The Operating Account Was Never Free

Open USD looks like an attack on Circle, but could it actually be a serious bid to buy back the float that has funded banking for fifty years?

Jul 05, 2026

On June 30, more than 140 companies announced they would back the same stablecoin. Visa, Mastercard and American Express in one consortium. Stripe, Adyen and Checkout.com alongside them. BlackRock, BNY, Standard Chartered, DBS, BBVA, U.S. Bank. Google, Shopify, Samsung. The vehicle is Open Standard, an independent company run by Bridge co-founder Zach Abrams, whose last business Stripe bought for $1.1 billion. The product is Open USD, launching later this year across Solana, Base, Stellar and Polygon. Circle’s stock fell 16% on the news, and the commentary settled quickly into a familiar groove. A challenger stablecoin, a threat to USDC, a fight between issuers.

I wonder if we might reframe it though. Perhaps a better question might be what product OUSD actually is, once you strip away the various chains and the governance language and look at what a business receives when it signs up.

So here is the proposition in a nutshell. You can mint and redeem at no cost, at any volume. You transact from the balance. And nearly all of the income earned on the reserves backing your balances flows back to you, less a small management fee. A dollar balance you make payments from, that pays you roughly the Treasury bill rate.

Of course, most bankers would have a name for that product. It’s an interest bearing corporate operating account. And the entire architecture of American banking, from a Depression era statute to a clause buried in last year’s stablecoin law, has been organized around making sure that product does not exist.

What 140 logos are actually asking for

Let’s start with the arithmetic, because I think the arithmetic is the product here. A stablecoin’s distributable income is fixed by what its reserves earn, and reserves earn as a function of dollars held over time. Yield accrues to float, not to flow. That sounds like it should kill the payments use case, but at corporate scale it doesn’t, because even transactional money has dwell time.

Ok, so let’s run the numbers on a back of an envelope for a treasurer pushing say, $1 billion a day through OUSD, with an average four hour gap between minting and the supplier’s redemption. That’s a time weighted balance of roughly $167 million, continuously. At, I dunno, a 4% reserve rate, that’s about $6.7 million a year, earned on money that today sits in settlement accounts earning nothing. This is the type of volume those 140 partners are doing, and the treasurer hasn’t repositioned anything into an investment. The same operational cash does the same job, except someone now pays rent on it.

Now in that example every supplier redeems on receipt, but I’m not sure the data nor the design suggests they will. The existing stablecoin market moved roughly $33 trillion in 2025 on about $300 billion of supply, an average dwell of around three days, and Tether’s $145 billion circulates for years without redeeming. If you do the math on three days then every number multiplies by eighteen. So now a business that does $1 million a day earns $120,000. That’s essentially the entire upper middle market, so we’re not talking JPMorgan's global treasury services desk here, these are the customers of super regionals and regional banks.

I think the word partner matters, because the yield follows the partner agreements, not the token. Reserve income accrues to the issuer, and a holder has no claim on it without a contract, that’s true of every fiat backed stablecoin, and GENIUS makes paying a mere holder unlawful in any case. Open Standard’s announcement says partners receive all of the earnings from the reserves, less the management fee. Put those together and the supplier’s position follows. A partner collects a share, a holder on a partner platform may get some kind of platform rebate, and a holder outside the network gets nothing while the income their float generates goes into the pool the partners divide. How that pool is split is the piece Open Standard hasn’t published. But held by a non partner, OUSD works exactly like the model it set out to replace, someone else keeps the interest on your money. Tether keeps everyone’s float income. Open Standard, on its own description, keeps the outsiders’ and pays it to the insiders.

That’s the design insight inside Open USD, and I reckon it’s why the partner list looks the way it does. Circle’s business model has the interest on roughly $73 billion of USDC reserves. Tether keeps the interest on $145 billion. Open Standard inverts it: the float income goes to whoever brings the balances, which answers the question that has hung over stablecoin distribution from the start. Why would Stripe or Shopify spend the next decade making someone else’s issuer rich? The OCC expects payment stablecoin issuance could reach $500 billion in 2026, and BNY, an Open Standard partner, has said publicly it anticipates stablecoins alone may grow to $1.5 trillion by 2030. Whoever holds the float on that pool collects the income on it. The consortium is a machine for redirecting where it lands.

The float was always the fee

To see why this is a banking story rather than just a crypto story, ask a question that might sound a bit naive, why doesn’t your operating account pay interest?

It’s not the law. Although it was, once. Regulation Q, born in the Banking Act of 1933, flatly prohibited interest on demand deposits, and for decades the corporate operating account paid zero because it had to. Banks and their clients then did what regulated parties always do, they engineered around the rule. The earnings credit rate was the workaround, a soft credit, calibrated to your balances, that offsets your treasury services fees. Functionally interest. Legally never interest. Not even taxable. Corporate cash management has run on this lawful fiction for probably half a century.

