Arc Shipped and Funded the Beat. The Reaffirmed Guide Prices a 600bp Q4 Margin Cliff.

Results vs. Consensus

Q3 2026 Scorecard

Consensus figures carry unusually wide dispersion this quarter. Four independent providers put revenue between $778.1M and $824.8M and EPS between $0.32 and $0.42, with a fifth at $0.23, because contributors disagreed about how fast Circle would recognize the ARC token presale rather than about the operating business. The table uses the most widely cited set, whose $801.1M revenue figure also happens to sit at the midpoint of the four. There was no published consensus for adjusted EBITDA, adjusted operating expenses or RLDC.

The five line-item estimates that were published, all from a single provider, tell the story more cleanly than the headline does. The Street had reserve income to within 0.3%, had the reserve return rate exactly right, and had both circulation measures slightly too high. It had other revenue at $140.8M against the $207.3M reported. The $66.5M other-revenue beat is larger than the $63.8M total revenue beat only because the line-item estimates do not sum to the headline figure. Either way, essentially all of the beat came from one accounting event, and reserve income, the rest of the revenue line, landed about $2M above its own estimate.

Year-Over-Year Comparisons

The divergence between basic and diluted EPS is worth a sentence because it runs the opposite way from the net income line. Net income rose 5.4%, basic EPS fell 4.3% on a 10.1% increase in the basic share count, and diluted EPS rose 31.3% because the diluted count was essentially flat while the prior-year diluted numerator excluded a fair-value gain on convertible debt that no longer exists. The $0.84 figure is the one the Street scored against, and it is the correct one, but three of the four comparisons in that pair are artifacts of capital structure rather than of operating performance.

Quarter-Over-Quarter Comparisons

Dollar figures in the sequential table are the company's own quarterly series, which is presented in rounded millions and may not foot precisely. Adjusted operating expenses and adjusted EBITDA come from the non-GAAP reconciliations, which are stated to the thousand, so those growth rates are exact. Both adjusted EBITDA definitions used here are the post-Q1 2026 definition that excludes payroll tax on stock compensation, applied consistently across every period in this report.

Revenue Composition and Platform Performance

Circle reports as a single operating segment, so the meaningful decomposition is by revenue line and by network. The revenue mix shifted more in one quarter than it has at any point since the IPO, and almost all of that shift is one transaction.

Revenue Lines

Reserve income: the rate cycle turned, the float did not quite

Reserve income of $657.6M fell 7.5% year-over-year as a 61 basis point decline in the reserve return rate overwhelmed 8% growth in average circulation. The sequential picture is the more interesting one. Average circulation fell 4% from Q2's record to $73.3B, but the reserve return rate rose 6 basis points to 3.54% following a Federal Reserve rate hike in September, the first sequential increase since Q3 2025. Reserve income landed within 0.3% of consensus.

"The reserve return rate was 3.54% for the quarter, down 61 basis points year-over-year. Sequentially, average rates rose 6 basis points following the Federal Reserve's rate hike in September." — Jeremy Fox-Geen, CFO

Assessment: The rate bear case, which we carried as the primary risk through four quarters and marked Contained in August, has now inverted. A rising policy rate is a direct tailwind to this line, and it arrives exactly as the float stabilizes. The constraint on reserve income has moved decisively from the numerator to the share of it Circle keeps.

ARC token presale: a $242M financing event recognized as revenue

Circle sold 807.5 million ARC tokens at $0.30 in Q2 for $242.2M of gross proceeds, which implies a $3.0B value for the ten-billion-token supply and $750M for Circle's retained 25%. With the genesis mint and the September 16 Mainnet launch, the company recognized $181.8M, roughly 75% of the total, in a single quarter. The remaining $60.4M sits in deferred revenue inside other current liabilities and is contingent on the final performance obligation, which the 10-Q describes as "the enhancement of its functionality on the Arc network through transition to a Proof-of-Stake consensus mechanism." Management puts that transition in 2027. Within the presale, $25.0M of tokens were sold to an entity affiliated with a holder of more than 5% of Circle's capital stock.

Assessment: This is revenue under GAAP because token sales are part of Circle's ongoing operations, and we do not dispute the accounting. It is not recurring revenue, it carries no cost of any kind, and capitalizing it at a business multiple double-counts a financing event against an asset Circle still holds 25% of off balance sheet. Every valuation line in this report therefore runs on a clean adjusted EBITDA that excludes it, which is the single most important modelling decision we carry on this name.

Other revenue excluding the presale: the diversification pillar is now a 2027 story

Strip the presale and the other-revenue line was $25.5M, down roughly 11% year-over-year and down about $8M sequentially. Management attributed the decline to monetizing fewer blockchain integrations than in Q2, and guided several for Q4. This is the second consecutive quarter in which the line that was supposed to prove Circle is more than a rate-sensitive reserve manager has gone backwards.

"Excluding Arc, other revenue was $25 million, down 11% year-over-year and down approximately $8 million sequentially as we monetized fewer blockchain integrations in the quarter compared to Q2." — Jeremy Fox-Geen, CFO

Assessment: At $25.5M a quarter this line is 2.9% of revenue and is shrinking. The forward case rests entirely on CPN monetization and Arc staking, both of which management dates to 2027 and neither of which has a disclosed take rate or a dollar figure attached. The diversification pillar is not broken, but it is now a promise with a date rather than a trend with a slope.

Network Metrics

USDC: the float stopped shrinking and usage kept its lead

End-of-period circulation rose about 1% sequentially to $73.9B, which reverses Q2's decline and clears the explicit test we set in August. Meaningful wallets rose to 7.3M from 7.0M, and market share held at 27%. Both circulation measures still came in below consensus, average circulation fell 4% from Q2's record, and management acknowledged the year-over-year position directly.

"USDC in circulation ended the quarter at about $74 billion, flat year-over-year in a soft digital asset market. I want to be direct about that. Balances were flat, but usage continues to grow." — Jeremy Allaire, Co-Founder, CEO and Chairman

The usage figures are genuinely striking. Onchain transaction volume grew 49% to $13.9 trillion, USDC carried 79% of all stablecoin transaction volume against 66% a year earlier while holding roughly 27% of the supply, and primary-market activity set a record at $208B of mints and redemptions, about $2.3B a day against $1.9B a day in Q2. Management translated that into a velocity figure of roughly 190 turns in the quarter, about ten times the next largest stablecoin. Onchain volume did slip about 6% from Q2's $14.8 trillion, so the turnover rate was roughly unchanged sequentially.

