I’ve been watching the stablecoin space long enough to know that raw circulation numbers can paint a misleading picture. USDC has kept expanding, average balances climbed roughly a quarter year over year in the latest reported quarter, yet the real question most people quietly ask is simpler: after the partners take their cut, how much of that interest income actually stays with Circle? The answer sits in the filings, and it is more nuanced than a single headline percentage.
How Circle Turns Reserve Assets Into Reported Income
Circle earns the bulk of its money the same way many reserve-backed issuers do. Dollars and short-term instruments sit behind every USDC in circulation. Those assets generate interest and dividends. In the second quarter the company recorded reserve income of $667.733 million. Add the smaller stream of other revenue and the combined total reached $701.315 million.
That figure looks impressive until you remember it arrives before any distribution or transaction costs are subtracted. Circle itself is careful to note that reserve income is booked gross. The amount left after partners are paid is what matters for the business that actually operates the company.
Average Circulation Versus Quarter-End Snapshots
Average USDC circulation across the quarter sat at $76.524 billion. The final day of the period showed a lower $73.269 billion. The distinction is not academic. Interest accrues on the daily balance, so the average figure drives the income calculation far more than any single closing number.
Year earlier the average had been $61.039 billion. Growth of that scale should have produced a much larger jump in reserve income. Instead the rise was only 5.3 percent. The reason is straightforward: the published reserve return rate slipped from roughly 4.1 percent to about 3.5 percent. Lower yields on the underlying assets offset a good portion of the circulation gain.
I find that dynamic particularly interesting because it shows how sensitive the model remains to the broader rate environment. A bigger float does not automatically translate into proportionally bigger earnings when overnight funding rates move lower.
What Circle Actually Kept Before Operating Costs
After subtracting distribution, transaction and other costs of $412.470 million, Circle retained $288.845 million. That works out to a retained-revenue margin of approximately 41.2 percent. In other words, a little over forty cents of every dollar of combined revenue and reserve income stayed inside the company before staff, technology, legal and other operating expenses were paid.
Compare that with the same quarter a year earlier. Total revenue and reserve income then stood at $658.078 million and the relevant costs were $406.942 million, leaving $251.136 million or about 38.2 percent. The retained amount grew roughly 15 percent while the overall top line rose only 6.6 percent. The margin itself improved by three percentage points.
That improvement is worth underlining. Distribution costs did rise in absolute dollars, yet they grew far more slowly than the income they were charged against. The narrative that every extra dollar of USDC simply disappears into partner payments does not match the numbers from this particular quarter.
| Measure | Q2 2025 | Q2 2026 |
| Revenue and reserve income | $658.078 million | $701.315 million |
| Distribution, transaction and other costs | $406.942 million | $412.470 million |
| Amount retained before operating expenses | $251.136 million | $288.845 million |
| Retained-revenue margin | 38.2% | 41.2% |
The Coinbase Relationship and Its Weight
A large share of those distribution costs flows to one longstanding partner. Coinbase-related payments totaled $324.6 million in the quarter, down slightly from $332.3 million a year earlier. That single relationship accounted for roughly 78.7 percent of the entire distribution, transaction and other cost category.
The commercial arrangement is more layered than a simple percentage of balances held on one exchange. Circle retains an issuer portion first. Coinbase then receives an allocation linked to USDC sitting on its platform and a further share tied to broader ecosystem growth after third-party participants are paid. Any quick calculation that multiplies visible exchange balances by the reserve yield will therefore understate the actual expense Circle records.
Management has indicated the current terms run through 2029. That longevity brings predictability, yet it also means the cost structure is largely fixed for the next several years unless the parties reopen the conversation.
Where the Tokens Sit Changes the Math
Circle tracks the daily proportion of USDC held on its own platform. That mix includes Circle Mint accounts, company-held tokens and certain non-custodial wallets offered through its services. At quarter end the on-platform balance was $12.442 billion, about 17 percent of total circulation. The weighted average share across the full quarter was higher at 19.5 percent, up sharply from 7.4 percent a year earlier.
The rise matters because balances that never leave Circle’s own channels generally carry lower or zero third-party incentive payments. Growth that occurs through commercially incentivised partners produces the opposite effect. Same increase in circulating supply can therefore deliver very different retained income depending on the channels that drive it.
I’ve always thought this balance-mix dynamic is one of the more under-appreciated levers in the stablecoin business. Circulation headlines grab attention; the location of those tokens quietly decides how much of the yield the issuer keeps.
Interest-Rate Sensitivity Cuts Both Ways
Circle publishes a simple sensitivity table. Starting from an average June reserve yield of 3.49 percent, a one-percentage-point parallel move is estimated to change annual reserve income by $737 million. The same move is modelled to alter distribution and transaction costs by $360 million. The net difference before other items is therefore $377 million.
