---
name: marlin-jt
description: Full deep dive on MARLIN / J.T. Marlin (jtmarlin.trade) on Robinhood Chain, 8 Sep 2026 - the Pons V2 launch that rotates its own 4% trade fee into protocol-owned stock-token LP. Read before any claim about its yield, its book, or its "paid daily" promise.
sources: [cowork]
aliases: [MARLIN, JT Marlin, jtmarlin, JTMarlinTrade, marlin-rotation]
---

# MARLIN / J.T. Marlin, screened 8 Sep 2026 at ~04:30 UTC (age 9.0h)

Token `0xaB35D04E1Ee39c789c4A65D522417b3C7F2E4723`. Top pool `0x59a626b46f61a117f13a4f3d31264dc9afcafff4c95c3e79d094f07cc050a3f0` (v4, MARLIN/ETH). Site jtmarlin.trade, X @JTMarlinTrade (268 followers). Named after the brokerage in the film Boiler Room.

## THE MECHANISM, derived not assumed
- **The token contract carries NO tax.** 6,498 bytes, no `mint`, `pause`, `owner`, `transferOwnership`, `blacklist`, `setFee`, `setMaxWallet`; `burn` present. Immutable, unowned. So "every $MARLIN trade carries a 3% protocol fee" is NOT a transfer tax: it is the **Pons V2 pool fee**, claimed by the creator from `PonsV2FeeEscrow`.
- Launched through `PonsV2LaunchFactory` `0x7eD598BcEf8bd9Edd8C97A195C6d13f40801EC7e` at 15:26 UTC 7 Sep by operator EOA `0x1350bbb6d6ed3f60053a00caeddee6cd82bb8997`.
- **The flow, fully traced:** operator calls `claim` on `PonsV2FeeEscrow` `0xd3AFEB2a57f70eF218Aa82451c51B2fb0416Ac9e`, then forwards ETH to desk EOA `0x2Faf625bF5bfB6b09deEE3b1B62D142ADF4578cb`, which swaps to a stock token and mints/increases a Uniswap **v3** LP position paired against USDG.
- **Fees claimed in the first 8.5h: 21.3796 ETH = ~$53,150** (6.4278 + 0.3685 + 0.8450 + 10.6383 + 3.1000). **Forwarded to the desk: 16.0251 ETH = ~$39,838. Retained in the operator EOA: 5.359 ETH = ~$13,323.**
- **Effective fee rate CONFIRMED: $53,150 of fees on $1,488,957 of 24h volume = 3.57%**, consistent with the site's published 4% routing (2.75% LP, 0.25% ops, 1% Pons rail).

## THE BOOK IS REAL. IT IS ALSO TINY AND IT IS IN A BARE EOA.
- Desk holds four position NFTs: v4 POSM id 2125052, and **v3 ids 1077487 (SGOV/USDG), 1074987 (HIMS/USDG), 1073924 (AMC/USDG)**, all fee tier 3000, all in ~360-tick (~3.6%) bands. Idle: 909.31 USDG, 65.39 HIMS, 0.0041 SGOV, 0.159 ETH.
- **Book value by cash accounting: $39,838 in minus ~$3,100 idle = ~$36.7K deployed.** This independently confirms the team's "~$36K book" claim to within noise. CONFIRMED.
- **`0x2Faf625bF5bfB6b09deEE3b1B62D142ADF4578cb` is a bare EOA (codelen 2). No locker, no timelock, no contract.** The whole "protocol-owned, never unwinds" book is one `decreaseLiquidity` away from being the operator's. This is the fatal column from [[onchain-forensics]] and it applies in full.
- Host pools the book sits in: HIMS/USDG v3 `0xC8C90d3a1c1a24967E773ac2aD0d456BA3E31F64` ($2.08M liq, $2.93M v24) · AMC/USDG v3 `0xaA34feA710a1A737840329051D81D3B0B7C564d5` ($2.64M liq, $2.93M v24) · SGOV/USDG v3 `0xfAb520051f96F4D2a32c22B6a3dD7fFfdf231bFe` ($3.89M liq, $7.27M v24).

## CLAIM AUDIT
- "Every $MARLIN trade carries a 3% (site header) / 4% (site body) protocol fee" — **CONFIRMED at 3.57% measured**, but the site contradicts itself on the number in two places.
- "3 pools live (AMC, HIMS, SGOV), book ~$36K" — **CONFIRMED on chain**, both the pools and the size.
- "Holders earn 80% of the LP fees, paid daily at 18:00 UTC" — **NOT ONE PAYOUT HAS EVER HAPPENED.** The desk's 36 transactions are entirely approves, swaps, mints and increases. There is **no distribution contract anywhere**; any payout would be a manual airdrop from an EOA. The split was also raised from 20% to 80% by tweet on day one, which means it can be lowered by tweet.
- "One published split. Visible onchain." — **FALSE as published.** The site prints exactly one address, the token itself. Neither the operator, the desk, nor the escrow appears anywhere on jtmarlin.trade, so nothing on it is verifiable from it.
- "$MET FDV is $198M, $MARLIN FDV is $700K, let that sink in" (@rantaiuang, 1,710f) — the MET number is **TRUE** ($198.9M FDV, $109.8M mcap) and the comparison is **FALSE as a comp**. Meteora is a DEX earning a take rate on other people's flow; MARLIN is a passive LP in other people's pools with no take rate on anything but its own token.
- "80% of these stats will be for $MARLIN holders once TVL and volume on @JTMarlinTrade can par Meteora" — MARLIN has no DEX. There is no venue to route volume to.
- Copycat live: **MRLN `0xc344bbe085E4b9F6E5AB7ff61EEee475bCE3D282`**, printing a $2.36M mcap on $97 of liquidity.

