Circle built its payments network around independent payment providers: Circle would coordinate and settle, and they would keep the customers. On September 8 it announced an agreement to buy one of them.
Why would a company whose pitch is coordinating payment providers decide to own one?
A Strange Purchase
The deal is for Tazapay, a Singapore-based cross-border payments company, for $400 million in stock; it is expected to close in 2027 and needs regulatory approval, including from the Monetary Authority of Singapore. Nothing has closed, in other words, so what we have is a statement of direction, not an operating result.
The direction is what's interesting. Circle did not buy another token, blockchain, or settlement engine. It bought a company with more than 60 banking and fintech partners and local payout rails in more than 100 markets, concentrated in Asia-Pacific and emerging markets, that processes over $25 billion a year, about 60% of it already on stablecoins. And it bought a company it knows well: Tazapay has been a design partner for Circle Payments Network (CPN) since 2025, and Circle Ventures led its Series B extension in March.
USDC can move value globally. What it can't do is supply the banking relationships, the currency conversion, and the payout that put local money in a recipient's account. Paying $400 million for those routes suggests Circle has concluded that settlement technology is not the bottleneck; compliant local execution is, and it is much harder to reproduce than an on-chain transfer.
It also changes Circle's role. CPN connects independent payment service providers and helps them settle in USDC. Owning an operator at the edge of that network moves Circle from coordinating those providers to competing alongside them.
The Earlier Promise
To see why Circle would accept that trade, it helps to remember how it got here. Crypto’s first payment ambition was Bitcoin, and it ran into a simple problem: when people expect a scarce asset to appreciate, they hold it instead of spending it, and businesses need a unit they can quote prices in. Dollar stablecoins supplied that unit. USDT and USDC could expand with demand while staying anchored to the dollar, and they became central to crypto by connecting on-chain value to transactions priced in dollars, not by being scarce.
Issuing USDC was never enough to make it useful, though. Circle depended on Coinbase, other exchanges, DeFi, and traditional finance for liquidity and distribution, and CPN extended the same approach to payments: bring payment institutions into USDC settlement without owning the customer-facing routes. Neutrality was the point; it gave intermediaries a reason to join instead of treating Circle as a competitor. The closer stablecoins get to ordinary commerce, however, the less their success depends on anything crypto-native, and the more it depends on the institutions that can price, distribute, and legally deliver money.
Permission First
The same thing that makes stablecoins useful makes them politically sensitive. A dollar stablecoin gives users another way to reach foreign currency, which is most valuable exactly where foreign-exchange controls restrict the existing ways; once stablecoins are settling business payments, they are touching the channels governments use to manage their currency and capital flows. That a network works technically settles nothing about whether it is allowed.
Brazil: demand and resistance
Brazil is the clearest case. Demand is not in question: stablecoins account for around 90% of the country's crypto volume. Part of the appeal is fiscal. Buying dollars through a bank incurs the IOF, Brazil's tax on financial operations, which since 2025 has been 3.5% on most outbound transfers; moving dollars as stablecoins does not. Reuters reported in November 2025 that the Finance Ministry was weighing an extension of the tax to close what officials saw as a loophole, but so far the IOF has not been applied to stablecoins.
The central bank has moved faster than the tax authorities. Resolutions published that same month, and in force since February 2, 2026, classify stablecoin purchases, sales, and transfers as foreign-exchange operations, with licensing for service providers, reporting to the central bank, and a US$100,000 cap per transaction with unauthorized counterparties. A further resolution, effective October 1, 2026, bars regulated cross-border payment providers from settling the offshore leg of a payment in stablecoins or any other crypto asset.
Read those together. Brazil has not banned stablecoins; Brazilians can still hold and trade them. What the central bank has done is make sure that when a stablecoin is used as money for a cross-border payment, it either goes through institutions the central bank supervises or doesn't go through. Demand for digital dollars and permission to use them commercially are moving in opposite directions, which is why circulation is a poor measure of adoption.
