Where did pump.fun's $100M runners go? Month by month, and the pools that print them

Since PumpSwap, coins that really held $100M went from 4 in 6.7 months to 1 since BOOST. The raw count says 20 since BOOST, because 19 are bought-out pools: one buy takes 99% of the pool and the few tokens left set the price.

The long view · pumpdata.fun · published October 28, 2026 · 4 min read · markdown

Quick answer

Fewer pump.fun coins really reach $100M than a year ago, and the raw numbers hide it. We count a coin as holding a level after 12 hours at or above it, each hour with 50 or more traders, and $5M traded. By that rule, 4 coins held $100M in the 6.7 months from PumpSwap to the Oct 10 crash, 4 in the 9.3 months from the crash to BOOST, and 1 since BOOST: ANSEM. Count every coin with a $100M hour instead and the BOOST months look like a boom: 1,796 coins launched July to October 2026 touched $100M. 1,781 of them are one shape, a bought-out pool, and their price is set by trades in the few tokens left in a nearly emptied pool.

What the thread argues

LexaproTrader's thread says the $100M drought isn't one event. Not TRUMP draining attention in January 2025, not the Oct 10 crash, but a year of pump.fun's own choices: PumpSwap, AI market-making, a small circle of winners. It calls the last three $100M coins (ANSEM, Penguin, Remus) highly controlled, with PVP copies, ladder bots and airdrops to fake accounts.

We can't see intent on the chain. We can see every trade, so here is what the chain shows, period by period.

Coins that held $10M and $100M

period	months	held $10M, other shapes	held $10M, bought-out pools	held $100M, other shapes	held $100M, bought-out pools
PumpSwap to the crash (Mar 20 to Oct 9, 2025)	6.7	67 (10.0 a month)	9	4	3
the crash to BOOST (Oct 10, 2025 to Jul 20, 2026)	9.3	48 (5.2 a month)	10	4	2
BOOST on (Jul 21 to Oct 27, 2026)	3.3	23 (7.0 a month)	20	1	19

The $10M rate halved after the crash and came part of the way back after BOOST. The $100M count went 4, 4, 1. These are single digits, so treat the trend with care. But nothing in the record looks like the spring of 2025, when House and RFC held $100M for weeks.

Coins that held $1M are steadier: about 67 a month after PumpSwap, 49 after the crash, 58 since BOOST.

The pools that print $100M

Since BOOST, most "$100M coins" follow one script. The deployer buys the whole bonding curve in one buy of about 85 SOL and the coin migrates to PumpSwap in the same second. Then a single buy, typically 1,000 to 20,000 SOL, takes 90% or more of the tokens in the new pool. What's left is a few tokens in a pool full of SOL, so every small trade prints an enormous market cap.

AROS (kFWgzdFBd6DERYGSWUrT3ZUUjoREcv14rHJCBSbpump) is the extreme: in its best hour its price said the coin was worth $7,851M, nearly eight billion dollars, yet only $10M ever traded in it. In the $100M hours of the September and October coins, the median trade was 0.0003 SOL. Of the 20 coins that held $100M since BOOST, 19 are this shape. A chart of $100M coins that doesn't filter them out is mostly charting these pools.

What the real ones looked like at launch

For coins that held $10M and aren't bought-out pools, the launch changed (medians by period):

period	supply bought in the first 10 s	of that, sold by the launch wallets within 24 h	same-ticker launches within 24 h	traders in the first 24 h
PumpSwap to the crash	25%	80%	11	4,702
the crash to BOOST	51%	100%	19	3,508
BOOST on	79%	100%	23	15,190

79% is the whole curve. Since BOOST, the median coin that became a real runner had its entire curve bought in its first 10 seconds, and the wallets that bought it had sold all of it within a day. The copies with the same ticker roughly doubled. That's the "less room for anyone outside" the thread describes, measured, without a word about who.

The three coins in the thread

ANSEM (9cRCn9rGT8V2imeM2BaKs13yhMEais3ruM3rPvTGpump) held $100M for 2,474 hours, with a best hour of $429M. Its deployer bought the whole curve and it migrated in the same second, but its biggest pool buy took 13% of the pool, so it isn't a bought-out pool. 550 coins with its ticker launched within a day of it.

PENGUIN (8Jx8AAHj86wbQgUTjGuj6GTTL5Ps3cqxKRTvpaJApump) held $100M for 44 hours, with a best hour of $154M. 23 wallets bought 20% of the supply in its first 10 seconds, it migrated after 3,956 seconds, and those wallets had sold everything they bought within an hour.

Remus never held $10M by our rule. The biggest, "Remus, Official Fomo Mascot" (aLVHdrVSpzWhBorVncNKrJ5SASkDmjHYp6k7N1Hpump), launched Aug 9, 2026: its deployer bought the whole curve, it migrated in the same second, and one pool buy of 3,912 SOL took 98% of the pool. Its best hour was $96M, held for no full hour at $100M. It's a bought-out pool, and so were the other three Remus coins that came near $100M that weekend.

What this record can't say

Before PumpSwap, graduates moved to Raydium, which this record doesn't hold, so the 2024 era of big runners isn't in these counts. Counts of $100M coins are single digits in every period. And coins launched in late October have had less time to hold anything.

How each number is made: methodology. Every coin above has its own check: paste the address at pumpdata.fun/check.

Every number here comes straight from the blockchain, the same data behind the paid calls: the calls ($0.005 to $0.01 each, USDC over x402, no account), yesterday's /stats and the daily wire are free.