972b90dc44b98db3a84e3b578cdf7c6a3e2ac6

These are near-simultaneous snapshots, not atomic cross-chain state. Base and Ethereum closing hashes were cross-checked on independent providers (more.json, links_metadata.json). Full request URLs, parameters, timestamps and responses are retained as hashed evidence captures inside raw-captures.tar.gz; raw-index.csv maps archived names to requests. Evidence filenames below are relative to artifacts/evidence/; large captures may have .json.gz suffixes. All analyses were read-only. No wallet, transaction signing, approval or broadcast was used.

Supply, float and price

Float here means token claims outside identified controllers, associated restricted balances, protocol reserves and the selected LP exclusion. sIMD holders retain their underlying claims. The high end excludes controller-owned LP principal; the low end additionally excludes all inventoried pools and funded associates. The choice about publicly owned LP inventory is explicit rather than hidden in an insider label.

FP genesis mint was 10,000,000 to the deployer at Base 3,332,899. Twenty-two 2,000-FP burns followed at 3,332,900, and 9,956,000 FP entered the launch pair at 3,333,068. The seed receipt contains the accompanying WETH transfer. Thus the initial public distribution route was the pool, not an airdrop of the genesis supply; this does not assert that no later gifts occurred. The 22 burns are verified; the pet identities behind each burn were not independently reconstructed. Deployment, liquidity seed; launch_fp.json, launch_liquidity_receipt.json.

At the snapshot, FP supply is 7,097,702.58458774, including bridge collateral. Its buy and sell fee getters both return 4%, subject to exemptions; the deployment source initialized a different fee, so 4% is not assumed for 2023 purchases. Current getters are in snapshot.json at Base 51,770,376.

Item	Minimum deduction	Maximum deduction	Treatment
Burned from FP totalSupply, relative to genesis	2,902,297.42	2,902,297.42	deduct once from genesis
Dead-address FP still included in totalSupply	45.38	45.38	deduct separately; still in ERC20 totalSupply; remote current dead balances verified zero in remote_dead.json
Verified controller wallet balances + sIMD claims	825,296.87	825,296.87	deduct from both scenarios
Funded associates	0.00	46,754.29	deduct only from low float; affiliation inferred or supplied
Protocol inventory, including pending burns	285,935.15	285,935.15	Diamond plus hook wallet and hook LP/ERC6909 inventory; subtracted from pool row to prevent overlap
LP / pool inventory, excluding hook overlap	189,668.73	313,602.44	low deduction=known team LP; high deduction=all known pools; never both

Start from genesis and subtract the first row, or start from live FP totalSupply and skip that row. All remaining deductions are then applied once. Detailed disjoint amounts and historical rows are in float.csv, with wallet amounts in holders.csv.

The adapter holds 3,885,013.38876300 FP. Verified remote supplies are Ethereum 3,854,335.55593100, Robinhood 30,553.86976200, Arbitrum 9 and HyperEVM 11. HyperEVM’s token is 0xc9dded1fb2d1e1c872f53a2a2548b587c5350d65, found through Base adapter peers(30367). Robinhood’s endpoint reports EID 30416, and Base’s corresponding peer matches the supplied token. Ethereum and Arbitrum peers also match. details.json, more.json, robinhood_peer.json contain the calls.

Do not subtract the adapter again from FP totalSupply, then omit the remote claims. Conversely, when adding the five token contracts’ supplies, subtract adapter collateral once. That latter reconciliation gives 7,097,598.62151774 represented circulating contracts versus 7,097,702.58458774 underlying FP: 103.96307000 tokens remain unreconciled across the selected peers and asynchronous snapshots. In-flight messages, other peers and dust were not exhaustively reconciled. The headline uses the underlying FP claim basis; this small gap is disclosed rather than labelled a verified bridge loss.