Assessment: The decoupling of usage from supply that we flagged in August persists: transaction share widened year-over-year even though volume slipped sequentially. That is a real competitive position and the primary-liquidity infrastructure behind it is the hardest part of this business to replicate. It is also, so far, unmonetized: Circle earns on balances held, not on volume moved, and the velocity story only converts to earnings if settlement activity eventually parks its working capital in USDC. Management said as much when it described the current pattern as a "stablecoin sandwich" it wants to make stickier.

Circle Payments Network: the fastest compounding thing in the business

CPN annualized payment volume on a trailing 30-day basis reached $41B at quarter end, nearly triple the level at the end of June and more than fifteen times a year earlier, with enrolled financial institutions up more than fourfold to 205. Management then gave a post-quarter figure that is larger than the quarter's own disclosure: more than $58B as of October 30, against $23B reported on July 31.

"Sitting here today, as of October 30, annualized payment volume has already surpassed $58 billion. That is more than double the $23 billion we reported as of July 31 on our last earnings call, delivering over 150% sequential growth." — Jeremy Allaire, Co-Founder, CEO and Chairman

Assessment: This is the one line in the business that is growing fast enough to change the thesis, and it is also the one with no disclosed economics whatsoever. Management confirmed monetization has begun and promised "material revenue from CPN in 2027," but gave no take rate, no revenue dollars and no 2027 figure. A $58B annualized volume at even five basis points is $29M of annual revenue, which would add roughly a quarter to the current non-presale other-revenue line at its annualized rate; at twenty-five basis points it is $145M. The spread between those two outcomes is most of the 2027 debate, and Circle chose not to narrow it.

Arc: the network shipped on the committed date and is carrying real volume

Arc public mainnet launched September 16, the exact date management committed to in August, with 11 founding validators, more than 100 institutions and applications, native USDC integration and fees payable in USDC. Late-October metrics given on the call: roughly $1.5B of USDC, cirBTC and EURC on the network, more than $30B of USDC moved since launch and recently about $1B a day, around 500,000 transactions a day with subsecond finality and fees under a cent. The developer funnel behind it includes more than 75,000 Arc House members and 1,200 projects built during private mainnet.

Assessment: We set "Mainnet ships September 16 with observable transaction volume" as the first of eight August watch items, and it is the one that passed most cleanly. A validator cohort drawn from the institutions that clear and settle global finance is not replicable on any relevant timeframe, and 500,000 daily transactions about six weeks after launch is a live network rather than a press release. What Arc has not yet produced is recurring revenue. Staking rewards, which management calls the primary recurring source, require the Proof-of-Stake transition, which is a 2027 event and is also the gate on the remaining $60.4M of presale recognition.

Key Topics & Management Commentary

Overall Management Tone: Management was confident on strategy and conspicuously selective on economics. Prepared remarks ran long on platform milestones and the January GENIUS Act inflection and allocated almost no time to the margin trajectory, the workforce reduction or the European supervisory change, all of which were disclosed in filings the same morning. The only quantitative forward signal was given indirectly, as an invitation to back-solve the Q4 margin step-down from a reaffirmed annual range. Tone was more assured than Q2's defensive posture and less forthcoming on numbers.

1. The ARC Token Presale Became 21% of Revenue in One Quarter

Circle recognized $181.8M of the $242.2M presale in Q3, about 75% of the total, triggered by the genesis mint of all ten billion tokens and the public mainnet launch. That single entry is 21.0% of quarterly revenue, carries no distribution cost, and is the entire reason reported RLDC margin printed 52.4% against 41.2% in Q2. The CFO was explicit about the mechanics and about the recognition trigger.

"As we noted last quarter, ARC Token presale revenue is recognized as certain performance milestones are met. With the minting and delivery of ARC Tokens and the public mainnet launch in Q3, we recognized approximately 75% of the total presale revenue in the quarter." — Jeremy Fox-Geen, CFO

The pace matters more than the amount. In August the full-year other-revenue guide of $310M to $330M contained roughly $180M of Arc recognition against an unspecified milestone schedule, and the Street spread that across two quarters: published other-revenue consensus for Q3 was $140.8M. Circle took almost all of it at once. The full-year guide did not move, which means the entire $66.5M other-revenue beat was pulled forward out of Q4 rather than added to the year.

Assessment: A beat produced by accelerating the recognition of a fixed pool of non-recurring proceeds is not a beat, and the market priced it that way within hours. The more useful fact is the residual: $60.4M remains deferred against the Proof-of-Stake transition, so there is one more recognition event of known size sitting in 2027 that is not in any current guidance.

2. The Reaffirmed Guide Prices a Q4 Distribution-Cost Cliff

Every element of full-year guidance was held: other revenue at $310M to $330M including the presale and $130M to $150M excluding it, RLDC margin at 41.7% to 43.7% including the presale, adjusted operating expenses at $570M to $585M, and the multi-year 40% circulation CAGR framework. The CFO then added the ex-presale margin anchor and told the Street to do the arithmetic.

"Excluding Arc revenue, we expect full year RLDC margin to come in near the midpoint of our original guidance range of 38% to 40%. As I mentioned earlier, Q4 will be the first full quarter, reflecting both our Hyperliquid arrangement and our renewed Binance partnership." — Jeremy Fox-Geen, CFO

Assessment: This is the most consequential number in the quarter and it was never stated aloud. A reaffirmed annual range sounds like stability and in this case encodes a one-quarter collapse in the economics of the core business. We have moved the distribution-cost risk to the most severe tag we carry on the back of it.

3. Binance: Five Years of Distribution, Bought With Margin and Equity

Circle signed a new five-year agreement with Binance alongside a $100M equity investment by Binance into Circle. The 10-Q gives the terms the call did not: 1,237,011 Class A shares issued on September 17 at $80.84, with transfer restrictions running to the earlier of the second anniversary or Binance's termination of the commercial arrangements. Binance brings more than 300 million registered users, 77% of them in emerging markets. Circle disclosed no pass-through rate, no incentive structure and no revenue-share terms.

