The Next Run on the Bank Won't Look Like a Bank Run

There is a seductive elegance to what the United States is constructing around stablecoins. America has an enormous and growing financing requirement; traditional foreign buyers of Treasury debt are changing their behavior, and the composition of Treasury ownership is shifting toward private investors, leveraged funds, money-market funds, and other price-sensitive buyers.

At the same time, the country has created a new class of institution that could become a structural buyer of government debt: the stablecoin issuer.

Important

The GENIUS Act formalizes payment stablecoins as privately issued digital dollars backed by highly liquid dollar assets, including short-term U.S. Treasuries. The mechanism it creates is beautifully simple. A person anywhere in the world wants a digital dollar, so they buy a stablecoin. The issuer receives the dollar and buys reserve assets, and those reserve assets include Treasury bills.

Global demand for digital dollars → stablecoin issuance → Treasury demand.

The person in Argentina, Turkey, Vietnam or Nigeria does not have to decide to finance the United States government. They merely have to decide that they want dollars, and the issuer performs the transformation automatically.

This could be one of the most important changes to the architecture of the dollar system in decades. It could also contain the seeds of its next crisis.

The disappearing buyer

Start with the Treasury market, where China provides a useful illustration of how the buyer base is changing. Chinese commercial banks have recently been purchasing U.S. Treasuries after raising rates on dollar deposits, and Chinese foreign-exchange deposits reached approximately $1.18 trillion by July, increasing about $121 billion during the first seven months of the year.¹

The economics are straightforward. Domestic Chinese yields are extremely low, so Chinese banks can attract dollar deposits at higher rates and deploy those dollars into higher-yielding U.S. securities. Meanwhile, China's Treasury holdings recorded through U.S. custodians fell to roughly $633 billion in June, their lowest level since 2008. Japan holds $1.12 trillion, and the United Kingdom, at $940 billion, now sits ahead of China.²

Hedge fund Treasury exposures doubled to roughly $4.0 trillion by late 2025 — $2.4 trillion long, $1.6 trillion short — with their share of outstanding Treasuries rising from about 4.5% to 8.5%. Nearly half of those long positions are leveraged arbitrage: the cash-futures basis trade and swap-spread trades. Net hedge fund repo borrowing reached about $1.8 trillion at the end of 2025, roughly 6% of outstanding marketable notes and bonds.³⁴

Note

For decades, foreign central banks accumulated Treasuries because they accumulated dollars. They were not asking whether 4.2% rather than 4.7% represented a compelling investment; they needed dollar reserves. The buyers who replaced them are yield-sensitive, levered, and financed overnight. Treasury now competes for capital, and capital has a price. That price is the yield.

Enter the digital dollar

This is where stablecoins become enormously important. Treasury's own advisory committee has modeled stablecoin issuers' Treasury-bill holdings potentially rising from more than $120 billion in 2024 to roughly $1 trillion by 2028.⁵

That isn't a crypto story. It is a sovereign-financing story.

Imagine several trillion dollars of dollar-denominated stablecoins circulating globally. Every additional dollar of supply requires corresponding reserves, and if a large percentage of those reserves flow into Treasury bills, the United States has created something remarkable:

a global, private-sector distribution system for Treasury demand

Instead of persuading foreign central banks to accumulate American debt, America allows private companies to distribute digital dollars around the world. Users create the demand for dollars, and regulation converts much of that demand into demand for government securities.

It is almost monetary alchemy, and it may work extraordinarily well.

That's the problem

Financial crises rarely originate in systems that obviously don't work. They emerge from systems that work so well that everyone begins depending upon them.

Stablecoins currently represent about $302 billion, with Tether at $183 billion and USDC at $74 billion. The market sits roughly 4.5% below its May 2026 peak and has contracted slightly over the past ninety days.⁶⁷ The flatness is more revealing than growth would be. Stablecoin supply has decoupled from the crypto price cycle: it fell more than 30% during the last bear market, but it has held near record highs through the 2026 downturn, up 14.3% year over year.⁷

That is what a monetary instrument looks like as it stops being a speculative one.

Caution

At $300 billion, a major issuer failure would be painful, mostly for crypto markets. At $2 trillion or $3 trillion, the situation changes entirely. Stablecoin issuers would no longer merely participate in the Treasury market — they would become part of its plumbing. And the mechanism that creates Treasury demand during the boom works in reverse during the bust.

