The $73 Lottery Ticket: More Than Half of the TRUMP Coin's $3.81 Billion in Losses Were on the Books by Day 19

Two documents landed four days apart, and together they let us audit one of the most lucrative products a sitting American president has ever put his name on. On June 30, the White House released Donald Trump's annual financial disclosurefinancial disclosureThe annual form federal officials must file listing their income, assets, and business interests; it is a legal document, not a press release., a filing with the Office of Government Ethics that officials sign under penalty of law. Among its line items, reported by PBS NewsHour, Forbes, and other outlets: $636 million in 2025 income routed through CIC Digital LLCCIC Digital LLCA Trump-affiliated company that receives payments for licensing the Trump name; the disclosure attributes $636 million of 2025 income to it., the Trump-affiliated company that licenses his name for use in connection with the $TRUMP memecoinmemecoinA cryptocurrency created around a joke, celebrity, or moment, with no underlying business, revenue, or asset backing its price, it is worth only what the next buyer will pay..

On July 4, the New York Times published an analysis by the blockchainblockchainA public, shared ledger that records every crypto transaction permanently, which is what lets analysts reconstruct exactly who bought and sold, and when.-data firm NansenNansenA blockchain-data firm that analyzes public ledger records; it performed the wallet profit-and-loss analysis published by the New York Times on July 4, 2026., run through June 30 and reported the same day by The Block, that counts who paid for it. Of 1.48 million wallets that ever traded the coin, 988,905, two out of three, are down a combined $3.81 billion. The token stood near $1.78 at the analysis cutoff and traded at $1.69 on July 5, according to CoinGecko: about 98 percent below the $73.43 peak it touched on January 19, 2025, two days after launch and one day before the inauguration.

The easy story is that these buyers were casualties of this year's broad crypto drawdowndrawdownThe fall from a peak price to a lower point, usually given as a percentage; TRUMP's roughly 98 percent drawdown means the price lost about 98 cents of every dollar from its peak., the defense any promoter would reach for, and one with surface plausibility, since crypto prices broadly fell in 2026 and the token fell with them. But because every trade lives permanently on a public blockchain, this is one of the rare consumer-loss stories where the alibi can actually be tested against a ledger.

So we asked one narrow question: were the $3.81 billion in losses caused by the 2026 downturn, or were they already on the books during the token's first three weeks? The question is falsifiable because it is about timing. If the downturn did the damage, the losses should have accumulated when the downturn hit, in 2026. If the losses were structural, a transfer from late buyers to early buyers and to the coin's issuer, they should show up almost immediately, while the broader market was still fine.

Answering it takes two dated snapshots of the same public ledger, and both exist. The first was taken on roughly February 5, 2025, day 19 of the token's life, when the New York Times commissioned the blockchain-forensics firm ChainalysisChainalysisA blockchain-forensics firm used by governments and media; it produced the February 2025 loss snapshot and traced the creator fee flows. to run a full walletwalletA blockchain account that holds and trades crypto; one person can control many wallets, so wallet counts are not exact people counts. accounting; Fortune reported the results that week. The second is the Nansen analysis of July 4, 2026, which totals the same ledger through June 30. Set one against the other and the lifetime losses split into two windows: what was already recorded by day 19, and everything that accumulated in the 510 days after.

The split is not close. Chainalysis found 813,294 wallets down a combined $2 billion by early February 2025. That is 52 percent of the eventual $3.81 billion, recorded in the first 3.6 percent of the token's life to date, sixteen months before the 2026 downturn began. The remaining $1.81 billion accumulated over the following seventeen months, a window that includes the entire downturn.

Two honest caveats about that split. The snapshots come from different firms, whose accounting methods may differ at the margins. And the crowd churned between them: some wallets that were losers in February later sold, recovered, or were replaced by new losers, so the 52 percent is arithmetic between two totals, not a count of the same people. The true early share could sit somewhat above or below half. What it cannot be is near zero, and near zero is what the downturn-only explanation requires.

Rewind to February 5, 2025, and the scale of the early damage comes into focus. The token had launched on January 17, spiked to $73.43 within two days on a wave of inauguration-week buying, and then fallen to about $16, a 78 percent drop in under three weeks, at a time when the wider crypto market had not crashed. The $2 billion in wallet losses Chainalysis counted at that point were not the residue of a bear market. They were the mechanical result of hundreds of thousands of people buying near a two-day-old peak from people who had bought hours earlier.

