In the third quarter of 2024, Circle Internet Group reported revenue north of $700 million on an annualized basis, and retained roughly seven cents of net profit for every dollar that entered its books. The headline — "one dollar in, seven cents kept" — circulated through crypto media as if it were a scandal. It is not. It is an accounting disclosure dressed as a revelation, and the two are not the same thing. The ledger remembers what the hype forgets, and what the ledger shows here is a company whose cost structure, not its business model, is consuming the margin.
The distinction matters because the two interpretations lead to opposite conclusions. If seven cents is a net margin, Circle has a distribution problem it can solve through vertical integration and scale. If seven cents were a gross retention rate, Circle would be selling dollars at a loss and the entire enterprise would be structurally insolvent. Only one of those readings survives contact with the numbers.
I have spent the better part of a decade auditing DeFi protocols and stablecoin mechanics, starting with an obscure 2017 ICO whose token minting function contained an integer overflow I found with a custom Python script. The team never answered my email. The token launched anyway. That experience taught me the first rule of this work: never accept a metric until you have located the denominator. Seven cents out of a dollar is meaningless until you know whether the dollar is gross revenue, net revenue, or net interest income after distributor rebates.
Circle's revenue base is overwhelmingly reserve interest, not payment fees. Roughly 95 percent of the company's top line derives from the yield earned on short-duration U.S. Treasury bills and overnight repo agreements backing the USDC float. The remaining sliver — API fees, CCTP cross-chain transfer charges, Programmable Wallets subscriptions — is growing but remains immaterial to the consolidated picture. Circle is not a payments company that happens to hold reserves. It is a money market fund with a blockchain front-end, and its income statement should be read accordingly.
This is where the seven cents originates. Every dollar of reserve interest must first satisfy the distribution agreements that place USDC into circulation. Under the arrangement disclosed in Circle's S-1 and subsequent 10-K filings, USDC held on Coinbase earns 100 percent of its reserve interest for Coinbase, while USDC circulating off-platform splits roughly fifty-fifty between the two firms. Coinbase also receives a share of on-chain transaction fees generated by USDC on its own networks. When you compress these terms into a single blended rate, you arrive at a distributor take of somewhere between 50 and 60 percent of gross revenue.
Layer operating expenses on top of that — compliance staff, licensing fees across dozens of U.S. state money transmitter regimes plus the French EMI authorization, auditing, engineering, marketing — and you consume another 30 to 40 cents. The residual is the seven to ten cents that Circle keeps. The math is not mysterious. It is contractual.
What the floating headline obscures is that this number is not a constant. It is a function of three variables, two of which Circle cannot control. Net profit roughly equals average USDC circulation multiplied by the short-term risk-free rate, minus distributor share, minus operating expense. Circle controls circulation growth and operating expense. It does not control the Federal Reserve's policy rate, and it does not unilaterally control the distributor share.
This makes Circle a high-beta interest rate instrument wearing the clothing of a technology company. When the Fed cut rates by 100 basis points, Circle's revenue declines by roughly 25 percent on a static float. Distributor terms contain fixed components that do not compress proportionally. In a scenario where short rates fall from 4 percent to 1 percent — a path the market has priced as plausible — Circle's net margin could compress toward zero, and the seven cents becomes three, or negative. That risk appears nowhere in the popular framing of the story, because the popular framing treats the metric as structural rather than cyclical.
There is a second, quieter risk that almost no one discusses: the political durability of the float itself. Circle's income is, in literal terms, a transfer of U.S. Treasury interest expense to a private issuer. The company is a meaningful buyer of short-dated government paper, which makes it useful to the Treasury while rates are high and issuance is smooth. But if a legislative coalition ever forms around the proposition that stablecoin issuers should not capture seigniorage rent privately — and the banking lobby has strong incentive to build that coalition, since stablecoin float directly cannibalizes deposit funding — the entire revenue base becomes a policy variable rather than a market one.
Understanding why Circle cannot retain more requires looking at where it sits in the value chain. It occupies a position that is simultaneously essential and weak. The system needs USDC for its liquidity depth and its regulatory standing; institutions have built settlement rails around it. But the system is not loyal to USDC. A holder can swap into USDT or DAI in a single on-chain transaction at negligible cost. Aave and Uniswap collect fees on USDC pools while Circle captures nothing from the downstream activity. Trust is a variable, not a constant, and here it is a variable set by switching costs that round to zero.