A two-percentage-point swing produces roughly double those figures. The model assumes circulation stays fixed at the June 30 level and that the rate change applies evenly. It is not a forecast. It simply illustrates that partner payments move with the underlying yield, so a drop in gross reserve income does not translate one-for-one into lower retained income.
The reverse is also true. When rates rise, Circle does not keep every extra dollar. A meaningful slice still flows outward under the existing commercial contracts.
Beyond Reserve Income: Payments and Infrastructure
Other revenue reached $33.582 million in the quarter, up 41.1 percent. Even after that growth it still represented less than five percent of the combined total. The company is clearly trying to diversify, yet the interest earned on reserves remains the dominant engine.
On-chain transaction volume for USDC hit $14.8 trillion, a 151 percent jump. That number counts every transfer; it is not revenue. A single dollar can move dozens of times while the reserve behind it stays put. Circle’s managed payments product and the early activity on its newer blockchain network introduce additional fee streams, but those streams are still small relative to the interest income.
Perhaps the most interesting near-term variable is how the company will recognise proceeds from earlier token sales linked to its infrastructure network. Cash received in the second quarter was largely recorded as deferred revenue. Recognition depends on the satisfaction of specific performance obligations, so the timing of future revenue could differ from the timing of cash collection.
Operating Expenses Still Sit Between Retained Revenue and Profit
The $288.845 million retained after distribution costs is not profit. Operating expenses of $254.486 million brought operating income down to $34.359 million. Net income from continuing operations settled at $48.214 million after other items and tax.
Year-over-year comparisons of operating expenses are complicated by one-time stock compensation tied to the earlier public listing. On an adjusted basis the company has indicated ongoing costs rose. That pattern is consistent with a business that is still investing heavily in technology, compliance and product expansion rather than maximising near-term margins.
Management has repeatedly stated a preference for reinvestment over dividends. That stance is understandable for a company still building out payments rails and custody capabilities, yet it also means shareholders should not expect large cash returns in the immediate future.
What the Next Earnings Report Could Clarify
The third-quarter results are scheduled for early November. Several items will be worth watching closely. First is the updated retained-revenue figure after the latest distribution costs. Second is any shift in the weighted average share of USDC held on Circle’s own platform. Third is the size of Coinbase-related payments relative to total costs. Fourth is the contribution, if any, from newer payments and infrastructure products once deferred items begin to convert into recognised revenue.
I expect the conversation around competition for large balances to continue. Platforms that control meaningful USDC holdings have leverage to negotiate higher shares of the yield. Whether that pressure ultimately compresses the 41 percent retained margin or whether growth in lower-cost channels offsets it remains an open question.
One thing the second-quarter numbers do establish is that higher circulation and lower yields can still produce both higher absolute retained income and a modestly better margin. The model is not as fragile as some simplified critiques suggest. At the same time, the heavy reliance on a single distribution partner and the sensitivity to interest rates leave clear areas of concentration risk.
Looking further out, the mix of balances, the evolution of commercial terms with existing and new partners, and the gradual contribution of non-reserve revenue streams will determine how much of the income Circle ultimately keeps. The filings give a clear baseline. The next few quarters will show whether that baseline can keep improving or whether the cost of distribution begins to catch up again.
For anyone trying to understand the economics of a major stablecoin issuer, the retained-revenue measure after distribution costs is currently the cleanest window available. Gross reserve income tells only half the story. The other half is written in the contracts that decide who gets paid when the yield arrives.
Putting the Numbers in Everyday Perspective
Think of the reserve assets as a large money-market portfolio that happens to back a digital dollar. Every day the portfolio earns a return. Part of that return is contractually owed to the firms that help keep the digital dollar circulating widely. What remains is the issuer’s working capital before it pays its own people and systems.
In the latest quarter that residual was just over forty cents on the dollar. A year earlier it had been closer to thirty-eight cents. The improvement came even while the overall yield environment softened and while total distribution costs edged higher. That combination is harder to achieve than it looks.
Whether the same pattern holds when the next set of numbers appears will depend on three moving parts: the path of short-term interest rates, the share of USDC that stays inside lower-cost channels, and the commercial terms struck with the largest distribution partners. None of those variables is entirely under Circle’s control, which is precisely why the retained-income figure deserves ongoing attention.
I’ve found that stablecoin discussions often fixate on total market capitalisation or daily transfer volume. Those metrics matter for adoption, yet they say surprisingly little about the issuer’s own economics. The gap between gross reserve income and the amount left after partner payments is where the real operating leverage, or lack of it, lives.
For now the evidence from the second quarter points to a business that managed to keep more of its income even as the underlying yield declined. That outcome is not guaranteed to repeat, but it does demonstrate that the distribution cost structure is not an automatic absorber of every incremental dollar. The coming earnings report will tell us whether that resilience continues or whether the pressure from partners and rates begins to show more clearly in the retained margin.
