Circle has agreed to pay $400M for a business whose revenue it can keep.

What's happening: On September 8, Circle agreed to buy Tazapay, a Singapore company that moves money across borders for businesses, for $400M in Circle stock. The deal is expected to close in 2027. Investors were unimpressed. CRCL fell 5.75% to $96.18 that day.

Why it matters: The usual explanation is that Circle bought the last mile of payments. That describes what Tazapay does. It doesn't explain why Circle wants it. Circle retains about 39 cents of each dollar its reserves earn after distribution and related costs. Tazapay's fees aren't subject to that same split.

This is a sensible use of $400M, even if it is too small to change Circle's earnings much this year. Both things can be true. The stock's reaction tells us little about which investors were thinking about.

Let’s unpack.

Start with the terms:

These terms give Circle some protection if the business changes before closing. They don't eliminate the risk. If Circle's stock falls, it must issue more shares to pay the same $400M. A fixed dollar price protects the purchase amount, not shareholders from dilution.

What Circle is buying:

By the numbers: The price is 20-32x FY2025 revenue, depending on which estimate you use, and 1.6% of annualized payment volume. It is also about 3.3x the $122M post-money valuation PitchBook records for the round Circle helped price six months ago. Tazapay raised about $60M: a $16.9M Series A led by Sequoia Southeast Asia, now Peak XV, in February 2023, followed by a $36M Series B in two parts. Peak XV led the first in August 2025, with Ripple and Circle Ventures participating. Circle Ventures led the March 2026 extension, with Coinbase Ventures and CMT Digital. Because the $400M excludes Circle's existing stake, the implied value of the whole company is somewhat higher.

Analysts disagree about what matters. Mizuho kept its Underperform rating and $45 target, arguing that Tazapay's payment volume is too small relative to Circle's to change much. (Investing.com) TD Cowen emphasized the opportunity to integrate USDC more closely as adoption grows. (Banking Dive)

Jeremy Allaire: "Combining USDC with Tazapay's world-class banking relationships, local payout rails, and institutional customer base will accelerate worldwide USDC adoption." Rahul Shinghal: "Circle has the dollar infrastructure in USDC and the regulatory standing to take what we've built further than we could alone."

Those are reasonable descriptions of what the companies can do together. The more interesting question is what Circle gets paid for doing it. More USDC moving through a system doesn't necessarily mean much more money for Circle. A fee on each payment might.

The 39-cent problem. USDC's reserves earned $668M in Q2. Circle paid $412M in distribution, transaction, and other costs. Most went to partners that hold USDC on their platforms. Circle's net reserve margin was about 39%.

Coinbase gets the largest share. The arrangement summarized in this breakdown of Circle's S-1 gives it 100% of reserve income on USDC held on its platform and 50% of the residual. Circle issues the coin. The company that brings and keeps the customer captures much of the income.

Tazapay charges for collections, payouts, and foreign exchange. Those fees don't come with the same obligation to share revenue with USDC distributors. Before Tazapay's own costs, a dollar of fee revenue is equivalent to about $2.56 of gross reserve income at Circle's 39% retention rate. That is a comparison of retained revenue, not profit. Tazapay still has to pay to move the money.

This helps explain why Circle might pay a high revenue multiple. There is also a reasonable argument that the historical multiple overstates the price. Our estimate starts with $12.4M of FY2025 revenue and roughly $10B in annualized volume, implying a fee of about 12 basis points, or 0.12%. Apply that to $25B and you get roughly $30M in annual revenue, putting the price near 13.3x. At 10 basis points, it is 16x. The periods don't align perfectly, and the fee rate may have changed. This is a 51 Insights estimate, not a disclosed run-rate.

Yes, but: Mizuho is right that $25B is small beside the $14.8T of USDC that moved on-chain in Q2. The trouble is that these figures measure different things. Circle doesn't earn reserve income each time a coin moves. It earns it while the coin exists.

Suppose all $25B of annual payment volume settled in USDC and remained there for a day. That would imply an average balance of about $68M. At a 3.5% reserve return, it would produce roughly $2.4M of gross reserve income a year, before distribution costs. Even this generous assumption adds little beside $668M in a single quarter. Let the money sit for three days and the conclusion is the same. In practice, only part of Tazapay's volume settles in stablecoins, and some of that is USDT.

The case for the acquisition therefore rests on fees. On our assumptions, Circle is paying roughly 13-16x estimated annual fee revenue for a growing payments business. That can be a reasonable price. It isn't made reasonable merely by comparing payment volume with on-chain transfers, or by treating revenue as profit.

