While the consortium era grabbed the year’s headlines, Circle Internet Group spent it compounding. The Circle Payments Network ended the second quarter at $14.7 billion in annualized transaction volume, measured over the trailing 30 days, up 76% on the prior quarter, according to the company’s results. Enrolled financial institutions reached 175, up 29%.
In April the company shipped Managed Payments, a version of the network in which clients never touch a digital asset. On Sept. 16, Circle launches Arc, a blockchain of its own, with a whitepaper for a native ARC token.
Each piece is a product announcement. Together they describe a company rebuilding itself into something the payments industry has seen before, a network, and the flows it is absorbing come out of correspondent banking’s ledger.
What CPN replaces
The incumbent deserves a sketch. A cross-border institutional payment today traverses correspondent chains: messages over SWIFT, settlement posting bank by bank, hours to days later, inside banking windows, with every hop adding fees, FX margin and reconciliation work. The system works, which is why it has lasted, and its costs are why treasurers pre-fund accounts on three continents and Friday-afternoon payments arrive Tuesday.
CPN collapses the chain to one hop. The sender’s institution converts to USDC, value moves on-chain in minutes at any hour, the receiver’s institution converts out, and Circle’s network layer handles counterparty standards, screening and orchestration. Correspondent banking wins on breadth; it reaches everywhere, and no token network will match that coverage for years. CPN wins on physics, and networks like this grow by making the physics gap matter more than the breadth gap, corridor by corridor, starting with the lanes where the incumbent service was worst.
What the curve says
The 2026 numbers answer the launch-year chorus of impressive pipeline, unproven flow. From $8.3 billion annualized at the end of March to $14.7 billion at the end of June is the shape of a network past its cold-start problem, according to the company’s quarterly disclosures. Members begin transacting with each other, each addition creates pairs rather than a node, and 76% volume growth against 29% membership growth is that signature.
The metric deserves one caveat: annualizing a trailing 30-day window flatters momentum. It also updates honestly, since a stall shows up within a quarter, which is presumably why Circle keeps publishing it.
The absolute number is small against the trillions in correspondent channels. The composition points where every serious stablecoin payments story points: emerging-market corridors, Latin American, African, Gulf and Southeast Asian institutions, remittance and business-to-business lanes where settlement in minutes against dollar liquidity is the product. CPN is not displacing Fedwire. It is absorbing flows whose incumbent alternative was pre-funding, patience or the informal market.
Circle’s own tell for the composition shift is the client names it now says out loud. “The institutions using USDC today, like BlackRock, BNY, and Standard Chartered aren’t piloting, they are expanding,” said Jeremy Allaire, Circle’s co-founder and chief executive, in the results announcement.
Managed Payments hides the token
April’s Managed Payments launch may matter more than the curve. It lets a payment company, fintech or bank route flows through CPN while Circle runs the entire digital-asset lifecycle: minting and burning USDC, orchestration, compliance controls, chain infrastructure. The client sees fiat in and fiat out, faster and cheaper. The token works below the waterline.
Three years of institutional pilots taught every issuer the same lesson, and this product is Circle acting on it: the customers do not want to hold crypto, they want what crypto does. Clearing banks wrapped token legs inside existing platforms, card networks buried tokens in settlement and Circle built the fully managed version.
The defensive value is real too. The firms most likely to route around USDC are processors and platforms building their own token capability, and a managed service converts their build decision into a buy decision at a lower operational price than building, which keeps them inside the tent while the consortium era tries to pull them out.
The membership number to watch is not the total. The mix so far skews to payment companies and cross-border specialists, and the milestone that would change CPN’s category is the arrival of household-name commercial banks, the constituency Managed Payments was built to onboard. Watch the composition of the next few disclosures more than the totals.
Arc completes the vertical
Then comes Arc. Circle’s chain arrives Sept. 16 carrying privacy features for institutional flows, which is the workload banks name first when asked why they avoid public chains, an agent stack for machine-initiated payments, support for tokenized real-world assets and a native ARC token contemplated for governance and network operations, according to the company’s roadmap.
The strategy reads straight off the stack. Circle already issues the asset, operates the settlement network and sells the managed platform. A proprietary chain completes the set, with fees, finality and compliance surfaces under Circle’s control rather than rented from Ethereum’s or Solana’s roadmaps. Vertical integration is the oldest play in payments, and Circle is running it in public, roughly one layer per year.
The risk is just as legible. USDC’s distribution has always depended on neutrality; everyone could build on it without enriching a competitor. A Circle chain with a Circle token tilts that, in the same season the card networks stood up Open USD on the explicit argument that no supplier should own this much of the stack. Arc will be judged on who deploys real workloads there in the first 90 days.
Circle’s answer to the neutrality question so far is a count: more than 100 teams building in the Arc ecosystem ahead of launch, according to its results. Counts are not commitments, and the names that run production flows in the first quarter will settle the reading.
Read more: New payment chains court banks with rails built for dollars • Visa opens a stablecoin mint for its bank clients
The token, the delicate part
The ARC whitepaper is the document Circle’s lawyers will have sweated. A native asset for governance and network operations sits in territory American regulators spent a decade litigating, and the industry’s most compliance-branded company cannot afford ambiguity next to its licensed core.