We keep looking at the wrong side of the balance sheet

The GENIUS Act attacks one of the obvious weaknesses of early stablecoins: questionable reserves. That's good policy, and a stablecoin backed by short-term Treasuries is vastly safer than one backed by opaque commercial paper, speculative assets or unsecured obligations.

But safe assets do not automatically produce safe liabilities. Banks learned this centuries ago. A bank can own excellent assets and still collapse if its liabilities are redeemable immediately and everyone asks for their money simultaneously.

Federal Reserve Bank of New York researchers have already measured the behavior empirically. Stablecoin holders exhibit the same flight-to-safety dynamics as money-market fund investors, moving from riskier to safer coins during stress, with a discrete break-the-buck threshold around $0.99 below which redemptions accelerate. Those flows tend to stay within blockchains, the way money-market flows stay within fund families.⁸

The asset side can be pristine. The liability remains runnable.

The run does not need speed. It needs congestion.

Here the newest research becomes genuinely uncomfortable. In June 2026, Federal Reserve economists published a model showing that runs can occur even when digital money is backed by perfectly safe reserves; the paper was updated on August 31.⁹¹⁰

The mechanism has nothing to do with reserve quality.

Digital money becomes more valuable as adoption rises — a network externality — while blockchain transaction fees move in the opposite direction, rising as the network gets busy and making the token more expensive to move. Push fees high enough and small payments stop making sense. The token loses utility, holders gain reason to redeem or migrate to another chain, and those decisions reinforce each other.

The result is a run produced entirely by the cost of transacting.

This reframes the popular intuition about digital bank runs. The usual version says redemptions move at software speed: there is no line around the block, no bank closing at 5 p.m., no weekend, only an API. That's half right, and it's the less dangerous half.

Exit intent forms at software speed; exit capacity does not.

The moment everyone wants out is precisely the moment the network is most congested, and the cost of moving is highest. Fees spike, settlement slows, and the run accelerates because leaving has become expensive rather than because it has become easy.

Important

The GENIUS Act requires one-to-one reserve backing, public redemption procedures, fee disclosure, and regular reporting. It sets no capacity or price standard for the public blockchains that carry the tokens, and Treasury's proposed rulemaking in August 2026 also targets issuers rather than network fees.¹⁰ The reserves are regulated. The rails are not.

The repo loop

The redemption itself is only the first move. Follow the cash.

An issuer facing redemptions needs dollars immediately, so it sells bills, lets repo run off, or both. Reserve composition has already been shifting in that direction: after Silicon Valley Bank's failure, one of the largest issuers cut the average maturity of its reserves, surged into repo concentrated at FICC, and moved its bank deposits toward the largest institutions — trading interest-rate risk for counterparty risk.¹¹

Now place that against the leveraged Treasury complex described above.

Hedge funds finance $2.4 trillion of long Treasury positions with roughly $1.8 trillion of repo borrowing, most of it overnight, on thin haircuts and low futures margin.³⁴ A stablecoin redemption wave withdraws cash from the short-term funding market at the exact moment the basis trade needs it rolled.

Redemptions → issuer liquidation → repo cash withdrawn 
      → funding rates widen → margin pressure on levered longs 
      → forced Treasury selling → wider spreads → more redemptions.

That is not a crypto contagion channel. That is the March 2020 channel with a new trigger attached.

By some estimates, hedge funds now hold about $2.6 trillion of outstanding Treasuries, more than 8% of the market — a less stable equilibrium in which small yield moves can trigger margin calls and forced deleveraging.¹² Stablecoins would be supplying short-term cash into a market already structured around leverage that must be refinanced every morning.

The boom makes the bust larger

Suppose stablecoins grow from roughly $300 billion to several trillion. That produces enormous Treasury demand, and the Treasury sees the demand and rationally responds: more bills can be issued because a seemingly structural buyer is waiting for them. Stablecoins in turn become more useful because liquidity improves. More wallets integrate them, more international users adopt digital dollars, more financial institutions build products around them, and every step makes the ecosystem appear safer.

Adoption → liquidity → confidence → adoption.

That's the boom. Network effects work in both directions.

Redemptions → asset liquidation → market stress → fear → more redemptions.

That's the bust. And by then the stablecoin market isn't sitting outside traditional finance. It is woven through it.

