Methodology — the whole chain, so you can attack any link
What this is and isn't: this clock counts gross sell-side outflow — dollars leaving pump.fun positions — allocated geographically. It is a measure of the scale of the casino, not of net retail losses. Read footnotes [9]–[11] before quoting the big numbers as "damage": the honest net figure is roughly 20× smaller.
[1] Launch date. Pump.fun launched January 19, 2024. The clock's baseline anchor is set at Jan 1, 2026 (see [3]) and the tick extrapolates forward from there.
[2] Fees. Cumulative protocol revenue crossed $1 billion by March 2026 — the first Solana app to do so (Yahoo Finance). 2025 full-year revenue was ~$971M; the 2026 run-rate is ~$320M/yr as of April (CoinDesk).
[3] Volume anchor. Cumulative lifetime trading volume passed $150 billion during 2025 (Phemex); we anchor $150B at Jan 1, 2026 and grow it at the selected run-rate. Cross-check: $1B cumulative fees at a blended fee of ~0.65% (1% bonding curve blended with 0.25–0.30% PumpSwap) implies ~$150B — the two public numbers agree with each other.
[4] Sell share. Every trade has two sides; over the life of a token that goes to ~zero (as ~98.6% do), roughly half of volume is sells — the dollars coming out. Default 50%, adjustable.
[5] US share. SimilarWeb (May 2026) attributes 47.3% of pump.fun web traffic to the United States (Similarweb). Web traffic understates total activity — the great majority of trades run through Telegram bots, Photon/BullX/Axiom and APIs rather than the site (~2.6M monthly visits vs. millions of on-chain trades per day) — so we assume off-site flow has the same geographic mix as on-site. That is the load-bearing extrapolation; adjust it below if you think bot traffic skews elsewhere.
[6] New York share. NY population (~19.6M) over US population (~335M) = 5.85% of the US slice. This is conservative: crypto trading activity overweights the NYC metro relative to population.
[7] Comparator. New York State's announced $100 million law-enforcement modernization investment (Gov. Hochul, July 2026). The counter divides NY extraction by $100M.
[8] Tokens. ~12 million tokens created since launch as of March 2026 (Yahoo Finance); ~98.6% end worthless or rugged (reported analyses). Tick rate is a labeled estimate, adjustable.
[9] Gross sell volume isn't extraction — it's churn. A degen who round-trips $1,000 fifty times generates $25,000 of "sells" while maybe losing $300. The same SOL cycles through the casino hundreds of times. That's how you get $83B of "extraction" out of a system where nothing like $83B of outside money ever entered. The clock is counting the speed of the roulette wheel, not the money leaving players' pockets.
[10] It double-counts both sides of a zero-sum game. Every dollar a winner "extracts" came from another trader — including other New Yorkers. If you attribute all sells to NY by population share, you're counting NY's snipers extracting from NY's retail as damage to NY twice. Net damage to a population is (their losses − their winnings + fees they paid), and if NY traders are roughly average, that nets out to their share of the rake, not their share of volume.
[11] The rake is the real number, and it's ~20× smaller. In a closed trading system, aggregate net trader losses ≈ what leaves the table: platform fees (~$1B cumulative, verified), Telegram-bot fees (Photon/BullX/Trojan take ~1% of the bot-routed majority of volume — plausibly another $1–2B), validator tips/priority fees, and insider/sniper net profits (real but hard to bound; hand-wavy hundreds of millions to low billions). Call it $3–5B global lifetime net extraction. Run that through the same geographic chain: US ≈ $1.5–2.4B, NY ≈ $90–140M lifetime.
Run-rate modes
2026 run-rate (default, conservative): ~$320M/yr fees ⇒ ~$32B/yr volume at the same blended fee ⇒ ~$16B/yr gross sells.