Circle already pays most of its reserve earnings away. The argument is about who gets to negotiate for the rest.
Start with a number from a filing, not from a press release. In the three months to June, Circle reported $701 million of total revenue and reserve income against $412 million of distribution, transaction, and other costs in its second quarter results. Fifty-nine cents of every dollar that came in went back out to the companies that put USDC in front of users, before Circle counted a profit.
That ratio is why Open USD, a token backed by more than 140 companies, believes it can take a serious share of the dollar stablecoin market away from USDC. It is also why Circle’s chief executive believes it cannot.
Where stablecoin reserve income comes from
The mechanism takes one paragraph to explain, which is part of why the fight over it is so direct.
A customer hands an issuer a dollar and receives a token. The issuer puts that dollar into short-dated Treasury bills, overnight repo, money market funds, or bank deposits, and keeps what those assets pay. The token holder receives no interest from the issuer, which American law now prohibits. The float is the product.
Scale turns an ordinary yield into a large business. Circle averaged $76.5 billion of USDC in circulation over the second quarter and earned $701 million on it, which works out to roughly 3.7 percent annualized. At that rate, every billion dollars of average float is worth about $37 million a year.
Two things follow from that arithmetic. Reserve income tracks short-term interest rates rather than usage, so a quarter can go backwards on falling yields while transaction volume climbs. And whoever decides where the float sits can dictate terms to whoever issues the token. The second point is where this argument actually lives.
What Open USD proposes
Open Standard announced Open USD at the end of June with a partner list that reads like a directory of the payments industry.
Visa, Mastercard, American Express and Discover are all in it. So are Stripe, BlackRock, BNY, Standard Chartered, Google, Samsung, Shopify, IBM, U.S. Bank, BBVA, Coinbase, Ripple and Galaxy, among more than 140 companies in total. Circle, Tether and PayPal, the three largest dollar stablecoin issuers, are not.
The design has three parts, set out in the announcement introducing Open USD. Partners mint and redeem at no cost and with no volume caps. Partners receive all of the earnings from the reserves, less a small management fee covering operating costs. And it is run by an independent company whose board is drawn from the partners, so reserve policy and redemption rules are set collectively instead of by one issuer.
Zach Abrams leads Open Standard as founding chief executive. Stripe’s Will Gaybrick said the token is meant to become “the default stablecoin for businesses running on Stripe,” which is the clearest statement anyone has made about how distribution is supposed to work. The coin is expected to go live later this year, and as of mid-September it has not.
The part that gets misread
Read quickly, the pitch sounds like an issuer giving away its revenue. Read against Circle’s accounts, it looks more like a formalization of what the market already forced.
Circle’s distribution costs were $407 million in the first quarter and $412 million in the second. Those are not marketing budgets. They are payments to exchanges, wallets and platforms for holding and pushing USDC, and they have run at nearly three-fifths of revenue in both quarters this year. The issuer that supposedly keeps the reserve income has been handing most of it over for some time.
What differs is who does the negotiating. Circle’s payouts are struck deal by deal, which means they reward bargaining power. A platform with millions of users extracts a share. A payments company with a few thousand business customers gets an integration and no economics at all. Open USD’s proposition is that a small distributor receives the same terms as a large one without having to win an argument it cannot win.
Moody’s analyst Stephen Tu framed the appeal in structural terms, saying consortium models “may have advantages over single-issuer approaches by aligning incentives, broadening distribution and supporting interoperability.” That is a rating agency’s way of saying the incentive to switch is real.
Read more: Tether posts a record quarter while its freeze powers go to court • Why US banks are splitting between tokenized deposits and stablecoins
What the Coinbase agreement actually shows
One relationship demonstrates the problem better than any argument about it.
Coinbase receives roughly half of USDC’s reserve income under its agreement with Circle. At the end of the second quarter, Coinbase held 30 percent of USDC in circulation. Circle keeps an issuer allocation first, then the remainder is split according to where balances sit, and Coinbase also collects a share of income from USDC held on neither company’s platform.
Thirty percent of the float, about half of the money. That gap measures what a single dominant distributor can extract when no other partner holds the same position.
Distribution deals come in a few recognizable shapes. The simplest pays a share of income on balances held on the partner’s platform, which rewards custody. A second pays on minting and redemption volume, which rewards moving the token instead of holding it. A third pays a flat integration fee and no reserve economics at all, which is what most partners get. Circle’s arrangement combines the first with a share of income from balances held nowhere near either company, and that last element is why the payout runs ahead of the custody share.
CoinShares analyst Luke Nolan called Open USD the “biggest threat yet” to USDC on precisely this basis, arguing that the consortium puts pressure on Circle’s reserve margin and on the cost of buying distribution at the same time. Jefferies made the narrower point that Circle draws about 95 percent of its revenue from reserve interest and leans on a single partner to place most of it.