Circle's weakest negotiation happens upstream, with its own distribution channel. Coinbase holds the leverage because it controls the retail and institutional funnel through which most USDC enters circulation. The asymmetry shows up in the terms: full interest capture on platform balances, zero on the reserve management burden. Coinbase earns a substantial share of USDC's economic value while bearing none of the custody, audit, or licensing cost. This is not malfeasance. It is the natural outcome of a commodity issuer bargaining against a differentiated distributor.
The competitive landscape sharpens the point. Tether operates with a fraction of Circle's compliance overhead, which is precisely why Tether's retained margin per dollar of float dwarfs Circle's. In the stablecoin market, compliance is a cost line, not a revenue line. That is an uncomfortable truth for anyone who believes regulatory clarity confers pricing power. It confers legitimacy, market access, and institutional trust — none of which translate into the ability to charge more, because the underlying product is a dollar and a dollar is worth a dollar everywhere.
The real competitive threat is not Tether. It is tokenized money market funds. A yield-bearing on-chain dollar instrument would render a zero-yield stablecoin strictly inferior for any holder with a time horizon longer than a week. That threat is currently blocked by the GENIUS Act's prohibition on paying interest to stablecoin holders — a provision that functions as an invisible subsidy to Circle's business model. Most market participants read that clause as consumer protection. It is more accurately read as a structural guarantee that Circle will never have to share its float income with the people whose money generates it.
Strip away the legislative protection, and the seven cents becomes structurally negative. If Circle were required to pass even half of its reserve yield to holders, the gross margin would collapse below the distributor take, and the model would require a complete redesign. Any legislative movement toward permitting yield-bearing stablecoins is an existential event for Circle, not a competitive annoyance.
What does a rational operator do when every link in the chain holds pricing power over it? It attempts to remove the links. Circle's product roadmap reads less like technological ambition and more like a map of escape routes. CCTP removes dependence on third-party bridges and captures cross-chain transfer fees directly. Programmable Wallets place USDC issuance and custody inside Circle-controlled interfaces, reducing the reliance on exchange distribution. And Arc, the company's proposed layer-one settlement network, is the most consequential bet of all: if Circle controls the settlement layer where USDC natively lives, it becomes the channel rather than the commodity flowing through it.
In my audit work on cross-chain infrastructure, I have repeatedly found that the economic architecture of a protocol usually predicts its vulnerability surface better than its code does. A bridge that must trust three external validators has a different risk profile than one that trusts itself. Circle's strategic logic follows the same principle: the company is attempting to convert a trust-heavy, margin-thin position into a trust-light, margin-rich one. Whether Arc delivers that outcome depends on adoption dynamics that remain unproven. The attempt itself, however, tells you what management understands about its own position.
The contrarian reading of the seven cents is that it is not a scandal at all, but rather an unusually honest disclosure in an industry that specializes in obscuring unit economics. Most token projects would never publish a metric this unflattering. Circle does so because regulators require it, and because the alternative — a prospectus that flatters the margin and invites litigation — is worse. Clarity precedes capital; chaos precedes collapse. The presence of an unflattering number in a public filing is evidence of a functioning disclosure regime, not evidence of a failing business.
What the number does reveal, however, is a valuation risk that the market has largely ignored. Circle is being priced, in large part, on the total addressable market for tokenized dollars — a multi-trillion-dollar narrative about the future of settlement. But the actual economics of the business at current scale resemble a narrow-spread financial intermediary with concentrated distributor relationships and acute rate sensitivity. There is a meaningful gap between the multiple a payments network deserves and the multiple a float-margin business deserves. When that gap closes, it will close abruptly.
There is an early signal worth tracking. Analysis that shifts from total addressable market to unit economics tends to appear near the end of a valuation expansion, not the beginning. When market participants start asking how many cents per dollar a company keeps, rather than how large the market it addresses might become, the narrative has already begun to fatigue. This does not mean the price falls tomorrow. It means the marginal buyer has changed from someone pricing a story to someone pricing a cash flow, and those two buyers do not pay the same multiple.
Data does not lie; people do. The seven cents is real, and it is disclosed, and it is defensible as long as rates stay elevated and the distributor terms hold. But nothing in Circle's structure guarantees either condition. The company's future turns on a single question it cannot fully answer alone: whether it can build a settlement layer that makes its own distribution partners optional before the interest rate cycle makes its current margin untenable. Every line of code is a legal precedent, and every architectural decision here is a financial one.
The bug was there before the launch. In Circle's case, the bug is not in the contract. It is in the assumption that a commodity dollar can command a durable margin without owning the rails it travels on.