## THE STRUCTURAL FINDING: there is no toll booth
- MARLIN does not sit between any flow and any counterparty. Its only revenue source is **the trading of its own token**, which is pure reflexivity: fees exist while the chart moves and stop when it does not.
- The LP book earns a **pro-rata passive share** of AMC/HIMS/SGOV pool fees with no privileged position of any kind.
- Yield math on the book, honest bounds: those three pools generate roughly $8.8K + $8.8K + $21.8K per day of gross 0.30% fees; a $36.7K book is 0.13% to 0.96% pro-rata of pools holding $2M to $3.9M. **Pro-rata that is ~$149/day gross, ~$119/day to holders = ~10% annualised on a $438K cap.** With the tight 360-tick bands giving perhaps 3x to 5x in-range concentration, the optimistic bound is ~$400/day = ~33% annualised. Neither number is a 10x.
- **The only thing that can produce a 10x is the book compounding**, and that is funded by MARLIN's own volume: 2.75% of $1.49M/day = ~$41K/day added today. At 10% of launch-day volume it still adds ~$4.1K/day and doubles the book in nine days.

## MARKET AND HOLDER STATE, 8 Sep ~04:30 UTC
- Price $0.0004386, **mcap $438,615**, 1B supply. Aggregate liquidity across 17 pairs $72,000, aggregate v24 $1,566,028. Turnover 21.7x.
- Change m5 +7.8% / h1 **-39.2%** / h6 +1,137% / h24 +743%. X posts put the intraday peak near $700K to $778K mcap around 22:30 UTC, so roughly **40% off the high** with the pullback still live. GeckoTerminal has no OHLCV for this pool, so no candle-verified ATH exists yet: source gap, stated.
- Txns h1 129 buys / 144 sells; h6 2,177 / 1,865; h24 4,455 / 4,013.
- **Holder ledger reconciles exactly** (positives sum to 1,000,000,000, zero address at minus supply), 2,660 addresses from 21,627 Transfer logs.
- **`0x267444d099b10fb5ed7c3cc7b7c767adca574952` holds 81,632,653 = 8.16%.** This is the same **Pons V2 `V2LaunchLocker`** recorded in [[onchain-forensics]] on BOW at the identical 8.16%: no ERC-20 exit path, structurally immobile. Credit it.
- `0x8366a39cc670b4001a1121b8f6a443a643e40951` (v4 PoolManager) holds 8.07%, i.e. the pooled float.
- **No whale.** Largest real holder 2.24%, then 2.09%, 2.09%, 2.00%, 1.93%, 1.75%, 1.41%, 1.38%, 1.26%, 1.22%. **Top 10 non-infrastructure = 17.37%.** Clean for a 9-hour-old launch.
- **The operator holds ZERO MARLIN** in `0x1350bbb6d6ed3f60053a00caeddee6cd82bb8997`. His take is the ETH fee stream, not a token bag, so there is no supply overhang from him. That cuts both ways: he has no token exposure to align him.

## X AND COVERAGE
- **Tier-one coverage is ZERO.** Direct `from:` queries on blknoiz06, theunipcs, pingucharts, CrashiusClay69, 0xkioto, _0xghost_, metaversejoji and dontbuytops all returned 0 posts.
- The visible bid is a Telegram-call layer and small accounts: @ussharmony (15,779f) "RH needs a good dex, and this is it", @rantaiuang (1,710f) the Meteora comp, @meliboi_sama (12,245f) a neutral description, @insiderdigi (999f) "5x cook from my tg call", plus the usual "20x in 6 hours" posts.
- **@treyerl (15,475f) is the honest one: "don't fully get why it's moving, saying that plainly."**
- @JTMarlinTrade is 268 followers with its first MARLIN post on 4 Sep ("Old vision's out. New one: fees to protocol-owned LP"), so the product was pivoted three days before the token existed.