The agreement renewed in August on existing terms for three more years, which Circle confirmed on its second-quarter call. Nothing was renegotiated, despite Coinbase having joined the consortium weeks earlier. Read that either as stability or as a reminder that the terms were already generous enough not to need improving.
Circle’s answer
Jeremy Allaire answered on July 1, the day after Circle’s stock fell more than 17 percent.
His argument was that stablecoins are platform businesses built over years out of developer integrations, liquidity depth and regulatory licenses, and that a partner list is not any of those things. He cited about $30 trillion of onchain USDC transactions in the first quarter, which he put at roughly 80 percent of dollar stablecoin transaction volume.
On the economics, he was blunter. Giving away all income, he said, “is a recipe for starving your infrastructure, systematically underinvesting and ensuring that your platform will remain limited in scope.” On governance, he said the record of multi-company products reaching scale “is absolutely dismal.”
Both points land. An issuer that keeps nothing funds nothing, and a board of 140 competitors agreeing on reserve policy is a coordination problem with a long history of going badly. His defense of USDC’s network effects is the strongest part of the case, because liquidity is the one thing a launch announcement cannot manufacture.
CoinShares conceded the same ground from the other direction, noting that Open USD is still an idea with unknown details, not a product. A model that pays better is worth nothing until there is something to hold.
The model has already been tried
Neither side has to speculate about whether paying distributors works, because a smaller version of this idea has been running since 2024.
Paxos launched Global Dollar in November of that year with the Global Dollar Network behind it, offering partners up to 100 percent of the returns generated by USDG held on their platforms. Robinhood, Kraken, Galaxy, Anchorage, Bullish and Nuvei signed on as founding members. The pitch was the one Open Standard is making now, with a shorter list of names attached to it.
USDG circulation sits at about $3.2 billion by DefiLlama’s count. USDC finished the second quarter at $73.3 billion. Close to two years of the most generous revenue share in the market bought roughly four and a half percent of the incumbent’s float.
That is not proof the model fails. Paxos did not have Visa, Stripe, Google, and four card networks, and the regulatory ground has moved under both of them since. What it does show is that better economics for distributors do not move balances on their own, which is the substance of Allaire’s objection and the part most of the launch coverage skipped.
The far end of the spectrum deserves a mention for contrast. Tether runs no comparable revenue share and still posts the largest profits in the sector, as our reporting on Tether’s quarterly profit set out. Three issuers, three answers to the same question about who keeps the float, and the one that keeps the most has the biggest token.
What this changes for a card program
For anyone running a card, wallet or payments product, the practical question is narrower than the market-share fight, and it comes down to the float.
Every program carries a balance. Customer funds waiting to be spent, prefunded settlement positions, treasury held for authorization coverage. Under the current arrangement, that balance earns for the issuer and for whichever distributor negotiated hard enough, and a mid-sized program is usually neither. A program averaging $50 million of float is sitting on roughly $1.85 million a year at present yields that it does not see.
So the choice of default token stops being a technical decision about chains and liquidity and becomes a revenue line. That is a different conversation, with a different person in the room, and it is why the consortium bothered signing four card networks instead of two.
Notice that the card networks are hedging in public. Visa sits in the Open USD partner list and separately runs its own route for banks to mint and settle, which we covered in our story on Visa’s stablecoin platform. A network holding both positions is not forecasting a winner. It is making sure it earns a fee whichever token the volume arrives on, and that is a reasonable posture for a program to copy.
Three cautions belong next to that. Reserve income is rate-dependent, so a model built on sharing yield gets less persuasive every time the Federal Reserve cuts. A token with no live liquidity costs more to use in spreads and conversion than it pays in shared reserve income, which means the arithmetic only works after adoption, not before. And a program that supports a second token doubles its reconciliation, treasury, and compliance surface, which is a real cost that no announcement has priced.
The reasonable position for most programs is to keep settling in what is liquid, negotiate reserve economics into the next contract renewal whichever token wins, and treat launch timing as the trigger for a decision, not the decision itself.
What to watch between now and launch
Four things will separate a business from a press release.
Whether Open USD actually ships this year, since the launch window narrows with every week of September that passes. Whether the management fee gets published as a number, because “small” is not a term anyone can model. Whether a large partner commits volume instead of a logo, with Stripe’s default-token statement the obvious one to test. And whether Circle’s distribution costs move in either direction over the next two quarters, which is the cleanest available measure of whether its partners have gained bargaining power.
The deeper question sits underneath all four. Reserve income exists because holders are not allowed to receive it, so the only argument is about which intermediary keeps it. A consortium sharing the money more evenly is a better deal for distributors and no different at all for the person holding the token. That is worth remembering when the marketing arrives, and it is the part of the design nobody in the 140 has any reason to bring up.