The Future of Money
The Dollar That Forgets It's a Bank
Stablecoins were supposed to solve the volatility problem in crypto — but in solving it, they accidentally recreated the most dangerous feature of traditional banking: the bank run.
The Idea
A stablecoin is, at its simplest, a promise. You hand over a dollar, the issuer mints a digital token worth exactly one dollar, and you can redeem that token for a dollar at any time. This sounds trivially safe. The danger is in the word 'any'. That promise of instant, universal redemption is precisely what makes banks fragile — and it turns out stablecoins are no different. The classic bank run works like this: a bank lends out most of its deposits, keeping only a fraction in reserve. If depositors lose faith and all try to withdraw at once, the bank cannot pay them all. The fear of insolvency becomes self-fulfilling. Banks solved this, imperfectly, through central banks acting as lenders of last resort, deposit insurance schemes, and reserve requirements that limit how recklessly money can be lent out. Stablecoins recreated the architecture of banking without inheriting most of those safety mechanisms. The fiat-backed variety — like USDC or Tether — hold reserves in cash and short-term government bonds to back every token one-to-one. That sounds like a narrow bank: fully reserved, supposedly run-proof. But even a fully reserved institution can face a run if its reserves are locked, inaccessible, or concentrated with a single counterparty. And the algorithmic variety — which tried to maintain a peg through software logic rather than actual collateral — contained a more fundamental flaw: the peg mechanism itself could accelerate a collapse. The Financial Stability Board has repeatedly warned what researchers had long argued: even fully reserved stablecoins can face destabilising runs if the backing cannot be delivered to the right place at the right time.
In the World
Two collapses, one decade apart in structure but only a year apart in time, illustrate both failure modes with uncomfortable clarity. In May 2022, TerraUSD — an algorithmic stablecoin that held no dollars in a bank account — imploded over the course of a week, wiping out roughly 40 to 45 billion dollars in combined market value. UST maintained its peg through an arbitrage relationship with LUNA, the ecosystem's native token: burn UST to mint LUNA, or vice versa. The Anchor protocol sweetened the deal by offering depositors a yield of around 19.5%, drawing enormous inflows. On May 7, 2022, large sell orders pushed UST below its dollar peg. As the price slipped, the stabilisation mechanism kicked in — but it ran in reverse: more UST redeemed for LUNA meant more LUNA minted, which crashed LUNA's price, which made UST worth even less. The mechanism designed to defend the peg accelerated the collapse. Research later confirmed that sophisticated investors who left first suffered far smaller losses than those who hesitated. Then came the quieter lesson. In March 2023, Circle — the issuer of USDC, one of the most transparent and conservatively managed stablecoins — disclosed that roughly 3.3 billion dollars of its cash reserves, about 8% of the total, were held at Silicon Valley Bank when regulators seized it. USDC fell to around 87 cents on some exchanges within hours. The peg recovered within days, only after the US Treasury, Federal Reserve, and FDIC jointly guaranteed all SVB depositors. TradFi risk had been hiding inside crypto's dollar all along — and only a government backstop fixed it, the very institution stablecoins had been designed to route around.
Why It Matters
The stablecoin market hit a record of roughly 322 billion dollars in May 2026, with dollar-pegged tokens accounting for well over 99% of that total. This is no longer a niche experiment. Major banks have taken note: JPMorgan Chase launched a deposit token for institutional clients in late 2025 and by July 2025 its CEO confirmed the bank would pursue stablecoin involvement, with Citigroup executives signalling the same intent. The GENIUS Act, signed into law in July 2025, created the first US federal licensing regime for payment stablecoins, requiring one-to-one reserve backing with cash or short-term Treasuries. What makes this worth sitting with is the deeper irony: stablecoins are now being regulated into resembling the banking system they were conceived to challenge. Full reserve requirements, redemption guarantees, capital buffers — the vocabulary of prudential banking is colonising crypto. And the financial stability question is running in both directions: Tether alone held approximately 141 billion dollars in US Treasuries as of March 2026, confirmed by an independent audit, ranking it among the largest single holders of American government debt anywhere in the world. The oldest problem in banking — that a promise of instant redemption and productive deployment of funds cannot both be kept under stress — has not been solved. It has been ported to a new medium, at scale, and dressed up in new terminology. Understanding that is more useful than any amount of crypto enthusiasm or crypto scepticism.
A Question to Ponder
If the solution to the bank run problem in stablecoins turns out to be government guarantees and central bank oversight, what does that tell us about whether trustless money is actually possible — or whether trust just shifts to a different institution?
One reader's Thinkable for Finance Friday — a random draw from the library.
Start learning with Thinkable