Protocol inventory includes Diamond 277,637.64 FP, hook LP/ERC-6909 claims 7,677.28 IMD, pending burn custody 620.18, and fee custody 0.05634480. The hook’s ordinary IMD wallet balance is zero; counting only that wallet would miss its inventory in PoolManager. Pending burns in BurnExecutor and the Base receiver are excluded as custody, not counted again as completed burns. hook.json, burn_sink_balances.json, base_burn_receiver.json, fee_inventory.json, burns.json support these figures at the snapshot blocks.

sIMD has 24-decimal shares and 1,675,647.37 IMD backing. The complete share-transfer replay reconciles to totalSupply and 459 positive holders. Claims use the verified vault’s virtual-share conversion: floor(shares_raw × (totalAssets + 1) / (totalSupply + 10^6)), then divide by 10^18 for IMD. All holders and the top 30 ranks are in holders.csv. No blanket staking deduction is made. Vault ownership is renounced at the snapshot; its owner is zero and paused is false. simd_logs.json, snapshot.json, and vault source.

LP owner / custody	Token principal at snapshot	Evidence
frenpet, v4 NFT #374493	129,617.61 IMD	ownerOf, position liquidity, packed ticks and pool slot0 at Ethereum 26,053,702
0x047f606fd5b2baa5f5c6c4ab8958e45cb6b054b7 — Original Base pair	288.59 FP	LP balance / totalSupply × FP reserve; Base 51,770,376
0xe764da9bdea91d845aac2c7d53a8dfe59a369095 — Base pair 0xDE66…7337	1,037.30 FP	LP balance / totalSupply × FP reserve; Base 51,770,376
0x038d3417508aecf2889f79ebff61b32564282534 — Base pair 0xDE66…7337	58,725.22 FP	LP balance / totalSupply × FP reserve; Base 51,770,376
rhynotic vault, 13 enumerated v4 NFTs	90,000.00 IMD	lp_positions.csv; owner and liquidity reads at Ethereum 26,053,702
Other/unassigned inventoried pool amounts	33,933.70	Residual after the listed owner positions; included only in the wider LP deduction, not automatically team supply

Owner-principal formulas exclude uncollected fees; the broader pool-balance deduction can include unassigned fees and use token-side principal at the pinned price. PoolManager’s entire IMD balance is not assigned to the main LP owner. The rhynotic 90,000-IMD wall is independently re-confirmed through positions; its tick ranges and USD conversions are in lp_positions.csv. The earlier job’s $10–$59 labels are not treated as immutable because ETH/USD changes.

The spot quote comes from Ethereum StateView getSlot0 for pool 0xb07d…abfb3: token0 is native ETH, token1 is IMD, both 18 decimals, so ETH/IMD is 2^192 / sqrtPriceX96^2. USD uses the pinned ETH/USD Chainlink feed 0x5f4e…8419 and its update time in float.csv. Base’s larger WETH/FP pool has a different quote due to venue/tax effects. These are spot marks, not executable liquidation proceeds; multiplying float by a marginal price is not exit liquidity. StateView, ETH/USD feed, valuation.json.

Weekly history

Each row selects the last block at or before 00:00 UTC on the date. Current restrictions are applied to historical balances, not represented as historical public knowledge. Weekly member rows measure Base, Ethereum and sIMD directly. Historical holdings on other chains are not attributed: the low float additionally reserves all positive Base collateral minus Ethereum supply as an unassigned remote-claim bound; the high float leaves these claims in public float. This is conditional on the selected OFT backing model and also covers in-flight balances, so it can overstate the deduction. Current controller deductions include the directly read Arbitrum, Robinhood and HyperEVM balances. Current and weekly ownership calls verify the fee-distributor control chain throughout the sampled period. Where the main v4 pool is not yet initialized, the price falls back to the Base 0xDE66…7337 reserve quote; float.csv records the venue switch. Two identified older Ethereum v3 pools are inventoried in the wider deduction, but historical NFT ownership for every LP was not reconstructed.