"We want these companies to be very incented to lean in because we believe and history has shown that the greater the activity on those platforms, the greater the activity and growth of the USDC ecosystem overall, i.e., in USDC that is not held within those highly incented platform infrastructures." — Jeremy Fox-Geen, CFO

The structural logic is the Coinbase template extended to a second venue: pay away most or all of the reserve income on balances held at the venue, in exchange for the venue making USDC the default and generating ecosystem balances elsewhere that Circle keeps more of. It worked with Coinbase over several years. It has not yet worked with Hyperliquid, where 90% of the balances landed on Coinbase's platform at full pass-through, and Q3's 80 basis point decline in net reserve margin is that arrangement's first full-quarter cost.

Assessment: The equity cheque is the part worth weighing carefully. Binance paid $80.84, which is 16.3% above the current price, and the lock-up unwinds early if Binance walks away from the commercial deal. That structure converts a strategic investment into something closer to a collateralized commitment, which is a better alignment than a plain minority stake. What it does not do is answer whether the margin Circle is giving up buys float that stays. We will know in ninety days.

4. Tazapay: Buying the Last Mile With $400M of Stock

Circle agreed on September 4 to acquire Tazapay, a Singapore-headquartered business-to-business cross-border payments platform, for $400.0M payable in Class A stock at a 20-trading-day volume-weighted average price immediately preceding closing, subject to Monetary Authority of Singapore approval and expected to close in 2027. Tazapay brings more than $25B of annualized payment volume, more than 60 banking and fintech partners, local payout rails across more than 100 markets, and roughly 60% of its volume already involving stablecoins. It has been a CPN design partner since 2025.

"The last mile, where a payment lands in a local currency in someone's account is critical in delivering value in global payments. Tazapay expands CPN's footprint, deepens our fiat connectivity and adds new capabilities for the payment providers and financial institutions building on CPN." — Jeremy Allaire, Co-Founder, CEO and Chairman

Assessment: Strategically this is the right purchase: CPN's constraint is termination into local currency, and Tazapay is the termination layer with the volume already stablecoin-denominated. Financially it is a reminder of how the consideration structure cuts. Fixing the price in dollars and settling in stock at a trailing average means a lower share price produces more dilution, roughly 5.7 million shares at $70 and roughly 8.0 million at $50. Buying assets with a de-rating currency is the cost of a de-rating.

5. The CFO Is Leaving, With No Successor Named

The CFO's departure after five and a half years was announced in the prepared remarks, in the form of the CEO thanking him. No successor was named, no interim arrangement was described and no timeline was given. The only filed trace is a Post-Termination Restrictive Covenant Agreement attached as an exhibit to the 10-Q; no Item 5.02 Form 8-K had been filed as of this writing.

"The last 5.5 years have been the honor of a lifetime, and I could not be more proud of what we've accomplished together. While I'm excited about my next chapter, I'll leave Circle in a position of strength and with tremendous conviction in the company's future." — Jeremy Fox-Geen, CFO

Assessment: Nothing in the disclosure or the tone suggests anything other than a planned exit, and most analysts on the call opened with a personal good wish, which is not how a contentious departure sounds. The timing is the issue rather than the fact. The next earnings call is the one where Circle sets its first 2027 other-revenue, RLDC margin and operating expense guidance, on a business whose distribution economics are about to reset and whose growth case now depends on two revenue lines that have never been quantified. Handing that call to a new or interim finance chief raises the variance on guidance that was already going to be the most important of the year.

6. A 5% Workforce Reduction That Nobody Mentioned

A footnote in the release and the 10-Q discloses "charges related to a workforce reduction of approximately 5% designed to better align our organizational structure with our operating model and strategic priorities," costing $9.145M in restructuring expense, which the release describes as primarily severance and notice-period pay, against roughly $5.7M of offsetting stock-compensation reversals from RSU forfeitures. The 10-Q's compensation discussion cites an $8.2M increase in severance tied to the cut. The words "workforce," "restructuring" and "severance" appear nowhere in the prepared remarks or in Q&A.

The cut shows up in the numbers as the first sequential decline in adjusted operating expenses across the five quarters Circle has disclosed on the current definition, down 5.2% to $138.8M, though management attributed that decline to non-recurring Q2 costs rather than to headcount. No headcount figure and no annualized savings estimate were given in any document.

Assessment: Cutting 5% of staff in the same quarter you launch your largest product since USDC, sign a five-year distribution agreement and agree to a $400M acquisition is a deliberate reallocation, not distress, and the expense line confirms it is working. Declining to discuss it is the part that is hard to defend. A company guiding to the high end of its operating expense range while quietly shrinking headcount owes the Street the arithmetic connecting the two, and the implied Q4 adjusted operating expense of roughly $164M makes that arithmetic urgent.

7. MiCA: USDC Becomes Europe's First Significant E-Money Token

The 10-Q subsequent-events note discloses that on October 14 the European Banking Authority classified USDC, as issued in the EU by Circle's French subsidiary, as a significant e-money token under the Markets in Crypto-Assets Regulation, the first such classification ever granted. Effective November 12, supervision of own funds, reserve of assets, liquidity requirements and recovery and redemption plans transfers from the French prudential regulator to the EBA, which has also signed a memorandum of understanding with the New York State Department of Financial Services on cross-border stablecoin cooperation. The filing states plainly that the classification "may be subject to increased regulatory capital, reserve, liquidity, and other prudential requirements under MiCA, which could require us to hold additional capital or liquidity in the European Union, increase our compliance costs."

Assessment: This cuts both ways and only one side was presented. Being the first and only significant e-money token in Europe is a competitive moat of exactly the kind this thesis is built on, and the EBA-NYDFS memorandum is the regulatory reciprocity management spent the call's opening minutes describing as the unlock for institutional adoption. It also carries an unquantified capital and liquidity cost that takes effect eight days after the call, on a company whose own filing says the impact is still being evaluated. Neither the upside nor the cost was raised in prepared remarks or Q&A, which is a disclosure choice rather than an oversight.

8. GENIUS in January: The Inflection Management Is Underwriting

The CEO opened and closed on the January 2027 effective date of the GENIUS Act, framing it as the moment digital dollars become a federally recognized part of the US financial system and the precondition large institutions have been waiting for. He added a specific parallel development: the Financial Accounting Standards Board is near final on accounting treatment for payment stablecoins issued under GENIUS, creating a framework that treats them as cash or cash equivalents.