Then in 2011, Dodd-Frank Section 627 repealed Reg Q for business checking, and almost nothing changed. Operating balances kept earning zero, or a thin ECR pegged well below market rates, because the constraint was never really the statute. It was the bundle. The operating account is the pricing mechanism for a relationship. So payment execution, intraday liquidity, fraud and disputes, and above all credit. Your revolver is priced against your operational balances, and every treasurer knows what happens to the facility when the balances leave. The float is how the bundle gets paid for. That spread, between the roughly zero you receive and the bill rate the bank earns on your transactional cash, is the deposit franchise.

By the way, it’s also enormous. When Treasury analyzed the stablecoin yield question, it estimated that as much as $6.6 trillion in deposits could migrate out of banks if stablecoins pay interest through the side door, close to 30% of the US deposit base. Look at which deposits are at issue. Savings balances already compete with money market funds, they were always mobile. The money in play here is operational balances, the deposits Basel’s liquidity rules treat as the stickiest and cheapest funding in the system, the ones banks are allowed to assume won’t run. A payment instrument that rebates float makes them portable. That’s is what to play for here.

Regular readers will recognize the territory. In The End of Idle Money I argued that the settlement friction subsidy erodes once money moves at programmable speed, whether or not the yield loopholes get closed, because velocity alone lets capital rotate out of idle balances. OUSD feels a lot blunter than that. It doesn’t wait for velocity to erode the subsidy, it prices the subsidy and hands it back to the customer as a rebate. Erosion has become expropriation, with a partner agreement attached.

OUSD, seen through this lens, is an unbundling. Payments utility at cost, float income returned to the customer, and the rest of the bundle, the credit relationship, the intraday liquidity, the insurance, simply absent. Whether the bundle was worth its price is now, for the first time in fifty years, a question with a competing quote attached.

Congress saw this coming, almost

The GENIUS Act, enacted in July 2025, contains one sentence built for exactly this moment. Section 4(a)(11): no permitted issuer shall pay the holder of a payment stablecoin “any form of interest or yield (whether in cash, tokens, or other consideration) solely in connection with the holding, use, or retention” of the stablecoin.

Read it as a drafting exercise and you can see the fingerprints of the last fifty years on it. “Whether in cash, tokens, or other consideration” is an anti ECR clause, written by people who know precisely how banks laundered interest into service credits under Reg Q. And “holding, use, or retention” closes the door I suspect most treasurers would reach for first, the claim that they aren’t holding an investment, merely using a payment instrument. Use is enumerated. So good, Congress anticipated the operating account dressing and legislated against it by name.

But the prohibition hangs on two words, and both feel like they cut in Open Standard’s favor. The first is “issuer”, the prohibition binds nobody else, which is why Coinbase can pay 3.5% on USDC balances today, call it a loyalty reward, and fund it from its agreement with Circle, which hands Coinbase all of the reserve income on USDC held on its platform and half of the rest, worth $908 million in 2024, more than half of Circle’s revenue. Neither leg of that arrangement violates the text. The second is “solely”, if a payment to a partner is consideration for a bundle, integration commitments, distribution, acceptance building, governance participation, then it is not solely for holding or use, and the prohibition by its terms doesn’t reach it. Open Standard’s partner agreement could easily be an engineering exercise around one adverb.

The regulators know. The OCC’s proposed rule, issued in February 2026 with 211 open questions attached, adds a rebuttable presumption against conduit arrangements, but read the presumption carefully and it’s definitely drafted for the Circle to Coinbase chain, it fires only when the issuer pays an intermediary solely for holding and the intermediary pays holders solely for holding. A well drafted partner agreement breaks the first leg at the source. What remains then is case by case anti evasion review, where substance beats form, and where a partner whose compensation tracks balances rather than services will struggle. The OCC concedes it hasn’t defined “use” and has asked for comment on whether transaction based rewards fall inside the ban. That undefined word is really where this fight now lives.

Congress, meanwhile, is bidding to define it first. The CLARITY Act, the market structure bill that cleared Senate Banking 15-9 in May, nearly stalled in markup on exactly this question, and the Tillis-Alsobrooks compromise that got it through draws the line in statute, payments that function like deposit interest are blocked, activity based rewards are permitted. The banking trades objected to that compromise for the same reason a treasurer should study it. At corporate scale, activity based and balance based are the same money measured differently, my four hour dwell back of an envelope arithmetic above is an activity reward that pays the T-bill rate. The bill sits on the Senate calendar waiting for floor time before the August recess. If it passes as drafted, the OUSD partner model gets something close to a statutory safe harbor for use based economics. If the banking lobby wins the floor fight, the ban tightens toward substance over form.