Date	Base / Ethereum block	Float low–high	Float cap USD low–high
2026-08-01	49376526 / 25656292	5,150,729.85–5,498,498.60	$2,415,341.45–$2,578,421.31
2026-08-08	49678926 / 25706531	5,105,010.49–5,595,827.94	$2,460,942.42–$2,697,547.90
2026-08-15	49981326 / 25756720	2,706,556.93–5,757,225.07	$3,156,975.43–$6,715,328.20
2026-08-22	50283726 / 25806958	2,583,764.12–5,396,003.97	$3,308,150.59–$6,908,832.56
2026-08-29	50586126 / 25857181	5,251,027.33–5,319,704.67	$5,110,673.59–$5,177,515.27
2026-09-05	50888526 / 25907399	5,536,529.81–5,612,211.49	$19,021,364.56–$19,281,377.39
2026-09-12	51190926 / 25957605	5,790,004.74–5,804,535.59	$16,152,511.52–$16,193,048.57
2026-09-19	51493326 / 26007835	5,575,136.95–5,654,310.14	$18,307,239.17–$18,567,222.48

weekly_members.csv gives each supplied team/cluster address’s FP, IMD and sIMD claim at every weekly block, independently of whether its classification is verified or inferred.

Team map: what is established

Lead	Finding and evidence
0x047F…54B7	VERIFIED genesis deployer and current IMD, Base adapter, hook and Diamond owner. HyperEVM, Arbitrum and Robinhood owner reads also match. FP ownership is now mediated by the fee distributor. ENS “surfsurf.eth” is an explorer label.
frenpet 0xE764…9095 and Safe 0x038D…2534	VERIFIED current FP teamWallet; Safe owners are exactly frenpet and 0x047F, threshold two; v4 NFT #374493 owner matches. snapshot.json. The 7,382 IMD transfer to 0x047F at Ethereum 25,886,274 is verified (transaction); it is a transfer, not a sale.
0x34325…1266	VERIFIED fee-distributor controller. VERIFIED 90.5 ETH from deployer on Ethereum (89 + 1.5), 50,000 FP on Base, plus repeated frenpet ETH and FP distributions. The supplied 233.65 ETH is not repeated as an exact total: observed qualifying transfers include an additional 0.298 ETH. Treasury versus market-maker mandate remains unknown.
Diamond 0x0e22…443C	VERIFIED 0x047F owner; installed diamondCut selector resolves to an owner-gated facet. Owner can replace/add implementation logic and thereby move FP. Current gameplay paths were not fully audited. VERIFIED 51 + 1 ETH from frenpet. details.json, more.json, lead_checks.json; creation metadata in snapshot.json.
0x0e0D…04c9	VERIFIED 8.5 + 0.8 ETH funding and 14,000 FP transfer from frenpet. INFERRED affiliate, not proven common key. 14,000 FP transfer.
0x2591…beE7	VERIFIED 10,000 FP from deployer at Base 7,933,856. INFERRED associate; a gift alone does not prove insider control. Transfer.
0x200E…0fB1	VERIFIED 0.054 ETH from deployer at Ethereum 25,091,363. “Team messaging address” remains a supplied role label, not independently proven by funding. Funding.
0x4683…9c21 (extension)	Repeated distributions from frenpet make this an INFERRED associate worth following. Measured Base/Ethereum/Arbitrum/Robinhood token and sIMD balances are zero; HyperEVM was not read for this added address. No employment or common-key conclusion.
rhynotic three-address map	Vault LP principal is VERIFIED; wallet and sIMD balances are re-read. The named-person/common-control grouping remains INFERRED from the supplied lead plus captured funding edges. No team control link is established, so this cluster is not deducted as team supply.

All requested large-holder addresses are present in holders.csv. Most remain UNCLASSIFIED: being large, buying early, sharing a public router or staking does not establish insider status. In particular E9fc…, 4CDC… and hundo’s 2888… have EIP-7702 delegation bytecode at the pinned block (account_checks.json); a block explorer’s “contract” label is not evidence that they are protocol treasuries.

