Two business days is the exam stablecoins have to pass

Tron settles about $150 billion a week and charges seven cents a transaction to do it, which makes the chain the most-used stablecoin rail in the market. By the logic of this week's rulemaking, that position is badly exposed: the two decisions that will determine whether the volume survives are how fast an issuer must redeem and how much yield a platform may hand back to holders, both being written by bank supervisors who owe the chain nothing.

The stablecoin argument in Washington has moved past legality and into the exam that follows a charter, now that the GENIUS Act has put dollar tokens on a statutory footing and set a 120-day approval clock. This week produced the working parts of it. The Fed wrote a redemption standard into its draft rule at two business days, and the Fed and the OCC each wrote a stablecoin reward presumption into their own drafts, matching each other's construction and opening a sixty-day comment record in which Congress is no longer in the room; the Treasury, meanwhile, began selling the dollar-token model overseas through three agencies before the rulemaking that defines the product is finished.

Read together, the three settle what the business now is: a stablecoin is a spread franchise that has to fund redemptions at the speed of a wire and share reserve income at a rate a supervisor will bless, and the issuers built for payment volume instead of balance-sheet depth are the ones whose economics change most. Volume stops being a trophy and becomes an obligation, because every dollar of flow is a dollar that can be called on two days' notice.

A custody rule asks where the asset sits; a reserve rule asks what it is worth when everyone wants it back at once. Stablecoin issuers spent their first years answering the first question and are now being handed the second in draft form, with a comment clock attached and no legislative forum left to appeal to — a harder exam for a company whose product is a promise of instant liquidity backed by a portfolio most holders never see.

The 48-hour standard reaches the reserve book before it reaches the user experience, which makes it more consequential than any licensing rule: a stablecoin that promises fast redemption against short Treasuries and overnight repo is running a liquidity and duration position, and a two-day call forces the issuer to keep enough of the book in genuinely callable assets that the yield on the rest compresses. A bank sizes its buffer to its deposit base for the same reason, and a token that has never had to now has to do the arithmetic.

The stakes are easiest to see in the mix itself: an issuer holding short-dated government paper and repo can meet a two-day call without selling into a bad market and earns a modest spread doing it, while an issuer that chased yield to fund a richer reward earns more in calm weather and holds less that is callable in a bad one. The exam is built to make that trade legible to whoever supervises the charter. The redemption deadline and the reward presumption are two readings of one instruction: the reserve has to be boring enough to pass, and boring reserves pay less.

The reward presumption squeezes from the other end: if both the Fed and the OCC read the statute as capping how much value an issuer may pass back to holders, platforms that compete on yield fund that yield from a thinner spread, and the ones with a bank's cost of funds behind them, or reserves already assembled to bank grade, compete from the better seat. Yield to holders has been a marketing line; the drafts turn it into a compliance variable, and that is harder to manage: a firm can outspend a rival on rewards, and it cannot outspend a cap.

Distribution economics follow from there: if a platform cannot pay holders a competitive yield out of a constrained book, the fight moves to who can put the token in front of the most users at the lowest cost, which is a payments and banking contest. Issuers with a bank's balance sheet can subsidize distribution from elsewhere on the profit and loss statement, while the crypto-native firms competing on rewards have only the spread to spend; the draft rules, when they bind, transfer advantage from the marketing line to the balance sheet.

None of this is settled — these are draft rules and a sixty-day docket, and the comment record can move the presumption in either direction. What the drafts settle is what has to be modeled: an issuer that has not run a two-day redemption, and cannot, has less time than it thinks to rebuild a book that survives one. The GENIUS Act's 120-day approval clock means charter decisions will arrive in a tight window once they start, and the firms still treating liquidity as a treasury detail have the least runway.

The way to see the filter is to run the two loudest names through it. Tron and Qivalis are both dollar-adjacent settlement businesses with real distribution and no bank charter, and they fail the same test from opposite ends: Tron brings the volume — it settles someone else's dollar token, its throughput depends on a validator set that is concentrated, and a concentrated validator set is not a reserve a supervisor will recognize. Qivalis brings the structure — 37 banks and a pending license, and a trade-finance corridor with nothing settled in it. The rulemaking grades the reserve book, which leaves each holding half an answer to the only question being asked.