Seven cents. That is the number that surfaced this week, attached to Circle and its USDC franchise: for every dollar of revenue the issuer books, roughly seven cents survive as profit. The framing arrived as a provocation โ one dollar in, seven cents left โ and it did what provocation does. It travelled. It also, quietly, conflated two very different things.
I have spent enough time inside node logs and reserve attestations to distrust a ratio before I know its denominator. So before the argument, the accounting. Seven cents is not a gross margin. It is almost certainly a net margin. Circle booked roughly $1.68 billion in revenue in 2024 against net income in the $150 million range. That arithmetic lands at 7 to 9 percent. The number is real. The rhetorical wrapper around it โ only seven cents left, the grammar of a business that cannot help itself โ is not.
That distinction is the whole article. A 7 percent gross retention would describe a structurally broken commodity. A 7 percent net margin describes a fee structure. Those are not the same disease, and they do not take the same medicine.
Circle is not a protocol. It is a licensed money transmitter with a balance sheet. Every USDC in circulation is a liability: one dollar owed, one dollar held. The reserve sits in short-dated US Treasuries and repo, managed through BlackRock, custodied at banks, attested quarterly. Circle earns the interest on that reserve. It pays almost nothing out.
That is the engine. More than 95 percent of revenue comes from reserve interest โ not from payment APIs, not from CCTP fees, not from Programmable Wallets. Circle is a float business wearing a payments company's jacket. It is a net interest margin operation with a mint function.
Which means the revenue line is a rate. Average USDC circulation multiplied by the short-end yield, minus distribution, minus operating expense. Three variables. Circle controls one of them.
The distribution line is where the seven cents goes to die. Under the arrangement with Coinbase, USDC held on Coinbase generates reserve interest that flows entirely to Coinbase. USDC held off Coinbase splits roughly fifty-fifty. Add Binance, add market makers, add the exchange tier generally, and 50 to 60 percent of gross revenue exits through the channel before Circle pays a single engineer.
So the waterfall looks like this. One hundred cents of reserve interest. Fifty to sixty cents to distribution. Thirty to forty cents to operations, compliance, legal, audit, headcount. Seven to ten cents left. The headline is not a scandal. It is a subtraction.
Here is what the subtraction actually tells you.
Circle has no technical moat, and the seven cents is the price of that absence. Minting and redeeming a fiat-backed token is standardized engineering. Any licensed financial institution can ship it in six to twelve months. There is no cryptography here a competitor cannot replicate, no consensus mechanism to defend, no network effect that locks a treasury desk into USDC when a single swap moves it to USDT or DAI. The barriers are licenses, exchange contracts, and the operational competence to manage tens of billions in reserve. Those are commercial barriers, not technical ones. A protocol with a real moat โ Ethereum, for instance โ retains the overwhelming majority of the fees it generates. A commoditized issuer must share. The seven cents is not mismanagement. It is the market clearing price of being replaceable.
And note the inversion the industry rarely says out loud: Tether's retention is dramatically higher than Circle's, and the gap is not operational excellence. It is regulatory posture. Tether runs a lighter compliance structure, a narrower audit surface, and a distribution model that leans on emerging-market demand rather than negotiated revenue shares with US exchanges. Circle chose the licensed path. That choice bought it institutional trust, MiCA standing, and a multi-trillion-dollar addressable market โ and it bought those things with margin. Compliance, in stablecoins, is a cost line. It has never been a revenue line.
Second: this is not a Ponzi, and the distinction matters more than the headline. The cash funding Circle's distribution payments is US Treasury interest โ an external, real, third-party transfer. It does not require new entrants to pay old ones. A Ponzi collapses when inflows stop. Circle's model compresses when rates fall. Those are different failure modes with different timelines and different early-warning signals. Anyone who read seven cents and reached for the P-word read the wrong chart.