"And if you're a large global company, if you're a public company, if you're a treasurer or a CFO of a public company or you run a regulated financial institution, you are not going to start building on digital dollars and weaving them into your infrastructure, even if they're better, faster, cheaper, even if they have new innovation and utility until they are codified into the financial system." — Jeremy Allaire, Co-Founder, CEO and Chairman

The CFO's version was more operational and more useful, because it listed the conditions rather than the promise: codification into law, clear accounting and tax rules, and integration into the treasury management systems and ERP software that corporate finance teams already use. Circle pointed at the Kyriba and Tereina integrations, the latter reaching SAP customers, as evidence that the third condition is being built now.

Assessment: We give this more weight than we did at Q2, because the accounting-standards leg is new information and because the integration leg is observable rather than aspirational. Cash-equivalent treatment is what moves stablecoins from a payments use case to a treasury use case, and treasury balances are sticky in a way that trading balances are not. The relevant caution is duration. None of this shows up in a Q1 2027 print, and management's own circulation framework is a through-cycle CAGR with the explicit caveat that "growth will not follow a straight line."

9. The Agentic Economy and a Q1 2027 Commitment

Management devoted a substantial block of prepared remarks to agent-to-agent payments, and attached a date to it: by the first quarter of 2027, an agent built on Arc can be found, hired, paid and trusted for the work it does. The supporting data points are that more than 99% of agent payments on the x402 protocol settle in USDC and that nearly 2,800 endpoints are listed in Circle's agent marketplace. Arc's roadmap includes a dedicated agent sector with verified identity, mandates and audit trails, alongside a privacy sector and a payment sector targeting more than 100,000 transactions per second.

Pressed on how Circle fits alongside consumer AI assistants that use card rails, management drew a distinction between the first wave of shopping agents, which it expects to ride existing card products, and the step change it is building for, where agents transact with other agents in small amounts under programmatic contracts.

Assessment: The 99% settlement share on x402 is a real standards position and the cheapest optionality in the story, because Circle is not spending much to hold it. It is also unmonetized and unquantified, and the specific Q1 2027 claim is a product commitment rather than a revenue commitment. We carry it as free optionality inside the Arc pillar rather than as a separate pillar, and we will treat the first disclosed dollar of agent-driven revenue as the signal.

10. Operating Expenses Turned, and Then the Guide Re-Armed Them

Adjusted operating expenses fell 5.2% sequentially to $138.8M, the first decline across the five quarters disclosed on the current definition, against our August expectation of another step up. GAAP operating expenses fell 3.8%. Inside the GAAP line, compensation expense fell year-over-year as $16.0M of lower stock compensation from RSU forfeitures and legacy award roll-off more than absorbed $8.2M of severance and $2.7M of higher salaries. General and administrative expenses rose 30.9% year-over-year, driven by $10.0M of additional legal, professional and consulting fees.

"On adjusted operating expenses, we continue trending towards the high end of our original $570 million to $585 million range." — Jeremy Fox-Geen, CFO

Assessment: Nine months of adjusted operating expenses total $420.9M, so the high end of the guide implies roughly $164M in Q4, up about 18% sequentially and about 24% year-over-year, in the quarter immediately after a 5% headcount reduction. Either the guide is conservative, in which case Q4 clean earnings land at the upper end of our $70M to $90M derived range, or the Chelsea marketing commitment and Arc infrastructure spending genuinely consume the savings. Management had the opportunity to resolve that and chose the broad range.

11. Competition: Management Answered a Question Nobody Is Asking

Asked by a retail shareholder about returning circulation to sustained growth while protecting margins, management's competitive framing was entirely about transaction volume and liquidity depth.

"I think a recently launched stablecoin is doing maybe $5 million a day of onchain transaction volume compared to our tens of billions or maybe a couple of hundred thousand dollars of daily trading volume versus, again, sometimes our nearly $10 billion of trading volume. So liquidity and the infrastructure we have is a huge competitive advantage." — Jeremy Allaire, Co-Founder, CEO and Chairman

The bank and payments consortium that reopened our competitive bear point in August, whose premise is paying distributors nearly all of the reserve income rather than building a better product, was not mentioned once in prepared remarks or Q&A.

Assessment: Circle is winning the volume war decisively and that is not the war. The threat to this business model has never been that a rival stablecoin out-transacts USDC; it is that the existence of a published alternative that pays away nearly all reserve income resets the price of every distribution renewal. The Q4 margin step-down embedded in the reaffirmed guide is what that threat looks like when it arrives, and management answered a liquidity question instead.

Guidance & Outlook

Nothing moved, which is itself the news. The composition underneath changed substantially: the presale came in faster than modelled, the non-presale other-revenue line came in slower, and the distribution economics stepped down. Holding all four ranges against those offsetting moves is defensible arithmetic and a choice not to signal.

Implied Q4 ramp. Other revenue excluding the presale must land between roughly $29M and $49M against $25.5M in Q3, consistent with management's promise of several blockchain integrations after only one in Q3. Adjusted operating expenses must land between $149M and $164M against $138.8M. The ex-presale RLDC margin must fall to roughly 34% from 39.7%. Clean adjusted EBITDA therefore lands at roughly $70M to $90M against $134.9M in Q3 and $175.9M in Q4 2025, which puts reported full-year adjusted EBITDA near $682M to $702M and clean full-year adjusted EBITDA near $500M to $520M.

Street at. There is no published consensus for RLDC margin or adjusted EBITDA to compare against, which is part of why nine of eleven desks cut price targets on the print rather than re-rating in either direction. The eleven post-print revisions span $46 to $103 against a $69.48 close, a 2.2x spread and a fair reflection of how little the guide constrains 2027.

Guidance style. Circle has now held this set of ranges through two quarters of substantial composition change, which reads as a company unwilling to re-cut guidance mid-year rather than one with high visibility. The pattern to watch is that the ex-presale RLDC margin guide was set at 38% to 40% at the start of the year and held near the midpoint in August and again now, while the actual nine-month figure runs at 40.8%. The guide has been a ceiling all year and Q4 is where it becomes binding.