The house was already collecting. By that first accounting, roughly $100 million in trading feestrading feesSmall charges collected automatically on each buy or sell of the token; the TRUMP coin's infrastructure routes a share of these to creator-linked wallets. had flowed to entities including CIC Digital and Fight Fight Fight LLCFight Fight Fight LLCA company co-owned with CIC Digital that, together with CIC, controls about 80 percent of the TRUMP token's total supply, which unlocks for sale gradually over three years., the two Trump-linked companies that together control about 80 percent of the token's total supply. Fortune's framing at the time: for every dollar of fees the creators took in, buyers had lost about twenty.

The $2 billion in losses on the books by day 19 were not the residue of a bear market. They were the mechanical result of buying near a two-day-old peak from people who had bought hours earlier.

The full-lifetime ledger explains why the early losses were baked in. A memecoin has no revenue, no asset, no cash flow; it cannot make the whole crowd richer. Every dollar a winner took out is a dollar a later buyer put in, minus the fees skimmed on the way through. That is what the Nansen totals show: $4.04 billion in gains across 492,285 winning wallets, $3.81 billion in losses across 988,905 losing ones, and a crowd-wide net of just +$236 million, a rounding error on the money that changed hands, and a paper figure at that, since unsold coins are marked at the June 30 price of about $1.78.

Who won? Per the Times, the winners share one trait: they bought below $1 in the token's first hours, before the price ran toward its peak. The gains were also intensely concentrated well before the downturn, a Chainalysis count reported by CNBC in May 2025 found just 58 wallets that had each cleared more than $10 million, for a combined $1.1 billion, while 764,000 wallets sat underwater. In lottery terms, the crowd bought tickets topping out at $73.43; a thin layer of first-hour buyers held the winning numbers; and the operator collected a rakerakeThe cut a game's operator takes from every bet regardless of who wins, here, the trading fees routed to the token creator's wallets on every trade. on every ticket regardless.

The rake is the part the president's own paperwork now documents. The June 30 disclosure reports $636 million in 2025 income through CIC Digital, the license vehicle for the coin. That one line is 2.7 times the entire crowd's net paper gain. Separately, Chainalysis had traced more than $324 million in trading fees flowing on-chain to creator-linked wallets by May 2025, fees the token's code routes automatically on every trade. Add the two and the documented Trump-linked take reaches roughly $960 million; but the two figures come from different ledgers, one legal and one on-chain, and could partially overlap if some traced fees are part of the disclosed license income. That is why we say 'up to.' The conservative comparison, disclosure line alone versus crowd net, is already 2.7 to one.

Is this loss structure specific to the TRUMP coin, or is it just what political memecoins do? There is a control case. In February 2025, Argentina's president Javier Milei promoted a SolanaSolanaThe blockchain network on which both the TRUMP coin and Argentina's LIBRA coin were launched; known for fast, cheap transactions that suit memecoin trading. memecoin called LIBRALIBRAA memecoin promoted by Argentina's president in February 2025 that collapsed within hours; the same analytics firm found 86 percent of its meaningfully sized traders lost money.; it spiked toward a multibillion-dollar valuation and collapsed within hours. Nansen, the same firm behind the TRUMP accounting, ran the LIBRA post-mortem. Among the 15,431 wallets whose gain or loss exceeded $1,000, 86 percent realized losses totaling $251 million, while 2,101 winning wallets took home about $180 million.

The comparison holds at the level of structure: a supermajority of meaningful traders losing, gains concentrated in a small early cohort, all of it settled long before any market downturn. It breaks at the level of scale and role. TRUMP's accounting covers roughly one hundred times as many wallets, its losses ran fifteen times larger, and its arc took seventeen months rather than a night. And where Milei promoted a coin others issued, conduct for which he faced impeachment calls and investigations, the TRUMP coin's supply and fee streams sit with companies that license the president's own name, with the income now itemized on his own federal disclosure.

Here is what the record supports. The president's signed disclosure documents $636 million flowing to his license vehicle from the coin in 2025. The public blockchain documents that the crowd who traded that coin is, in aggregate, roughly flat on paper, with two-thirds of wallets down $3.81 billion and more than half of that damage recorded in the first nineteen days, when the wider market was untouched. The 2026 downturn deepened the hole and dragged the price to $1.69, but it cannot explain the half of the losses that predate it by sixteen months, and it had no effect on the license income, which was earned on 2025 activity and reported before the Nansen totals were public.

Here is what the record does not support. Wallets are not people: one trader can hold many wallets, so 988,905 losing wallets is not 988,905 losing households. The loss totals mix realized cash losses with paper losses that would shrink if the price recovered. And the two ledgers, a legal disclosure and an on-chain trace, cannot be reconciled line by line from the outside, which is why the combined take is a range topping out near $960 million rather than a single number.

What remains is the timing, and timing is the whole story. A market downturn is an alibi that requires the losses to arrive when the market fell. On this ledger, they arrived first, before the downturn, before the disclosure, in the nineteen days when a $73 ticket became a $16 one.