The argument, stated fairly
The game-theoretic case for national bitcoin reserves is simple enough to fit in a sentence, which is part of why it spread so fast. The supply is fixed. If one large state starts accumulating, every other state faces a choice between buying at today's price and buying later at a higher one, so the rational move is to buy early, so everyone buys, so the first mover was right. No one has to believe in bitcoin for this to work. They only have to believe that others might.
That argument is not stupid. It is the same structure that drives an arms race or a bank run, and both of those are real. It is also testable, and it has now been tested, because the thing it was waiting for happened.
The trigger fired
On 6 March 2025 the United States established a Strategic Bitcoin Reserve. The order capitalises it with bitcoin already forfeited to the government in criminal and civil proceedings and directs that those coins not be sold.
Read the mechanism rather than the headline and the move is smaller than it looks. An executive order can direct how the executive branch handles what it already owns. It cannot appropriate money, and the order says so itself: acquisition strategies must be "budget neutral", and implementation is "subject to the availability of appropriations". The United States did not buy any bitcoin. It stopped selling the bitcoin it had seized.
That distinction matters for the game, because the strategy the argument depends on is accumulate, and what actually happened was retain. A player who declines to sell is not bidding.
Eighteen months of moves
Congress has not supplied the missing half. The BITCOIN Act, which would direct the Treasury to buy a million coins over five years and hold them for twenty, was introduced on 11 March 2025 and has sat in the Senate Banking Committee since, with no hearing and no markup. The successor proposal introduced in May 2026, the American Reserve Modernization Act, is bipartisan and asks for a study of budget-neutral acquisition methods. The bill that would spend money stalled; the bill that would think about spending money is the one that attracted co-sponsors.
The states were supposed to be where the cascade started. New Hampshire authorised a treasury allocation in May 2025 and has made none. Florida's two reserve bills were postponed and then died in committee. Arizona's legislature passed one and Governor Katie Hobbs vetoed it, writing that the state retirement system was not the place for an untested investment. Texas appropriated 10 million dollars and spent it, in two tranches in November and December 2025, on shares in a bitcoin exchange traded fund (Bond Buyer, November 2025). One state out of fifty has put public money into the asset, and what it bought was a security.
Central banks moved least of all, and the exception proves the size of the move. In November 2025 the Czech National Bank became the first to buy any, and it said precisely what it was doing: a one million dollar test portfolio of bitcoin, a dollar stablecoin and a tokenised deposit, approved to gain practical experience with custody, approval workflows and compliance, and explicitly separated from the international reserves. One million dollars, against reserves measured in the hundreds of billions. That is not a position. It is a procurement exercise.
And the one country that went furthest went backwards. El Salvador amended its Bitcoin Law in January 2025, removing the acceptance obligation, the tax channel and the state's convertibility guarantee, as a prior action for an IMF programme. The first mover in the sovereign game traded part of its position for a credit line denominated in dollars.
Why the prediction failed
The argument was not wrong about scarcity. It was wrong about who is playing.
"The United States" does not have a utility function. A comptroller does, a treasurer does, a central bank governor does, and their payoffs are not the state's. A treasurer who allocates 5 percent of a general fund to bitcoin and is down 20 percent at the next budget hearing has a career problem that is not offset by the possibility of being right in 2040. A treasurer who does nothing has no problem at all, because doing nothing is what the job was designed to reward. The asymmetry is not irrational. It is the actual incentive structure of custodial public office, and it points the other way from the one the essay assumed.
Add the horizon. The scarcity argument pays out over decades. Legislative terms are two years, gubernatorial terms four, and appropriations annual or biennial. A payoff outside the decision-maker's tenure is, for game-theoretic purposes, not in the matrix.
Add the reversibility. Buying bitcoin is a decision that can be examined, priced daily and attributed to a name. Not buying leaves no record. Strategies with visible losses and invisible opportunity costs get selected against, whatever the expected values say.
The result is the pattern in the evidence: authorisation is cheap and common, allocation is expensive and rare. Passing a law that permits a reserve costs a legislator nothing and signals plenty. Writing the cheque costs somebody their reputation if it goes wrong.
The strongest version of the other side
Eighteen months is not a long time, and this is the honest objection.
Arms races are slow before they are fast. The relevant precedent is not "did everyone move in a year" but the shape of institutional adoption generally, which tends to run through permission, then pilot, then position, over years rather than quarters. On that reading, everything above is exactly what stage one looks like: the legal question settled in several jurisdictions, one state through the procurement, one central bank through the custody test, and a federal reserve in existence with an acquisition instruction attached to it. Nobody has had to be convinced twice.
There is also a version of the argument that never needed states at all. If the mechanism is scarcity meeting demand, exchange traded funds and corporate treasuries are already doing the buying, and whether the buyer wears a flag is a detail. The essay this post originally was leaned on that idea, treating a listed company's balance sheet as a proxy for a national reserve. That was sloppy, and the difference matters: a corporation can be compelled to sell by its creditors, and a state cannot.
I take the first objection seriously and the second less so. But both concede the point this post now makes, which is that the cascade was asserted rather than observed, and that the interesting question is what would actually cause one.
What would change the answer
Three things, and none of them is a law.
A purchase with appropriated money, at a size that shows on a balance sheet, by a government that then says so. A custody arrangement where the state holds keys rather than shares, because that is the point at which a reserve stops being a portfolio allocation. And a second buyer whose stated reason is the first buyer, which is the only evidence that would show the mechanism working rather than being described.
None of the three has happened. Until one does, the reserve race is a plausible model with no confirming instance, and a model that has been available for eighteen months without producing its central prediction is not a forecast anyone should be trading on. The scarcity is real. The compulsion was imagined.
Correction, 20 August 2026. This post originally predicted that a United States Strategic Bitcoin Reserve would trigger a global accumulation race and stated that "game theory says the rest of the world will follow". No such race has occurred, and the post has been rewritten to test that prediction against what happened instead. It also described a listed company's bitcoin holdings as "effectively functioning as early strategic reserves for the U.S. economy", which is not accurate, and carried a table of expected national reactions and a table of fiat outcomes, neither of them sourced. Both tables and repeated links to a third-party legislation tracker have been removed.