Analyst Q&A Highlights

The Economics of the Renewed Binance Agreement

The dominant topic in Q&A was what Circle is paying for the Binance relationship. Management declined to break out terms, characterizing the arrangement as a marketing, distribution and product relationship in which the counterparty has "a strong economic incentive to market, distribute and integrate USDC as a primary digital dollar." The finance side of the answer was the most informative moment of the call, and it was a direction to do arithmetic rather than a figure.

Q: "I wanted to touch on the Binance partnership. Could you speak to the drivers of how the updated partnership was structured as well as any detail on the economics to the degree that you're willing to comment on that? And finally, do any of the recent events at Binance impact your partnership at all?"
— James Yaro, Goldman Sachs

A: "On the economic specifics, while we haven't called out the details of the Binance agreement specifically, I did say in my comments on guidance that the full year margin, excluding the ARC Token presale was trending towards the middle of the original range we gave at the beginning of the year. You can do the kind of backward math from that, and you'll see the step down that we anticipate in Q4 due to the full quarter of the Binance incentive arrangement being recognized."
— Jeremy Fox-Geen, CFO

Assessment: Management confirmed the existence and the mechanism of a Q4 margin step-down while refusing to size it, and explicitly delegated the sizing to the Street. That is a defensible way to avoid issuing a quarterly guide, and it is also an acknowledgement that the number is large enough to need pre-warning. Our back-solve puts the Q4 ex-presale RLDC margin near 34% and the net reserve margin near 30%. Nobody on the call pushed back on that range, because nobody had it yet.

Why Non-Coinbase Distribution Costs Fell While the Coinbase Relationship Grew

A pointed question tried to reconcile the 8.5% year-over-year decline in distribution and transaction costs against a Hyperliquid arrangement whose balances sit on Coinbase's platform at full pass-through, and separately asked whether the Binance renewal carried any upfront cost. Management answered the second question cleanly and redirected the first back to the net reserve margin.

Q: "It looks like your distribution costs declined this quarter despite Coinbase revenue increasing with the hyperliquid agreement. Can you explain what's driving the decline in non-Coinbase distribution costs? And then as a quick one, is there any upfront costs associated with the Binance renewal?"
— Edward Engel, Compass Point

A: "There are no upfront costs associated with the Binance agreement. And we don't comment specifically on kind of the puts and takes that we've talked about in general terms many times in each of our different distribution agreements. We have many of them and some of them ebb and flow with different amounts of stock and different amounts of incentive through any quarter. What you saw though defining net reserve margin this quarter, we talked about at the beginning was the sequential decline in margin, which is largely due to the Hyperliquid agreement, which balances are on Coinbase. So you can kind of work through the math on that one."
— Jeremy Fox-Geen, CFO

Assessment: The confirmation that Binance carries no upfront cost is genuinely useful, because it means the entire cost is an ongoing margin item rather than a one-time charge, which is worse for the run rate and better for the balance sheet. The unanswered half matters more. The 10-Q supplies what the call did not: distribution costs fell $38.0M year-over-year because Coinbase's payment fell $30.0M on lower reserve income, partially offset by a slight increase in Coinbase's share of average on-platform balances, with $8.0M coming from reduced activity with other partners. The headline decline is a rate artifact, and the Coinbase share went the wrong way.

CPN's Revenue Model and the Timing of Arc's Proof-of-Stake Transition

The final question of the call asked the one thing the whole 2027 case rests on: how CPN actually makes money and when Arc moves to proof of stake. The answer named 2027 twice, tied the revenue to the Tazapay close, and deferred the figure to the next call.

Q: "Can you talk about the economic and monetary model on CPN and sort of how it will sort of help turn that flywheel with the rest of the USDC ecosystem and perhaps also comment on timing for Arc to move to proof of stake."
— Andrew Jeffrey, William Blair

A: "Again, we're not commenting on kind of 2027 guidance just yet. But looking back to what we said last quarter, we are now beginning to monetize CPN, and this is a priority for us going forward. So we have started to monetize some of the CPN flows, and that is going to grow over time. And then next year as well with the closing of the Tazapay acquisition and the integration of Taza into CPN overall, we expect to see material revenue from CPN in 2027, and we'll be talking about that more at our next earnings call when we give kind of other revenue guidance for the year."
— Jeremy Fox-Geen, CFO

Assessment: "Material" with no number, on an acquisition that has not closed and needs a Singaporean regulatory approval, is the thinnest possible support for the pillar carrying the most weight in the bull case. The commitment is real and dated, and the volume growth behind it is extraordinary, but it converts a 2027 earnings estimate into a judgement about take rates nobody has disclosed. This exchange is the single largest reason our 2027 range stays wide.

Why Payments Comparables Are the Wrong Frame for Circle's Economics

Immediately after deferring the CPN figure, management volunteered an unprompted and unusually direct warning about how the Street should model payment economics at all. It is the closest thing on the call to guidance about the shape of the long-term margin structure, and it points down.

A: "And so I think anyone who's sort of taking a comp and saying, “Hey, this payments company makes 10 to 15 basis points or this payments company makes 20 to 30 basis points,” that's the wrong comp..."
— Jeremy Allaire, Co-Founder, CEO and Chairman

A: "Or 2%..."
— Jeremy Fox-Geen, CFO

A: "Or 2% or whatever it is. That's the wrong comp. When this achieves scale, it's going to have enormous velocity and the unit costs are going to be much, much lower. ... We are not shooting to — let's go repeat the margin structure of the existing payments industry. This is about compressing it, compressing it materially, and that is what we're shooting for. And so when you think about your own modeling for what that looks like, we want these bit numbers to be very, very big with huge network effects and huge velocity with very, very low unit costs."
— Jeremy Allaire, Co-Founder, CEO and Chairman

Assessment: This is strategically coherent and financially sobering, and it deserves to be taken at face value rather than read as modesty. Management is telling the Street not to apply a twenty-basis-point take rate to CPN volume, which caps the most common bull-case calculation on this name at something closer to five basis points. On $58B of annualized volume that is roughly $29M, not $145M. Volunteering a downward anchor on your own fastest-growing line, in the same answer as declining to guide it, is either unusual candour or an early attempt to manage expectations ahead of a 2027 guide nobody will like.

