Washington Bet the Treasury Market on Stablecoins, and Stablecoins Stopped Growing
The United States Treasury spent the past two years building a fiscal argument on top of a product whose supply is now going the wrong way.
Stablecoin supply has fallen roughly $15 billion from its mid-May peak — the sharpest contraction since TerraUSD detonated in 2022, and the first sustained shrinkage the sector has posted in four years, as Forbes flagged this summer. The difference from 2022 is that nothing broke. USDT and USDC both held a dollar the entire way down. The money simply left.
That is a stranger problem than a depeg, and a worse one for the people in Washington who were counting on the opposite.
The pitch was that crypto would buy the deficit
Treasury Secretary Scott Bessent has floated the idea that stablecoin issuers could grow into trillion-dollar buyers of US government debt — the sector expanding roughly tenfold by the end of the decade to something like $3 trillion, absorbing a meaningful share of the roughly $7 trillion in outstanding short-term bills. Bloomberg reported this month that the retreat is now testing that thesis directly.
The mechanism was elegant on paper. A compliant dollar token has to be backed by short-dated, liquid, boring collateral. Bills qualify. So every dollar that migrates from a bank deposit into a regulated stablecoin becomes, at one remove, a bid for Treasury paper. Tether alone reported around $141 billion of direct and indirect bill exposure as of the end of March. Structural demand, manufactured by regulation, at precisely the moment the government needs to roll a great deal of short-term debt.
The Bank for International Settlements has studied the same transmission channel and found the effect real but modest, with inflows compressing short-tenor yields by a few basis points. Real, in other words, and nowhere near the size the fiscal narrative implies.
But the fiscal narrative was never really about basis points. It was about a story: that a new, politically friendly class of buyer was arriving to take down supply that foreign central banks were no longer eager to absorb.
The law that was supposed to grow the sector is the thing capping it
Here is the part that deserves more attention than it gets. The GENIUS Act, which takes effect on January 18, 2027, forbids payment stablecoins from paying yield to holders.
That provision exists for an obvious reason. A yield-bearing dollar token is a money market fund wearing a costume, and the banking lobby made a credible case that permitting it would drain deposits out of regional institutions at speed.
The consequence is just as obvious in hindsight. In a world with meaningful short rates, holding a stablecoin means voluntarily forfeiting the return on your own collateral so that the issuer can keep it. At three percent, that is a tolerable convenience fee. At five, it is an actively bad trade for anyone holding size.
So institutional cash did what institutional cash does. It moved into tokenized Treasury funds, which deliver the same on-chain settlement finality and pay the holder instead of the issuer. Same rails. Same collateral. Different answer to the question of who keeps the interest.
The Treasury Department wrote a rule that made stablecoins buy bills, and in the same stroke made large holders prefer a different wrapper for the identical exposure. The demand did not disappear. It relocated to an instrument that does not show up in the stablecoin supply figure anyone is watching.
Supply is the wrong number and everyone knows it
Defenders of the sector have a strong rebuttal, and it should be conceded. Supply measures idle balances. Volume measures usage. Those are different businesses, and only one of them is the actual product.
A dollar that sits in a wallet for a year contributes a dollar to market cap and almost nothing to the economy. A dollar that clears fourteen cross-border supplier payments in a month contributes the same single dollar to market cap while doing fourteen times the work. If stablecoins are succeeding as payment infrastructure rather than as a savings vehicle, you would expect exactly this: velocity rising while balances thin out, because nobody parks money in an instrument that pays nothing.
By that reading, the contraction is a sign of maturity. Speculative float is draining and working capital is what remains.
Fine. But that argument dismantles the Treasury pitch rather than rescuing it. Bills get bought out of balances, not out of velocity. A stablecoin that turns over thirty times a month and holds $400 million in float buys $400 million of collateral. The fiscal case required a large, slow, growing pile of idle dollars — precisely the thing the industry now says it never wanted to be.
You cannot simultaneously argue that stablecoins are payment rails rather than savings accounts and that they will absorb a trillion dollars of government debt. Those are opposite businesses.
Twenty-one banks are about to test it properly
The next act is already scheduled. Bank of America, Citi, Goldman Sachs, Deutsche Bank, UBS, Santander, MUFG and Fidelity are among twenty-one institutions planning to stand up a joint entity this half and issue a fully reserved dollar stablecoin in the first half of 2027, per crypto.news and confirmed across multiple accounts of the consortium's structure. Euro issuance is the stated follow-on priority.
The launch window is not a coincidence. It sits immediately after the GENIUS Act's effective date, which is when bank-issued tokens stop being a legal gray area and start being a licensed product.
Two possibilities follow, and they lead to opposite places.
In the first, the banks bring genuinely new balances on-chain — corporate treasury operations, settlement float, intraday liquidity that presently sits in accounts nobody tokenizes — and total supply grows without cannibalising Tether or Circle. Bessent's arithmetic starts to look less fanciful.
In the second, which the deposit-flight math favors, the consortium mostly converts liabilities the member banks already hold. A dollar moves from a Citi deposit into a Citi-affiliated token. The bank's funding profile changes. The Treasury's buyer base does not, because those deposits were already funding bill purchases through the banking system's own balance sheet. Tokenizing a claim does not create a new claimant.
The second scenario is the boring one, which is usually a reason to expect it.
What breaks the deadlock is a rate cut, not a regulation
Strip everything else away and the sector's balance growth is a levered bet on the front end of the curve. When short rates are high, forfeiting yield to hold a payment token is expensive and balances thin. When short rates fall, the opportunity cost of holding a non-yielding dollar collapses, tokenized bill funds lose their edge, and stablecoin supply expands again without anyone passing a law.
Which produces an unhelpful loop for the people making fiscal policy. The condition under which stablecoins buy the most Treasury bills is the condition under which the Treasury least needs them to. The condition under which the government most wants new buyers at the front end is the one that drives holders into the competing product.
Washington built a demand channel and then wired it backwards. Come January, twenty-one of the largest banks on earth will find out whether that wiring can be fixed by issuing more of the thing, or whether the market has already answered by choosing the version that pays.