Singapore: access is licensed
Singapore makes the same point from the permissive end. A regulated business needs more than a circulating token: it needs a permitted payment product, defensible accounting treatment, and a licensed route through which to use it. Singapore’s Payment Services Act brings digital payment token services inside the regulatory framework, but what a provider may do depends on its activities and its licence; settling on a global blockchain doesn't hand anyone local payment access. It is no accident that the one closing condition Circle named for the Tazapay deal is approval from MAS.
A compliant stablecoin network cannot simply become a second SWIFT. Circle is a private-sector issuer, not an intergovernmental alliance.
Governments can license, constrain, or refuse Circle and Tether even when their networks work perfectly; where a government does authorize broad stablecoin payments under rules that preserve its oversight, the ceiling rises. The national question, then, is who permits a stablecoin to enter ordinary commerce, and on what terms. Issuance and trading are only part of that test.
The National Test
Korea: trading is not adoption
Korea is the best illustration of why trading activity is the wrong test. It is one of the most active crypto markets in the world, and a business there still has no legal framework for stablecoin payments: the second phase of the Digital Asset Basic Act, which would cover stablecoin issuance and distribution, missed its December 2025 deadline and remains stalled.
What it is stalled on is telling. The dispute is over who may issue a won-denominated stablecoin: one proposal would require banks to hold a majority of any issuer, and the Bank of Korea has pushed for a say over approvals. The ruling party's April 2026 bill would treat stablecoins as a means of payment under the Foreign Exchange Transactions Act, which puts them under existing currency-control authority. Meanwhile the four major financial groups are each testing won-stablecoin payments with exchanges and fintech partners.
None of this is hostility to stablecoins. It looks like a country deciding that if stablecoins are going to become money, the money will be won, the issuers will answer to banks and the central bank, and dollar tokens will be treated as foreign exchange. The size of Korea's trading market tells you nothing about whether USDC can be used to pay a Korean supplier.
US–China rivalry is not one score
The US–China rivalry adds another layer, and it resists a single score. American institutions retain formal influence over finance, Chinese-linked founders and distribution channels play significant roles in crypto, and mainland restrictions sit alongside Hong Kong's intermediary position. Founder origin, infrastructure, policy jurisdiction, and customer market are four different sources of influence. The commercial test cuts through them: who controls lawful distribution and local rails, and can digital dollars reach businesses through them?
Taiwan: permission becomes concrete
Taiwan is another market where permission becomes concrete. The FSC’s published Virtual Asset Service Act identifies a July 22, 2026 promulgation date and warns that provisions are not yet in force. Promulgation must therefore be distinguished from operational authorization; implementation and approvals determine what a provider can actually offer.
The law puts domestic fiat-linked stablecoin issuance behind FSC authorization and Central Bank consent. VASPs may handle a stablecoin only when its issuance or trading has been approved.
That makes the adoption threshold concrete: stablecoins move beyond trading when companies can use licensed routes for settlement and accounting, and consumers can use regulated payment products.
A country can remain important to crypto production or trading while refusing dollar stablecoins a role in its monetary system. And even where permission exists, it doesn't mean USDC reaches customers, or that Circle captures their fees.
The Price of Neutrality
Permission is the first constraint. The second is commercial. CPN connected independent payment service providers and helped them settle in USDC, but those providers still controlled the merchant contract, the local payout route, the customer fee, and the choice of stablecoin.
That is what neutrality costs. USDT has the established liquidity, and conversion carries fees and thresholds; to get a provider to choose USDC in an existing corridor, Circle may have to subsidize it, and the provider keeps the customer relationship anyway. More settlement volume doesn't give the issuer a larger share of customer fees.
The distribution burden
Circle’s Q2 2026 Form 10-Q shows what that looks like in the income statement. For the three months ended June 30, Circle reported $667.7 million of reserve income against $412.5 million of distribution, transaction and other costs: 62 cents of every dollar.
The Coinbase arrangement shows where it goes. After specified expenses and Circle’s issuer portion, the parties receive platform-based allocations, and Coinbase receives 50% of the remaining broader-ecosystem amount after approved third-party payments; those payments are recorded as distribution costs. The circulation belongs to USDC. A large share of the economics belongs to distribution.
Payments are set up to repeat the pattern. If independent providers keep the paying customer and need incentives to use USDC, Circle has little control over the economics of the transaction, and moving downstream gets more attractive the longer customer revenue can't be shared on acceptable terms.