Whether AI-Driven Security Failures Threaten the Distribution Base

A recurring line of questioning on the call connected the previous week's large exchange breach to the durability of the exchange partnerships USDC's distribution depends on. Management's answer covered industry coordination, law enforcement and government response, and then pivoted to an argument that a permissioned validator set is itself a security feature.

Q: "From our standpoint, we see some of these hacking risks, especially driven by AI is the biggest concern for onchain bearer assets in 2027. So is this an industry wake-up call? And is there consortiums you guys can develop? ... And you guys are kind of less directly involved, but if you start seeing customers pull more assets from exchanges, it obviously impacts these partnerships around USDC."
— John Todaro, Needham

A: "... a blockchain infrastructure that is designed with prudential safety in mind that is operated by a known set of actors whose infrastructure are held to high standards of compliance and information security. ... It's like the early days of cloud. Cloud early on, the CIA wasn't going to run their stuff on it or major banks weren't going to run their stuff on it. And eventually, the standards got strong enough. We're going to go through that change in blockchain infrastructure itself."
— Jeremy Allaire, Co-Founder, CEO and Chairman

Assessment: The cloud analogy is the best articulation yet of why a permissioned chain operated by regulated institutions is a product rather than a compromise, and it reframes Arc's validator set from a governance concession into a differentiator. The exchange-concentration half of the question went largely unanswered. Management offered a reassuring same-day deposit figure for the affected venue but did not engage with the structural point, which is that Circle has just committed five years of distribution economics to the exact category of counterparty the question identified as 2027's largest operational risk.

Where the Industry Is Furthest From the Infrastructure Vision

The call opened with a question about which part of the stack remains furthest from the "blockchain fades into the background" thesis and what single bottleneck consumes the most resources. Management answered by listing what is already done rather than what is missing, and delivered an inventory of mainstream integrations as evidence.

Q: "I guess if we look across Circle's investments across USDC, CPN, Arc, agent stack, enterprise partnerships, what have you, where do you think the industry is still furthest from that vision? And what's the single biggest bottleneck that Circle is spending the most time and resources trying to solve right now?"
— Peter Christiansen, Citi

A: "I mean we didn't talk about some of this, but Lloyds Bank settling cross-border payments with USDC, Visa connecting their global money movement platform, Visa Direct to be funded in USDC. Walmart's finance subsidiary, OnePay, adding support for USDC. Samsung, putting USDC in all of their mobile phones and their mobile wallets. These are big milestones that, again, even when we went public 1.5 years ago, would have been watershed moments. These are happening every week, every month, lots of these things happening."
— Jeremy Allaire, Co-Founder, CEO and Chairman

Assessment: The integration list is impressive and verifiable and it is also a non-answer to the question asked. A bottleneck was requested and three completed checkboxes were supplied. The pattern recurs throughout this call: strategy questions receive expansive answers, and questions whose honest answer is a number receive redirection. The commercial substance stands on its own, though, and the Samsung and OnePay channels in particular reach consumer distribution that no stablecoin has previously had.

Corporate Treasury Readiness Ahead of January

A question about how aware large corporates actually are of the post-GENIUS opportunity produced the most concrete forward framework on the call, in the form of a checklist of preconditions rather than a forecast.

Q: "Just how aware do you think Fortune 500 Global 2000s are of this opportunity to integrate stablecoins? How do you see this kind of play out a little bit here after GENIUS is enacted. Just kind of looking for maybe your views of how this evolves over the next 6 to 12 months."
— Joseph Vafi, Canaccord

A: "Necessary things to happen before CFOs and treasurers will use digital dollars, Jeremy mentioned several of them. Codification into law bringing digital dollars into the financial system, clear rules for accounting and tax, that isn't — the first has happened. Those are in the works. Very importantly, integration into the existing systems that treasurers use today, right? We talked about integration into software that works with SAP, treasury management systems."
— Jeremy Fox-Geen, CFO

Assessment: This is the most useful thing a Circle executive said about 2027 demand, because it is falsifiable. Three preconditions were named, one is complete in January, one is near-final at the standards-setting level, and the third is observable in the ERP and treasury-management integrations Circle has already shipped. It also implies the honest answer to the question is "not very aware yet," and it sets the earliest plausible window for corporate treasury balances to appear in circulation data at the back half of 2027 rather than the front.

What They're NOT Saying

1. The size of the Q4 margin step-down: confirmed to exist, attributed to a specific cause, and deliberately left for the Street to back-solve.
2. Any 2027 guidance at all: explicitly deferred on two separate questions. The business is one quarter from a full-year guide on a cost base that is resetting and two revenue lines that have never been quantified.
3. CPN economics: $58B of annualized volume as of October 30, monetization confirmed to have started, "material revenue" promised for 2027, and not one take rate, revenue dollar or margin figure disclosed.
4. The 5% workforce reduction: $9.1M of restructuring charges, mostly severance, disclosed only in a release footnote and the 10-Q, with no headcount figure, no annualized savings and no mention in prepared remarks or Q&A.
5. The CFO succession plan: no successor, no interim arrangement, no timeline, and no Item 5.02 Form 8-K as of this writing. Only a restrictive-covenant agreement attached as a 10-Q exhibit.
6. The MiCA significant-token classification: granted October 14, effective November 12, with possible incremental European capital and liquidity requirements that the filing says are still being evaluated. Absent from the release and from the call.
7. Binance's economic terms: no pass-through rate, no incentive mechanism, no revenue share. The one disclosed fact, that there is no upfront cost, means the entire cost is recurring.
8. The remaining $60.4M of presale revenue: the deferred balance and its Proof-of-Stake trigger appear in the 10-Q. Neither the amount nor the expected recognition timing was raised on the call.
9. The competing distribution consortium: zero mentions. The competitive question on the call was answered with transaction-volume comparisons against a different and much smaller threat.
10. Last quarter's named partnerships: the DTCC tokenization work and the BlackRock BUIDL integration, both flagged in August as moving from "collaborating" to live volume, went unmentioned.
11. Capital return: no repurchase authorization anywhere in the release, the 10-Q or the call, against $1.70B of corporate cash and a stock down 37.5% in twelve months. This was a monitor item we set in August and it remains unaddressed.
12. Any mark on the ARC token: Circle holds 25% of a network it values at $3.0B on presale pricing, carried at zero under GAAP, with no secondary reference price offered.

