A stablecoin is only as solid as the bank accounts nobody has listed
US prosecutors have seized bank accounts at a payments firm working for Tether. Whatever that case turns out to be, it exposes a layer of stablecoin plumbing that no reserve rule reaches and no attestation shows.
A stablecoin is a digital token meant to be worth exactly one dollar, backed by real assets the issuer holds. Tether runs the biggest one. To turn tokens into actual dollars, issuers rely on outside payment firms holding ordinary bank accounts. US prosecutors have now seized the accounts of one such firm working for Tether.
The rules being written cover what backs a stablecoin, not the bank accounts the money passes through on its way in and out. Those accounts can be frozen overnight, and holders are not told they exist.
The hard part was never the token
Minting a stablecoin takes a keystroke. Freezing a bank account takes a prosecutor and a judge. That asymmetry is the whole story of this week’s news, and it is not a crypto story at all — it is a banking story wearing a crypto costume.
United States federal prosecutors have seized bank accounts belonging to Capstone, a payments business that works on behalf of the stablecoin issuer Tether and its sister exchange Bitfinex, according to the Financial Times. That is the extent of what is on the record in front of me. No charge is described. A seizure is not a finding of wrongdoing, and funds seized are sometimes returned. Assume, for the sake of argument, that everyone involved is eventually vindicated. The structural point survives intact.
What a payments agent actually does
A stablecoin is a token that promises to be worth one dollar, redeemable for one dollar. To honour that, the issuer needs two things: assets that are worth a dollar, and a way to move dollars. The first is the reserve — typically short-term government debt and bank deposits. The second is a chain of commercial bank accounts and third-party payment firms that take wires from large customers, convert them into newly minted tokens, and run the process in reverse on redemption.
That second layer rarely gets discussed, because it is boring and because nobody publishes it. Issuers that struggle to hold correspondent banking relationships in their own name often rely on intermediaries who hold accounts in theirs and act as agent. The dollars sit in those accounts for hours or days. They are not, in that window, in a reserve portfolio. They are in a bank account belonging to a company most holders have never heard of, in a jurisdiction whose prosecutors can freeze it without warning.
The rulebook is aimed at the stock, not the flow
This matters right now because American regulators are mid-way through writing the rulebook. Last week the Federal Reserve Board put out two proposals for payment stablecoin issuers it supervises: full backing with permissible reserve assets such as short-term Treasury bills, standardised capital requirements for credit and operational risk, risk management standards, rules for supervised firms that safekeep the assets backing stablecoins, and a tailored application process for banks that want to issue.
Read that list again and notice what it governs. It governs the pool — its composition, its custody, the capital held against it. It is a good list as far as it goes. But a seizure of a payment agent’s bank account is not a reserve problem. It is a flow problem. The pool can be immaculate, fully backed, custodied by a supervised trust, and the redemption queue still stops if the account that receives and sends the wires is frozen. Capital held against operational risk does not unfreeze an account. It just means someone can absorb the loss afterwards.
And the largest issuers in this market are not Board-supervised entities at all. The Fed’s framework reaches the firms that apply to sit inside it. The plumbing that moves the most dollars sits outside it, which means the one genuine lever left is disclosure — and nobody is pulling it.
Attestations show a photograph, not a pipeline
Reserve attestations tell holders what was in the pot on a given date. They do not tell anyone how many payment agents an issuer uses, which banks those agents rely on, how concentrated the flow is across them, or how much money is in transit at any moment. If one agent handles a large share of subscriptions, its loss is a single point of failure — and no outsider can size it, because no outsider knows the agent exists until prosecutors act.
That is a strange thing to tolerate in an instrument marketed on the strength of its transparency.
The case for leaving it alone
The strongest counter-argument is that none of this touches ordinary holders. Primary redemption — going to the issuer, handing back tokens, receiving a wire — is the preserve of large counterparties. Everyone else sells on an exchange at whatever the market offers, and those markets have kept prices near a dollar through worse news than this. The reserve is large and liquid. An agent’s accounts being frozen is an inconvenience to the issuer’s treasury operation, not a run.
It is a fair point and it is wrong about causation. The secondary price holds at a dollar precisely because a handful of professional firms believe they can take tokens to the primary window and get dollars back. They are the ones who use these agents. Break their confidence in that channel and the arbitrage that keeps everyone else whole is what breaks. The peg is not maintained by the reserve directly; it is maintained by belief in access to the reserve. Access runs through bank accounts.
The cheap fix
None of this requires new legislation. It requires a list. Naming the agents and banks through which fiat moves costs an issuer nothing but embarrassment, and tells the market exactly how concentrated the chokepoint is. The reason it is not published is that the answer would not be reassuring — which is the strongest argument for publishing it.
Tether should publish the identity of the payment agents and banking partners through which subscriptions and redemptions settle, and disclose within one business day any balance frozen, seized or otherwise unavailable. Firms that decline should be asked why by every counterparty that relies on the primary redemption window.
The Federal Reserve should use the comment period on its two stablecoin proposals to add a requirement it has so far left out: supervised issuers must name their fiat payment agents, report concentration across them, and notify supervisors of any seizure or freeze immediately. And the Department of Justice should state plainly whether the accounts it has seized held customer subscription or redemption money, because that single fact determines whether this is a corporate matter or a market-structure one.
What would change this view
Show me the chain and I will stop calling it opaque. If Tether publishes, within the next quarter, the identity of the payment agents and banks through which subscriptions and redemptions settle, plus a figure for any balances currently frozen or seized, the specific complaint in this piece is answered.
Equally, if court filings in the Capstone matter establish that the seized accounts held only the firm's own corporate funds, with no customer subscription or redemption money passing through them, then the operational risk I am describing is smaller than I claim and the case rests on disclosure alone.
Worked from
- US prosecutors seize Tether-linked payment firm's bank accounts - FT
Finextra · 2026-09-25