Circle’s own rate table sends half of a marginal reserve-income move into distribution costs

Circle's own interest-rate sensitivity table shows that a large part of any marginal reserve-income movement is modeled to flow into distribution and transaction costs. For the twelve months following 30 September 2026, the company estimated that a 100-basis-point change in rates would move reserve income by $739 million and distribution and transaction costs by $371 million. FXBrokers.ai calculates an implied cost pass-through of 50.2%, leaving a modeled $368 million pre-operating-expense difference before tax and any behavioural change.[1][2]

The calculation changes how Circle's rate exposure should be read. A headline sensitivity based only on reserve income would overstate the modeled earnings benefit of higher rates—and the modeled earnings damage of lower rates—because the commercial agreements that distribute USDC also make part of the revenue movement variable. The table is not a forecast. Circle held USDC circulation, asset allocation and the rate change constant in its model. But it is the clearest company-provided quantification of how reserve economics are shared at the margin.[1][2]

The sensitivity that matters

The figures appear in Circle Internet Group's Form 10-Q for the quarter ended 30 September 2026. The company modeled the following twelve months from an average September yield of 3.62%. It assumed USDC circulation remained equal to the period-end amount, the rate move occurred on 1 October and affected all reserve assets equally, and reserve allocation did not change. Circle expressly said it cannot accurately predict circulation or the resulting reserve income and costs because customer behaviour and market dynamics are uncertain.[1][2]

Those assumptions are not minor footnotes. Stablecoin circulation can respond to crypto-market activity, payments adoption, competitive pricing, regulation and the opportunity cost of holding a non-interest-bearing token. Reserve assets can mature or be reallocated. Distribution economics can change as the share of USDC held on Circle's own platform, Coinbase or other channels changes. The sensitivity isolates one variable—rates—while freezing the others. Its value is therefore structural rather than predictive: it reveals how the current commercial model divides a rate movement under stated conditions.[1][2]

Half the marginal movement is not the same as half of all reserve income

The 50.2% ratio applies to the modeled change, not necessarily to every dollar of existing reserve income. Circle's distribution and transaction costs are influenced by reserve income, where USDC is held and the applicable agreements. Fixed or threshold effects may mean the average cost ratio differs from the marginal sensitivity. It would be wrong to multiply total reserve income by 50.2% and claim that the result is an exact partner payment without reconciling the actual expense table.[1][2]

The direction is still revealing. A one-point rate increase does not produce a modeled $739 million improvement available to Circle shareholders before ordinary costs. Under the company's assumptions, $371 million of the movement is accompanied by higher distribution and transaction costs. A one-point decline also lowers those costs, cushioning the revenue loss by the same modeled amount. The arrangements therefore operate as both a revenue share in stronger-rate conditions and a partial shock absorber in weaker-rate conditions.[1][2]

This helps reconcile two truths that can appear inconsistent. Reserve income is Circle's largest revenue source and makes the company highly rate-sensitive. Yet the company does not retain the full economics of the reserves because distribution partners and channels contribute to USDC's circulation and use. The commercial cost is economically connected to the revenue it helps create. Ignoring either side gives an incomplete account of the stablecoin model.[1][2]

A second denominator changed the quarter

Circle reported $864.888 million of total revenue and reserve income in the third quarter, up from $739.759 million a year earlier. Reserve income was $657.624 million, or 76.0% of the total. Other revenue was $207.264 million, but $181.8 million of that came from recognition of ARC token presale revenue as the company progressed in satisfying its performance obligation. The ARC amount represented 87.7% of other revenue and 21.0% of total quarterly revenue.[1][2]

Subtracting ARC presale revenue from the total produces $683.088 million, 7.66% below the prior-year total. That is an analytical scenario, not a GAAP result and not a measure Circle reported. It does not imply that ARC revenue was improper or economically valueless. The company says the presale is part of its ongoing major operations and recognises revenue under the contracts as it mints and distributes tokens and enhances Arc's functionality toward a proof-of-stake model.[1][2]

The subtraction is useful because the reported year-on-year growth combines recurring reserve economics with a large, unusual revenue-recognition event. Without a denominator adjustment, a reader might conclude that the established reserve and service franchise grew by the full difference between $864.9 million and $739.8 million. In fact, reserve income fell $53.6 million year on year, while the ARC item transformed the total comparison. The filing provides both facts; the original calculation makes their combined effect explicit.[1][2]

Why the ARC treatment cannot simply be reversed

ARC token presale proceeds were initially deferred. Circle's performance obligation includes the mint and distribution of ARC and work to enhance its functionality through a transition in consensus mechanism. Revenue recognition therefore depends on management judgments about performance progress. A critic cannot responsibly label the $181.8 million a cash windfall or a completed recurring sale without reading the contract accounting. Nor can an analyst permanently remove it from Circle's business if token issuance is a central activity the company intends to repeat or expand.[1][2]

The proper distinction is between reported GAAP revenue and a scenario used to assess comparability. The GAAP number answers what Circle recognised under its accounting policy. The ex-ARC scenario asks how the quarter compares before that specific presale recognition. Both are legitimate when clearly labelled; only the first is the financial statement. The second should never be presented as 'adjusted revenue' endorsed by the company.[1][2]

What the evidence says about Circle’s economic engine

Circle's filing shows a business in transition. Reserve income still supplied three-quarters of quarterly revenue, and rate movements remain capable of changing annual revenue by hundreds of millions of dollars. At the same time, the marginal economics are shared through distribution costs, and a token presale supplied most of the quarter's non-reserve revenue. The diversification story is therefore real in accounting terms but concentrated in one item during this period.[1][2]

The reserve model also depends on circulation, not rates alone. Circle calculates its reserve return rate using reserve income and the daily average of segregated reserves. USDC circulation determines the scale of those reserves. A lower yield applied to a much larger circulation base can produce more income; a higher yield applied to a shrinking base can produce less. The company's sensitivity correctly avoids pretending it can forecast that interaction, but readers should not mistake a static-balance table for a complete earnings model.[1][2]

What the next filing should disclose

The findings do not establish that Circle's agreements are unfavourable or that its ARC revenue is overstated. They establish two narrower points that materially change interpretation. First, the company's own model assigns roughly half of a marginal rate-driven reserve-income movement to associated distribution and transaction costs. Second, the quarter's reported growth depended heavily on $181.8 million of ARC presale revenue. A defensible view of Circle must therefore model both sides of the rate equation and separate recurring comparison from contract-driven token recognition.[1][2]

Future quarters will test both mechanisms. A rate move can be compared with the actual change in reserve income and distribution costs, while additional ARC recognition can be traced against deferred revenue and disclosed performance progress. Until then, the most reliable measure is the one Circle itself provided: for every modeled $739 million annual reserve-income movement from 100 basis points, $371 million moved in the same direction through distribution and transaction costs.[1][2]

Report an error or submit an update