Market Reaction

• Pre-print setup: closed $75.22 on November 3, down 5.1% year to date from $79.30, down 9.7% over the trailing thirty days from $83.30 and down 32.4% over the trailing twelve months from $111.25, against a 52-week closing range of $50.23 to $132.84. The S&P 500 entered the print up 10.6% year to date.
• Reaction session, November 4 (before the open, so the print day is the reaction day): gapped down to $74.04, a 1.6% decline, traded $72.38 to $77.59 and closed $76.12, up 1.2% or $0.90. Volume of 14.5M shares was 1.4 times the 30-day average. The S&P 500 rose 1.6% the same session.
• Follow-through session, November 5: opened $73.50, traded $64.81 to $74.40 and closed $69.48, down 8.7% on 17.3M shares, or 1.7 times average volume. The S&P 500 fell 4.4% that session in a broad risk-off move.
• Analyst actions, November 5: of eleven published target revisions, nine were cuts ranging from 4% to 23%, one was a raise from the lowest target among the revisions while retaining a negative rating, and one was a rating upgrade. The revised targets span the mid-$40s to the low $100s.

The Print Day Marked the Beat to Zero

A 100% EPS beat, an 8% revenue beat and an 85% increase in reported adjusted EBITDA produced a 1.2% gain on a day the index rose 1.6%. On the cleanest available read, the market gave Circle less than nothing for the quarter. The stock opened down 1.6%, after both the release and the 8:00 a.m. call, then recovered through the session to close up 1.2%. That pattern is consistent with investors discounting the token revenue and paying for the platform metrics, but daily price bars cannot show what drove the recovery.

The Second Session Cannot Be Attributed to the Print

November 5 took the stock down 8.7% to $69.48, below the pre-print level, but three things hit that session simultaneously: nine target cuts in the morning, a market-wide decline of 4.4% in the worst index session since April 2025, and no new company disclosure. Circle fell roughly twice the index, which is consistent with its beta profile in a risk-off tape and does not require a company-specific explanation. We therefore treat November 4 as the market's verdict on the quarter and November 5 as noise with an unfavourable direction. The honest summary is that the print was worth approximately nothing and the subsequent de-rating belongs to the tape.

What the Target Cuts Were Actually Repricing

The pattern in the revisions is informative because the ratings did not move with the targets. Ten of eleven desks left their rating unchanged, and nine of them cut the number, which is what happens when an earnings base comes down without the story changing. The one upgrade came with an explicitly cautious framing about the speed of mainstream adoption, and the one target increase came from the lowest target among the post-print revisions and still sits roughly 34% below the current price. A revised target range running from the mid-$40s to the low $100s on the same quarter's disclosures is a direct measure of how little the reaffirmed guide constrains 2027.

Street Perspective

Debate: Is the ARC Token a Revenue Line or a Financing Event?

Bull view: The sale is revenue under GAAP because token issuance is part of Circle's central operations, and treating it as financing understates the business. Circle created a $3.0B network asset, sold 8% of it for $242M, retained 25% carried at zero, and is building toward staking rewards that management describes as the primary recurring source. The $60.4M still deferred is a known future recognition. Pricing the company as though the token franchise is worth nothing because the first monetization was lumpy misses an asset that no competitor has.

Bear view: A one-time sale of an asset at a price the seller set, to buyers including a holder of more than 5% of the seller's stock, recognized 75% in the quarter the seller chose to mint, is a financing event wearing a revenue label. It has no cost, no recurrence and no relationship to operating performance, and capitalizing it at a business multiple double-counts value Circle still owns. Strip it and revenue fell 8% and adjusted EBITDA fell 21%, which is the actual trajectory the stock has to discount.

Our take: The bear framing is correct for valuation and the bull framing is correct for strategy, and the two are not in conflict. We run every multiple in this report on clean adjusted EBITDA that excludes the presale, and we separately carry Circle's retained stake as an asset in the fair-value bridge at $750M to $1.5B. What we will not do is both. The genuine open question is not accounting treatment but whether staking rewards on a permissioned chain produce a revenue stream large enough to matter against a $500M earnings base, and nothing disclosed this quarter bears on that.

Debate: Do the Binance and Hyperliquid Deals Buy Growth or Just Buy Volume?

Bull view: The Coinbase relationship cost Circle most of the reserve income on Coinbase-held balances for years and in exchange built the deepest distribution and liquidity in the category, which is now the hardest asset in the business to replicate. Extending the same structure to the largest platform in emerging markets, where 77% of 300 million users sit, is buying the next three years of float at a known price. The counterparty put $100M of its own equity in at a 16% premium to the current price, under a lock-up that unwinds if it walks away from the commercial deal, which is not how a counterparty behaves when it expects to extract and leave.

Bear view: Hyperliquid is the test case and it failed: 90% of the balances landed where Circle retains nothing, and the first full quarter cost 80 basis points of net reserve margin. Q4 adds a second such arrangement and the reaffirmed guide implies net reserve margin near 30%, meaning Circle would be paying away roughly 70 cents of every reserve dollar. If circulation does not grow in response, this is not an investment in distribution, it is a transfer of the entire economic surplus to the venues, and the volume that arrives is worth nothing.

Our take: The bear case is ahead on evidence and the test is unambiguous and ninety days away. The Coinbase precedent is real but it ran during a period when the float was compounding, which converted a high pass-through rate into a growing absolute dollar take; the same structure in a flat-float environment is pure margin transfer. The one fact that moves us toward the bull side is the equity structure, because tying the lock-up to the commercial arrangement makes the counterparty's $100M hostage to USDC's success on its platform. We need to see Q4 circulation growth alongside the margin step-down. The margin step-down without the growth is a downgrade.

Debate: Does GENIUS in January Restart Float Growth?

Bull view: January is the moment regulated institutions can hold and transact digital dollars with legal certainty, and the accounting standard-setters are close to classifying GENIUS-compliant stablecoins as cash or cash equivalents, which is the condition that lets corporate treasurers actually use them. Circle enters that moment with a federal trust charter, the first significant e-money token designation in Europe, 79% of stablecoin transaction volume, and ERP and treasury-system integrations already shipped. Circulation stabilized this quarter in a soft market; it reaccelerates when the buyer base expands from crypto to corporate.