The hidden bank run

There is another consequence that receives much less attention: stablecoins compete directly with bank deposits.

Suppose $100,000 sits in a bank. The traditional chain runs

deposit → bank funding → private credit

Move that money into a stablecoin and the chain becomes

deposit → stablecoin → Treasury

Capital has moved from financing private credit toward financing the federal government.

This is no longer theoretical. New York Fed research on stablecoin disintermediation finds that partner banks hold more reserves and that their loan share contracts relative to peers.¹³ The Bank for International Settlements has shown that the damage route depends on reserve design: if issuers hold bank deposits, retail funding is replaced by concentrated wholesale money; if issuers hold government bills, banks lose high-quality liquid assets; if issuers could hold central-bank reserves, conversion would strip banks of both deposits and reserves.¹³

Each route weakens bank liquidity through a different mechanism.

Banks losing deposits have choices. They can pay depositors more, seek wholesale funding, or shrink lending. So the policy could simultaneously make federal financing easier while making portions of private credit more expensive.

That's an extraordinary trade: the government gains a structural buyer, and the banking system loses a structural source of funding.

And then there is the maturity problem

Stablecoins do not solve the entire Treasury-demand problem, because they overwhelmingly favor the shortest end of the curve. They can create enormous demand for 30-, 60- and 90-day paper without creating buyers for 10-, 20- and 30-year bonds. Treasury can respond by issuing more bills, which reduces pressure to issue long-duration bonds.

But a 10-year Treasury finances the government for ten years, while a three-month bill must be refinanced roughly forty times over the same period. The government has traded duration risk for rollover risk.

Important

As long as stablecoin demand keeps expanding, that looks brilliant. If it reverses, the funding structure looks very different — and it looks different every ninety days.

This is where today's Treasury market becomes important

The 10-year Treasury yields 4.79%, near its highest level since November 2023, and the 30-year sits above 5.25%.¹⁴¹⁵ Those levels change the hurdle rate for stocks, private equity, venture capital, real estate, and corporate borrowing. When the risk-free rate was approximately zero, investors were pushed outward along the risk curve. At these levels, they aren't.

So the United States faces two problems at once. It needs enormous quantities of financing, and the financing itself is becoming more attractive relative to competing assets.

Stablecoins solve part of the first problem while reinforcing the second.

The danger isn't that GENIUS fails

This is the part markets may eventually misunderstand. The interesting risk isn't what happens if stablecoins fail. It is what happens if they succeed spectacularly.

Imagine $3 trillion of regulated dollar stablecoins, ubiquitous in international commerce and settling instantly. Businesses hold them, consumers hold them, financial applications route through them. They own enormous quantities of Treasury bills, banks compete with them for deposits, and Treasury incorporates their demand into its financing decisions.

Caution

At that point, stablecoins are no longer crypto. They are part of the monetary system. And once something becomes part of the monetary system, its failure becomes everybody's problem.

We've seen this movie before

Modern finance repeatedly creates instruments designed to make capital safer and more efficient: money-market funds, mortgage securitization, repo, commercial paper. Each innovation solves a legitimate problem, each becomes increasingly interconnected, and eventually the system begins treating liquidity as permanent. Then something causes everyone to ask for liquidity simultaneously.

The mistake is rarely that the underlying innovation had no value. The mistake is assuming that because an asset is safe, the system built around it cannot run.

Stablecoins may extend dollar dominance, dramatically reduce international payment friction, create enormous structural demand for U.S. government debt, and make the dollar more important in a digital financial system rather than less. All of that can be true, and it can simultaneously create a new systemic vulnerability.

The trade

During the boom, the sequence runs:

digital dollar adoption → stablecoin growth → Treasury demand → dollar reinforcement

The same architecture contains its inverse:

congestion or confidence shock → redemptions → bill and repo liquidation 
        → funding stress → margin calls on levered Treasury longs
        → more redemptions

The irony is that regulation makes the boom larger precisely because it makes stablecoins credible enough for mainstream adoption. The United States may have found a remarkably effective way to turn global demand for digital dollars into demand for its debt. But the more Treasury depends on that demand, the more consequential the day becomes when those digital dollars run the other way.

The next crisis will not begin because the collateral was bad. It will begin on a morning when the collateral is pristine, the reserves are fully audited, gas fees on the largest stablecoin chain are running twenty times normal, and $1.8 trillion of overnight repo has to be rolled by noon.