What ownership actually changes
Take a deliberately simple example:
- A merchant sends $100 through an independent licensed PSP and pays it a $2 customer fee.
- The PSP keeps that fee, while the stablecoin issuer pays the PSP a $1 incentive to use its token.
- The issuer has bought usage, but it has not gained customer revenue.
If the issuer acquires the PSP, the combined operator keeps the existing $2 fee, and the $1 incentive no longer leaves the group. Note what did and didn't happen: consolidation saved the $1 that was being paid out; the $2 customer fee already existed, so Circle would be buying it, not creating it. (These are illustrative numbers, before the acquisition price and operating costs.)
Change the fee arrangement and ownership becomes less necessary. If the independent PSP remits $1 of its fee to the issuer, that offsets the incentive, and a partnership has solved part of the economics without an acquisition.
So ownership is one answer, not the only one. Faster settlement is productivity; a license is access; ownership is bargaining power; and profit shows up only after the resulting revenue exceeds acquisition, compliance, and operating costs. A partner that shared enough customer revenue could be a better deal than a purchased operator. Neutrality gets expensive when partners keep the scarce assets (licenses, contracts, pricing power) while the network pays to support their usage.
(I make a similar distinction in my analysis of industrial AI and producer profit: a system can become more productive while the firms controlling customer access and coordination keep most of the gain.)
Buying the Last Mile
That is the context for what Circle says it is buying. From Circle’s Tazapay acquisition announcement, co-founder, CEO and chairman Jeremy Allaire:
“Combining USDC with Tazapay’s world-class banking relationships, local payout rails, and institutional customer base will accelerate worldwide USDC adoption.”
The division of labor is explicit: Circle supplies USDC; Tazapay brings banking relationships, payout rails, and institutional customers. Tazapay's co-founder and CEO Rahul Shinghal was more specific in Tazapay’s announcement of the Circle agreement:
“We built the licensed infrastructure that helps connect that network to local money—market by market, in local currency, on local rails.”
That is a description of local execution, built one market at a time, and it is what Circle is buying: the connection between global settlement and local money. If the deal closes and is integrated, Circle gets more direct access to customer contracts, currency conversion, and local payouts, and more influence over the fees and terms attached to them.
Does that end CPN's neutrality? Not automatically. What matters is whether Circle steers activity toward Tazapay, gives it better terms, or withholds equivalent access from competing PSPs. Keeping Tazapay operationally separate and using it to fill geographic gaps could preserve most of the neutrality that matters; the operating rules count for more than the ownership label. (Tazapay has told its customers that nothing changes for now: same team, same licences, existing contract terms.)
(The same supplier-customer tension runs through my comparison of AI platforms and the App Store economy; in both cases the thing to watch is whether the platform leaves independent businesses room to earn durable returns as it grows.)
Two payment stacks converge
Circle is not the only one moving. Mastercard completed its acquisition of BVNK, a crypto-payment infrastructure provider, while Circle is going the other way with a conventional cross-border PSP. Traditional payment networks are buying crypto infrastructure and the stablecoin issuer is buying conventional payment access; the two stacks are converging, and whether that improves anyone's returns depends on the price and cost of putting them together.
The operating test
Moving into local conversion and payouts brings Circle closer to traditional foreign exchange, where revenue depends on currency spreads and pricing power in payment flows, alongside the economics of issuing USDC. Profit arrives only if retained fees and stronger bargaining power exceed the purchase price, the compliance burden, integration costs, and competitive pressure. Five operating choices will show whether the thesis holds: how payments are routed, whether competing PSPs get equal access, how much of the fee Circle retains, what compliance costs, and what customers are charged.
Circle's agreement marks the limit of the neutral network model. Settlement moves value; licenses, local rails, and customer relationships decide who gets paid. The acquisition brings those assets inside Circle, and Brazil and Korea are a reminder of how much of the outcome is still up to someone else: the regulators who decide whether a dollar token is allowed to act as money in their market.
Get the next note
in your inbox.
Free research notes, selected readings and publication updates. No spam, unsubscribe any time.