Third: the seven cents is a variable, not a constant, and that is the part the headline buried. It is a function of the rate cycle and of scale. When USDC circulation grows faster than headcount, operating leverage pushes the margin up โ the denominator effect. When the Fed cuts, the numerator breaks immediately, while the fixed components of distribution and compliance do not move down in step. Model it: short rates from 4 percent to 1 percent cuts gross revenue by roughly three quarters, holding circulation flat. The margin does not compress. It can invert. Depending on the quarter and the rate path, that seven cents could plausibly read anywhere from three to twenty.
Fourth: the competitive threat is misidentified. The market watches Tether. Tether is not the threat. Tokenized money market funds are. A yield-bearing on-chain dollar instrument, if permitted at scale, makes a zero-yield payment token look like a checking account next to a savings account. What blocks that threat is legislative, not commercial: the GENIUS Act framework bars issuers from paying interest to stablecoin holders. That prohibition is the invisible pillar holding up Circle's entire P&L. Remove it and Circle must share reserve interest with holders. At that point seven cents does not shrink. It goes negative.
I have watched this pattern before, in a different market. In 2021 I clustered wallets on a profile-picture project and found 60 percent of the community resolving to three addresses running wash trades. The data was clean. The conclusion was unwelcome. I wrote the report anyway and filed it. Data integrity is not a popularity strategy. The same discipline applies here: the seven cents is not an accusation, it is a structure, and structures can be read.
Which brings me to the strategic tell nobody is pricing. Circle's product roadmap โ Arc, its own settlement layer; CCTP for native cross-chain transfer; Programmable Wallets โ is not technical vanity. It is de-channelization. Every dollar of USDC that reaches a user through Circle's own rails instead of Coinbase's is a dollar of distribution cost recovered. If that vertical integration works, seven cents becomes thirty. If it fails, Circle remains permanently a toll collector for its own distributors.
The consensus reading of the seven-cent figure is that Circle is a thin-margin commodity business and therefore deserves a commodity multiple. I think that is backwards in one specific way, and it is the more dangerous error.
Circle is not a low-margin business. It is a high-beta one. The correct mental model is not a payments network. It is a levered position on the short end of the US curve, wrapped in a regulatory license, with a fixed cost base. Frame it that way and the valuation question changes entirely: what multiple should the market pay for a rate-beta instrument?
If CRCL trades on total addressable market โ the two-trillion-dollar stablecoin story โ then the seven cents is irrelevant to the price, and the stock is a narrative asset. If it trades on unit economics, the seven cents is everything, and the stock is a financial. The gap between those two frames is where the damage lives. And here is the asymmetry: the market can hold the TAM frame for years. It only has to break once, at the moment the rate cycle turns, for the repricing to be violent.
There is a second blind spot. Circle's dependency runs uphill. Against Coinbase it is the weaker party โ on-platform interest flows 100 percent to the exchange, which captures a large slice of USDC's economic value while carrying none of the reserve management or compliance cost. Against DeFi it cannot charge an on-chain fee; Aave and Uniswap monetize USDC liquidity and Circle sees none of it. Against end users, switching costs are effectively zero. Against regulators, it depends on licenses it cannot replace. Circle is weak at every link. The seven cents is not a cost problem. It is a bargaining-position problem.
And the lobbying risk is underweighted. Banks are the natural adversary here. Stablecoin issuance pulls deposits out of the banking system and redirects Treasury interest away from bank balance sheets. The lobbying incentive to restrict non-bank issuers is structural and well funded. It does not appear anywhere in the seven-cent discussion. It probably should.
So: seven cents is a net margin, not a gross one. It is cyclical, not structural. It is the price of replaceability, not evidence of incompetence. And it is protected by a legislative prohibition on yield that most of the market has not bothered to price.
The signal to watch next quarter is not the margin itself. It is the split between on-platform and off-platform USDC, and whether Arc and CCTP move the second number. That is the only line item that tells you whether Circle is becoming infrastructure or staying a distributor's subcontractor.
Silence is the most expensive asset in a bubble. The seven cents has been saying something for a while. The question is whether anyone is reading it as data, or as a headline.