Expect the design to lean toward utility mechanics, validator coordination, fee abstraction, resource pricing, and away from anything resembling profit participation, with distribution favoring institutions over a public sale. September’s documents will be parsed clause by clause, and not only by developers.
The licenses underneath
None of the expansion happens without the regulatory spine. Circle holds the industry’s broadest permission set: the American trust-charter track alongside state licenses, MiCA authorization carrying USDC across the European Union, and named-benchmark status in the Asian recognition regimes now being drafted. Singapore’s consultation and Hong Kong’s licensing wave both effectively pre-clear tokens of exactly Circle’s profile.
That portfolio is why Managed Payments can exist. A bank routing flows through CPN is transacting with a comprehensively supervised counterparty, not experimenting with crypto. It is also why the consortium threat has a near-term ceiling: every quarter Open USD spends acquiring what Circle already holds is a quarter of CPN compounding uncontested.
What a corridor conversion looks like
The unit of CPN growth is a corridor going live, so walk one. A payout provider serving Gulf-to-South-Asia flows enrolls both of its regional banking partners in the network. Compliance standards are the slow part: each institution clears Circle’s counterparty requirements, wires its screening into the network’s and maps its local reporting onto the flow. Then treasury flips a routing rule.
Transfers that pre-funded a correspondent account on Sunday now convert to USDC at the sending edge and out at the receiving one, whatever the day. Each conversion is months of compliance work followed by a step change in volume, which is why the network compounds in lurches rather than lines, and why membership composition predicts volume two quarters out better than volume predicts itself.
Public-company discipline as a weapon
The disclosure cadence is a competitive asset of its own. Circle has reported as a public company since its 2025 listing: quarterly results with named metrics, dated product launches, a roadmap that commits to timelines. That discipline matters in a category whose largest player discloses through attestations on its own calendar.
Institutions choosing a settlement network are also choosing whose numbers they will have to defend to their own boards, and audited quarterly reporting travels better in that meeting than any whitepaper.
The partner file deserves its own line in any fair account. Coinbase’s distribution arrangement takes a share of USDC economics and shapes where the float lives, and the exchange shelf decisions that once built the token’s liquidity now sit partly with firms holding seats in Open Standard. Circle’s response has been to make the institutions its own distribution: every CPN enrollment is a relationship the consortium cannot vote away, which is one more reason the membership number carries more strategy than the float does.
The failure modes
A fair account lists them. Concentration, because early network volume tends to run through a few heavy corridors and counterparties, and the curve flattens abruptly if one consolidates away. Margin, because reserve income on the roughly $75 billion USDC float still funds everything, and it is a rates bet Circle does not control. Partners, because Coinbase’s distribution arrangement and the exchanges’ shelf decisions still shape the float, and several of those partners now hold seats in a rival consortium.
Allaire named the margin exposure himself in the results, saying the quarter reflects “the current rate environment and a crypto market that has slowed,” conditions he called outside the network. The network strategy is the answer to that sentence as much as to any competitor.
Arc adds a fourth: a token-bearing chain invites questions a licensed issuer has so far avoided, and a neutrality stumble would hand Open Standard its best recruiting argument.
The board around it
Open USD courts Circle’s settlement franchise with zero-fee mint and redeem and the card networks’ distribution. Tempo, the Stripe-incubated payments chain live since March, courts the same institutional flows at the rail layer with a design-partner list that reads like Circle’s target market. The 21-bank consortium promises a bank-owned token for 2027, and Tether’s offshore franchise sits untouched by all of it.
Circle’s visible answer is speed and stack depth: grow CPN’s pair count before consortium tokens exist, convert institutions to Managed Payments before they consider building, ship a chain before the industry standardizes on someone else’s. It is the strategy of a firm that knows its moat is integration rather than the token, because after the GENIUS Act the token is a commodity anyone licensed can mint.
Against Tempo, the structural difference is ownership. Tempo is a processor’s chain, tokenless at launch and financed as shared infrastructure that makes its stakeholders’ other businesses cheaper. Arc is an issuer’s chain, completing a vertical rather than pooling one. The interesting scenario is not one beating the other but cross-listing, USDC native on Tempo or consortium assets on Arc, which would signal the grounds intend to interoperate rather than balkanize.
A note on measurement
Circle’s disclosures model something the category still does badly. The company reports network volume from observable windows, membership with growth rates and products tied to named capabilities, while its float has become almost a lagging indicator of the strategy.
Float measures money at rest. Circle is methodically converting itself into a company paid for money in motion, and if the flywheel holds, the industry’s scoreboard will eventually follow it, from supply tables toward throughput, enrolled institutions and corridor pricing. The supply chart that once defined the company is becoming a residue of the strategy rather than the strategy itself, which is exactly how processor economics have always been read.
Circle spent the year being written about as the incumbent under siege, and the quarter’s numbers describe a firm converting siege into flywheel, one enrolled institution at a time. Whether the flywheel outruns the consortiums is the defining commercial question of the next four quarters, and unusually for this industry, both sides have agreed to fight it in public, with dated launches and quarterly numbers, where everyone can keep score.