Cheap supply and sellers

The complete scan covers Base 3,332,899–4,628,898, exactly the first 30 days from deployment at two-second block intervals (UTC endpoints in lead_checks.json). It records 8,532 Swap events, 5,691 buy events and 3,460 unique pool recipients. Gross FP bought in this window is 22,559,329.45; this is repeated trading turnover and can exceed total supply. Net received and early sale attribution require receipts: 6,637/8,281 receipts are available. Missing receipts leave net totals blank rather than assumed tax-adjusted.

cheap_supply.csv contains every first-window recipient, exact buy blocks/transactions, gross amounts, available net amounts, pool ETH spent, early sales, present FP/IMD/sIMD claims and verified recent direct-pool sales. The measured present claims across this population total at least 880,431.86 tokens. Missing wallet reads by chain are {"fp_base_now": 0, "imd_eth_now": 0, "imd_arb_now": 1587, "imd_robinhood_now": 0, "imd_hyper_now": 0}. The unread Arbitrum addresses cannot add further tokens under the standard ERC20 supply invariant: the entire verified 9-IMD Arbitrum supply is already held by the measured deployer. Their individual RPC fields remain blank rather than being fabricated. An address’s remaining balance is not necessarily the same coins it bought cheaply: gifts, rebuying, staking, bridges and outgoing transfers break cost-basis identification. Pool recipients can also be routers or agents, not independent investors.

The alternative “bought below 0.00002 ETH at any later date” condition is not exhaustively scanned. The in-window price flags use gross pool FP output; first-30-day inclusion does not depend on those flags. Consequently the CSV is a complete launch-window recipient list, not the complete lifetime cheap-supply universe. All-time bought/sold columns are intentionally blank. Sales between the launch window and the last 30 days are not fully reconstructed.

For the latest window, both named Base pairs have complete Swap-log coverage across 50,474,376–51,770,376, but only 10,207/12,257 transaction receipts survived provider limits. Ethereum token-transfer coverage spans the exact 30-day cutoff block through 26,053,702 (recent_imd.json); receipt attribution is limited to supplied-map and launch-window addresses. Direct token transfers into a venue with a swap in the same receipt form a verified-sale lower bound. Router-mediated attribution, other venues, LP additions in mixed transactions, and historical off-chain executions need further tracing. Amounts below are tokens received by the pool, not gross wallet debits including FP tax.

Seller	Chain	Verified pool-received tokens	Classification
0x2888be6982957103ea88ac10f4179df0a294e877	eth	5,000.00	UNCLASSIFIED; launch-window recipient
0xde62c9d66884b78270d212a2c64a47cfd7641a4e	eth	4,990.21	INFERRED_RHYNOTIC_CLUSTER_NOT_TEAM; supplied map
0x29ed6881cb6698bffbf24bb3b8b3f20a8c51a2ca	eth	3,015.38	UNCLASSIFIED; launch-window recipient
0x14837a7fd681000b613ae034bee8960c1a46d086	base	1,303.68	UNCLASSIFIED; launch-window recipient
0xb6d46c5301c85e708780e5cafdb2bb4a61894d0a	base	782.90	UNCLASSIFIED; launch-window recipient
0xf3d756a3431182ee75e9e41652e21da76f2d97e0	base	25.50	UNCLASSIFIED; launch-window recipient
0x887b86b6b6957f7bbea88b8cefd392f39236a88c	base	9.60	UNCLASSIFIED; launch-window recipient
0x4a09b2ef6eee4635555e5e32b61b2412e76644fb	eth	8.81	UNCLASSIFIED; launch-window recipient

sellers.csv supplies every included transaction, venue, block and available UTC time. No verified direct-pool sale by the confirmed core controllers was found in the captured attribution set; this is not proof that the team made no sales. Transfers to PoolManager without a matched swap are not called sales. The rhynotic sIMD holder’s exit is distinguished from a staking withdrawal by the ensuing swap receipt.