Bear view: Legal clarity benefits every compliant issuer, including the bank consortium whose economics are explicitly built to pay distributors more than Circle can. GENIUS lowers the barrier for Circle's would-be competitors at least as much as it raises Circle's ceiling, and corporate treasury adoption runs on multi-quarter procurement cycles that will not show up in 2027 circulation data. Meanwhile USDC supply remains below its year-end level and management's only circulation guidance is an unfalsifiable through-cycle CAGR.

Our take: We are closer to the bull case on this one than we were in August, and the reason is specific: the accounting-standards leg is new information and it is the leg that converts stablecoins from a payments rail into a treasury instrument. Treasury balances are sticky where trading balances are not, which is exactly the problem Circle's velocity-without-balances profile has. The timing objection stands, though. Nothing in this debate shows up before the second half of 2027, and Circle has to get through a Q4 margin reset and a CFO transition first.

Model Update

Two changes matter and they point in opposite directions. The 2026 clean estimate goes up modestly, from $495M to $510M, because the first three quarters of ex-presale adjusted EBITDA ran at $151M, $143M and $135M against a path we had drifting lower faster, and because the operating expense line turned a quarter earlier than we expected. The 2027 estimate comes down materially, from a $500M midpoint to $425M, because the Q4 exit run rate on the new distribution economics is roughly $70M to $90M a quarter, or $280M to $360M annualized, and that is the base 2027 has to build from rather than the $135M quarterly rate Q3 printed.

Bridging from that exit rate to $425M requires three things we are explicitly crediting without disclosure: CPN revenue arriving at scale after the Tazapay close, Arc staking revenue beginning after the Proof-of-Stake transition, and the $60.4M of remaining presale recognition landing in the year. Against those we subtract a full twelve months of the Binance and Hyperliquid distribution drag and the operating expense of an acquired business. The $350M to $500M range is wide because management declined to narrow any of the three credits. A path where CPN monetizes at five basis points, staking is immaterial in its first year, and the float stays flat produces something near $340M. A path where circulation reaccelerates post-GENIUS to an $85B average, CPN and integrations reach $300M of other revenue, and net reserve margin recovers toward 34% produces something above $550M.

Valuation. At the November 5 close of $69.48 on 267.8M diluted shares, market capitalization is approximately $18.61B. Corporate liquidity is $1.70B of cash and cash equivalents plus $1.02B of cash segregated against Circle's own stablecoin holdings, which carries no corresponding holder liability, for $2.72B in total; a further $257M sits in investments and digital assets at fair value and is excluded. Deducting the $2.72B gives an enterprise value near $15.89B, and deducting Circle's retained 25% Arc stake at its $750M presale mark leaves roughly $15.14B attributable to the operating business. Against our clean FY2027E adjusted EBITDA of $425M, that is about 35.6x.

Fair value: $44 to $74 over 12 months. The low anchor applies 24x to the bottom of our clean FY2027E range ($350M), giving $8.40B of enterprise value, plus $2.72B of corporate liquidity and the Arc stake at its $750M presale mark, or roughly $44 per share on 267.8M diluted shares. The high anchor applies 31x to the top of that range ($500M), giving $15.50B, plus the same liquidity and the Arc stake marked at $1.5B on the basis that a live mainnet carrying roughly 500,000 transactions a day has roughly doubled the network's presale-implied value, or roughly $74. The $59 midpoint sits about 15% below the $69.48 close and the high end about 6% above it, so the stock again trades in the upper half of our range with no margin of safety. Bull case $95 to $110, roughly +37% to +58%, requires Q4 net reserve margin above 33% alongside circulation growth, a disclosed CPN take rate, Proof-of-Stake on schedule, and FY2027 other-revenue guidance above $350M. Bear case $35 to $45, roughly -50% to -35%, requires Q4 net reserve margin at or below 30% with no circulation growth and the ex-token other-revenue line holding near a $100M annual rate.

Thesis Scorecard Post-Earnings

Overall: the thesis bifurcated rather than weakened. Of the three August watch items that could have forced a downgrade, two resolved favourably. Arc shipped on the exact committed date and is carrying real volume, and the float stopped shrinking. Those are the two pillars whose failure would have made this an Underperform. The third resolved unfavourably and severely: the distribution economics management flagged for Q3 arrived, cost 80 basis points, and the reaffirmed guide says Q4 costs roughly another 780. The business that emerges is one with a stronger competitive position, a materially lower earnings base, and a growth case that depends entirely on two revenue lines nobody has quantified. Five of eleven pillars improved this quarter, three held, and three worsened, which is a genuinely two-sided quarter rather than a weak one.

Action: Maintain Hold. We are not upgrading because the clean earnings base is falling, the implied Q4 margin reset is severe, the stock trades at a higher multiple of 2027 earnings than it did in August despite falling 8% from the pre-print level, and the next call sets 2027 guidance with a departing CFO and no quantified CPN or Arc revenue to anchor it. We are not downgrading because the two pillars that would have justified Underperform both strengthened, the rate cycle has turned from headwind to tailwind, GENIUS goes live in eight weeks with cash-equivalent accounting treatment close behind, and Circle holds 25% of a live layer-one network carried at zero on its balance sheet. Fair value $44 to $74 brackets the $69.48 close with the stock in the upper half. Conviction unchanged at 5 of 10.

Explicit tests for Q4 2026. We will downgrade to Underperform on either of the following: Q4 net reserve margin at or below 30% with no sequential circulation growth, which would make the distribution-economics transfer structural rather than an investment; or FY2027 other-revenue guidance below roughly $250M excluding token recognition, which would break the diversification pillar outright. We will upgrade to Outperform on either of the following: Q4 net reserve margin above 33% alongside sequential circulation growth, which would demonstrate that the Binance and Hyperliquid margin is buying float that stays; or a disclosed CPN take rate with FY2027 other-revenue guidance above $350M. Secondary monitor items: a named CFO; the Proof-of-Stake date and the recognition schedule for the remaining $60.4M; the quantified MiCA capital and liquidity impact; Tazapay's share count at closing; whether adjusted operating expenses land at the $585M high end or below it; and any capital-return authorization.