Timing is not just the secret to comedy, it does a lot of work here too. GENIUS doesn’t take effect until January 2027 or 120 days after final rules, whichever comes first. Open USD launches into wet concrete, and 140 companies will have live economics in the structure before the boundary hardens. More than 3,200 bankers have signed the ABA’s letter demanding Congress close the pass through. The awkwardness is that several of the banking industry’s largest names now sit inside the consortium whose economics depend on it staying open. The lobbying coalition against the loophole just lost some of its heaviest members to the other side of the table.

We have run this experiment before

If this feels familiar, it should, I used the same history in The End of Idle Money to make a point about what happens when capital learns it has options. It’s worth one more pass here, because this time the endgame mechanics are the point. In 1977, money market mutual funds held under $4 billion. Reg Q ceilings capped what banks could pay savers while inflation ran past 10%, and Merrill Lynch had just launched the Cash Management Account, a brokerage product that looked suspiciously like a checking account with a market yield. By the end of 1982, money funds held roughly $235 billion, and the deposit base was bleeding out of the banking system toward instruments that arbitraged a yield prohibition from outside the regulatory perimeter.

Money funds were never banned. The endgame was Garn-St Germain in 1982, where Congress handed banks the money market deposit account, a weapon to compete with, and the old prohibition collapsed because it had become a subsidy to the banks’ competitors rather than a protection for the system. Same instinct, different instrument, different decade.

I can’t help but think the stablecoin yield ban is Reg Q’s grandchild and will meet the same fate by one of the same two roads. Either the loophole gets closed hard, which requires legislation that the consortium’s membership is now well positioned to soften, or banks get their own weapon. That weapon has, by now, an obvious name. A tokenized deposit pays a real rate on operational balances while keeping the credit relationship, the insurance, and the regulatory perimeter intact. It is the MMDA of this cycle, and every bank inside the Open Standard tent that is simultaneously building tokenized deposit rails is hedging exactly this bet.

The honest caveats, and why they don’t save the incumbents

Ok, but let’s be real here, consortiums have a graveyard, and it’s chock full of great press releases. Libra’s 2019 logo wall looked unbeatable and produced nothing. And Open USD is at least the fourth yield share coin, so the model has benchmarks. Paxos’ Global Dollar Network launched in late 2024 with 130 plus partners and essentially the same pitch, took 13 months to reach $1 billion, and sits near $3 billion today against Tether’s $145 billion. Agora’s AUSD is at roughly $181 million. USDF, the bank consortium version, is dead. A revenue share alone demonstrably doesn’t move balances. Nearly every marquee OUSD partner runs a competing strategy, Stripe owns Bridge, Coinbase’s economics remain tied to USDC with the commercial agreement reportedly up for renewal in August, and half the banks on the list are building the tokenized deposits I just described. Membership is additive, not exclusive. Dragonfly’s Rob Hadick put it plainly:

consortiums are hard and they break easily. OUSD may well end up as the best marketed letter of intent since Libra.

But price discovery doesn’t need the challenger to win, it only needs the quote to exist. The first time a corporate treasurer earns $6.7 million on float that previously earned nothing, the number goes into every treasury RFP in that sector, and the banks across the table have to answer a question they have never had to answer, what, exactly, am I being paid for the use of my operational cash? The ECR conversation stops being a fee negotiation and becomes a rate negotiation. That repricing happens whether OUSD reaches $500 billion or stalls at Global Dollar’s $3 billion, because the offer is now public and the arithmetic is not complicated.

And the repricing does not land evenly. Money center banks can absorb thinner deposit margins and build the tokenized rails to compete. Specialist banks can price the whole relationship deliberately. The banks in the middle, funded on corporate transaction balances they assumed were inert, holding loan books priced against funding costs that are about to become a negotiation, are the ones for whom the operating account franchise was doing the most work. Fewer of them exist every year already. You have to believe this accelerates the sorting.

The operating account was never free. The price was deducted from the yield before the customer ever saw it, which is why nobody negotiated it for fifty years. OUSD’s real contribution might end up being less about the coin, and more about it’s basic arithmetic. Every treasurer can now put a number on what their float is worth to someone else, and no treasurer who has seen that number will easily unsee it.

Banks may not lose the operating account over this, but they’ll start paying for it.

References

Previously on Atomic Settlement

The End of Idle Money, March 2026: the settlement friction subsidy, the 1980s rhyme, and why closing yield loopholes doesn’t solve the funding problem
The Orchestration Layer, June 2026: where value capture goes once rails and instruments commoditize

Open Standard / Open USD

Legislation and regulatory sources

Deposit economics and industry positions

Atomic Settlement

Discussion about this post

Ready for more?