## RECHECK, 8 Sep 2026 00:40-00:55 UTC. FIVE CORRECTIONS TO THE FIRST PASS.
1. **TIMESTAMP WAS WRONG.** The first pass was stamped "~04:30 UTC 8 Sep". The chain head was 1788827308 = **00:28:28 UTC 8 Sep**. Everything else in that pass stands; only the label was wrong. It matters because the daily 18:00 UTC payout window had **not come due at all** at screen time (the 80% split was announced 19:57 UTC 7 Sep, after that day's window), so "never paid" was true but was not yet a broken promise. First real test is 18:00 UTC 8 Sep.
2. **ATH WAS WRONG, and the number that misled me was the promoters' own quoted market caps.** I wrote "peak ~$700K-$778K, ~40% off" from X posts because GeckoTerminal had no candles at the time. **It has them now: ATH $0.00119229 = $1,192,294 mcap at 22:00 UTC 7 Sep, and the token is -59.7% off it.** This is exactly the failure my own standing rule forbids (tops and bottoms from intraday highs, never from someone's quoted number).
3. **YIELD WAS A GUESS, now measured.** I gave "~10% to 33% annualised". Computed properly from each position's share of the pool's in-range liquidity: AMC 1.44% of pool L, HIMS 1.16%, SGOV 0.06% -> **$241.69/day gross, $193.35/day to holders = 14.71% annualised on a $479,818 cap. The book itself earns 242.9% APR** because the bands are tight and all three are in range.
4. **BOOK SIZE upgraded from cash accounting to direct mark-to-market: $36,316** (AMC $11,248 · HIMS $20,228 · SGOV $4,839), all three IN RANGE, positions unchanged and still owned by the desk EOA. The team's "~$36K" is accurate.
5. **HOLDERS: 812**, not "2,660 addresses" (that was addresses ever touched). Verified sweep, 21,762 logs, reconcile exactly 1.0. Top 10 ex-infrastructure 17.37%, unchanged.

### THE TECHNICAL READ THE CORRECTED ATH FORCES, per [[pullback-comeback]]
- **This is a blow-off, 2.7 hours old, not a base.** The 22:00 hour alone did **$852,043 = 57% of all lifetime volume**, then 23:00 $92,912, then 00:00 $33,834. Volume is down 96% from the peak hour.
- **FIRST PUMP 22x to 88x** (15:00 low $0.00001355 / open $0.00005340 -> ATH $0.00119229). The study's single strongest tell says **3.2x-3.7x comes back and 4.8x-5.3x dies**. MARLIN is 5x to 20x beyond the "dies" median. This is the worst possible reading on the one feature that works alone on both data sources.
- **63% of all volume ever traded sits ABOVE the current price.** Trapped supply overhead, measured, not asserted.
- RSI(9) hourly **68.7**, not oversold, no divergence available on 10 candles.
- Launch-study placement is equally bad: hour-1 volume **$304,783** puts it in the "over $100k" bucket (13.0% 2x, 6.5% 10x, median 1.01x) and it was already up ~4x inside hour one, so it fails checks 3 and 4 of the four 60-minute checks.
- Factory `0x3711cea4feade896c913c68f01eda97cb06d1a42` (Pons V2 token factory) is **absent from the 199-row `dep_table.json` scorecard**, so it has no track record number and grades "brand new / unrated", base 4.93% 10x.
- **Conclusion on the zone, which is what Ben's standing rule asks for: the -60% band cannot be called a buy zone here because no base exists yet.** The study's decision point is five days after a base forms and the median run takes 21-24 days from the low. At 2.7 hours past a 88x blow-off with volume collapsing, this is mid-dump, not reaccumulation.

### UNCHANGED ON RECHECK
- All three v3 positions intact, same liquidity, `ownerOf` still the desk EOA `0x2Faf625bF5bfB6b09deEE3b1B62D142ADF4578cb`. No `decreaseLiquidity`, no NFT transfer.
- **Still zero payouts.** Desk still at 36 transactions, all approves, swaps, mints and increases. No distribution contract.
- Fee claims still 21.3796 ETH total, none since 23:39 UTC. At $2,491.62/ETH that is **$53,268 claimed, $39,924 forwarded, $13,352 retained by the operator**. Measured fee rate 3.56% of volume.
- Operator and desk both hold **zero MARLIN**. Fee escrow holds 1,101.73 ETH (all projects, not MARLIN's).
- Tier one still **zero** on all eight accounts. Team account 268 -> 272 followers in 45 minutes.
- Net buying since the first pass: PoolManager MARLIN balance 81.6M -> 74.6M while price went $0.0004386 -> $0.0004798 (+9%).
- Wording drift worth noting: the 17:31 post said "20% of the fees go to holders"; by 19:13 it was "20% of the fees **it** makes" (the LP) and by 19:57 "80% of the **LP** fees". The headline number quadrupled while the base it applies to narrowed.

## SOURCES THAT FAILED, stated
- **Dune: out of datapoints for the billing cycle**, so `seller_ledger.py` could not run its SQL. The venue-set and holder work was done instead from 21,627 RPC Transfer logs with a full supply reconciliation, which is the same method against a different source.
- **GeckoTerminal has no pool or OHLCV record for MARLIN**, so there is no candle-verified ATH.
- Blockscout v1 `txlist` and v2 `/addresses` both intermittently 429'd and returned "Something went wrong"; every figure above came back on retry.
- fomo.family and purealpha were not checked this pass (Mac bridge was timing out on ssh during the run).