Free seats and cluster limits

The NFT mint events run from Ethereum 25,090,094 at 2026-05-14 01:30:11 UTC to 25,090,304 at 02:12:11 UTC, exactly 42 minutes. Each minter is unique. The verified nonpayable mint function permits one per address, and the full transfer replay produces 2,000 live token IDs and 781 current holders. All mints were free of mint-price payment; gas was still paid. The supplied floor of 1.78 ETH is not re-certified: an executable NFT floor needs current signed listings, not merely historical transfers. Future reward/allocation entitlements are not quantified here. NFT contract.

nft_free.csv has one row per seat, with original minter, current holder, last acquisition transaction, retained-free/verified-paid/unknown classification, first-observed funder and token overlaps. Holder totals repeat on each of that holder’s rows; do not sum repeated holder totals. The 1,365 paid acquisitions require a Seaport 1.6 OrderFulfilled event matching this NFT/id, monetary consideration and recipient. The other 508 transfers are not casually relabelled as purchases or free gifts.

The additional funding scan covers full Ethereum blocks 25,089,894–25,090,304. It finds 363 successful native-ETH transfers to minting addresses, checked against receipts. This is a short pre-mint/mint window, not lifetime account funding history. Internal transfers and earlier exchange withdrawals are not fully captured. Shared service funders can mimic farms; candidate groups must not be counted as proven people or insider clusters.

Candidate shared funder	Minter addresses in captured group	Interpretation
0x2f92c25453818b02f8180d4cd2097c349bfc8850	103	INFERRED farm candidate; funding evidence in clusters.csv
0xdd6cebe9283079daffc06a112b7b848cca7e22fb	16	INFERRED farm candidate; funding evidence in clusters.csv
0x5cf9b3fc56305251f7737437c4ea9cd8bdbf8252	14	INFERRED farm candidate; funding evidence in clusters.csv
0xf58685e2dcf6700db6e7f302f697f2ead460261d	10	INFERRED farm candidate; funding evidence in clusters.csv
0x023d5dcc012288ae0d07dec3af3f998226130b36	7	INFERRED farm candidate; funding evidence in clusters.csv

Token overlap is address-level only: 3 seats were minted by launch-window FP recipients, and 67 current seats are held by those recipients. 215 current seats are held by the snapshot’s top-30 sIMD addresses. These counts refer to seats, not unique people; per-address totals are in nft_free.csv. All refer to the pinned Ethereum ownership replay and first-month Base scan.

clusters.csv separates verified funding edges, verified controller membership, inferred affiliates, candidate NFT farms and same-block buy coincidences. Overlapping screening groups must not be added together. Same-block buys and shared routers do not independently justify clustering. LP-principal rows are separate from wallet claims. First-ever ETH sources and bridge-timing/common-counterparty clustering are incomplete, and unknown group holdings are blank.

What remains unanswered and how to reproduce

The key unresolved work is: all later sub-threshold FP buyers; lifetime sales and routed beneficial ownership; every large holder’s role; complete first-funder and internal-transfer graphs; matching bridge messages across every peer; all historical LP-NFT owners; and the 508 NFT acquisition payments not matched to the supported Seaport event. The headline range covers the identified map and inventoried contracts, not every possible undisclosed insider or unknown protocol balance.

Public RPCs returned HTTP 403/429, batch-size limits, block-range caps and pruned historical-state errors; explorer legacy and paginated endpoints also rate-limited. Rotation used Base mainnet, Blast, Nodies and BlockPI; Ethereum PublicNode, Flashbots and MEV Blocker; Robinhood Ordofi/Blockmachine; HyperEVM Sentio/Thirdweb; and several Arbitrum providers. Raw errors are retained alongside successes. Log range completeness is tracked separately from receipt and balance completeness: a completed log scan does not turn failed receipts into zero activity. Broad scans that kept hitting limits were stopped rather than represented as finished.

research/build.py, research/position_table.py and research/report.py derive the deliverables offline from captured evidence, using only Python’s standard library. Collection scripts are separate and make read-only calls. README documents file units and the offline check command. Local consistency checks do not independently certify the chain data, provider integrity or inferred identities.