One rate cut costs more than Tazapay earns. Circle's Q2 reserve income grew 5% while average USDC circulation grew 25%. The gap came from a 66-basis-point fall in the reserve return, to 3.5%. At $76.5B of average circulation, a 25-basis-point decline in yield removes about $191M of gross annual reserve income. At a 39% net reserve margin, that's roughly $75M after distribution, before allowing for the time it takes Treasury yields to adjust.

Our estimate of Tazapay's entire annual revenue is $30M. It would take about two and a half businesses that size just to match the lost retained reserve revenue, and more to replace the profit once their operating costs are included.

Tazapay is a first step toward diversification. Circle still needs other sources of revenue. It says it will begin charging CPN fees in the second half of 2026. On the Q2 call, CFO Jeremy Fox-Geen said: "Now it is becoming the time for us to start to monetize that." Arc's token presale has already helped lift FY2026 other-revenue guidance from $150-170M to $310-330M.

Here is a useful scenario, not a forecast: charge 10-12 basis points on CPN's $23B of annualized volume as of July 31, and you get $23-28M a year. That assumes CPN can charge something like Tazapay's estimated fee despite providing a different service. It also shouldn't simply be added to Tazapay's revenue without checking for overlap. Either way, the scale is still small beside roughly $2.7B of reserve income.

The acquisition matters more as a clue to Circle's next moves. If it keeps buying fee businesses with stock, we will have evidence of a deliberate effort to change the revenue mix. Another issuer or another chain would suggest different priorities.

The big picture: Until now, the large stablecoin payments deals mostly followed the same pattern. A company with customers bought the infrastructure to serve them:

The four large deals add up to roughly $3.7B in headline consideration. The BVNK and Reap totals include contingent payments. Circle is approaching the same problem from the other end. It already issues the dollar. It wants a more direct relationship with the businesses using it.

There is another reason to move now. On September 1, 21 banks, including Bank of America, Citi, Goldman Sachs, Wells Fargo, UBS, Santander, Deutsche Bank, and MUFG, committed to a joint dollar stablecoin. They plan to form the company in the second half of 2026 and launch in the first half of 2027, with cross-border payments and digital asset settlement among the intended uses. Tazapay is expected to close in 2027 too.

The banks bring balance sheets and customers of their own. They still have to connect their new dollar to local payment systems. Tazapay gives Circle more of those connections before the banks arrive. If a business already collects and pays through Circle's subsidiary, Circle has another way to keep the relationship even as the business gains more choices of coin.

Circle's headline purchase price is low relative to Tazapay's reported payment volume. That is partly because collections and payouts earn small fees, and partly because the buyers in these deals were purchasing different capabilities. Price per dollar of volume is useful context; it isn't a valuation model. Sources: company announcements and SEC filings; Architect Partners on Bridge volume, August 2024; Ripple on Rail's share of the $36B B2B market, 2025; Mastercard/BVNK closing reports, August 2026; Circle's 8-K; 51 Insights analysis.

The catch: Tazapay also settles in USDT. Neither company has explained what happens to that after closing. About 60% of Tazapay's volume settles in stablecoins, and it supports both USDT and USDC. USDT has a particularly strong position in Asia. Its roughly $185B circulation compares with USDC's $73.3B at June 30, when Circle put its market share at about 27%.

Circle will therefore own a company that helps customers use its main competitor's coin. It can require USDC and risk losing customers. It can keep supporting USDT and earn fees from moving a rival's dollar. Or it can accept USDT and convert it to USDC for settlement where customers permit. That last option may support USDC balances, though the reserve-income benefit depends on how long those balances remain.

Mastercard's Michael Miebach can comfortably expect "a world of multiplicity, many coins, many chains." Mastercard doesn't need its own coin to win. Circle has to decide how much that distinction matters.

Our prediction: Circle keeps USDT for at least a year and converts at settlement where customers allow it. Customers choose payment systems partly because their counterparties already use them. Force a different coin on them and some will choose a different provider. The first customer terms after closing will tell us how Circle resolves that tension.

The USDT decision seems the most immediate strategic risk because it could change the business Circle is buying. Partner departures come next. The bank consortium is a longer-term competitive question. The terms and regulatory permissions still matter, and so does price if our revenue estimate is too generous. A business can fit the strategy and still disappoint the buyer.

Circle's core income depends on interest rates, and it shares much of that income with distributors. Payment fees offer a way to change both facts. They aren't mechanically tied to Treasury yields or covered by the same revenue-sharing arrangement, though they remain exposed to volume, competition, and operating costs.

Meanwhile, distributors have committed roughly $3.7B to the large infrastructure deals above, and banks are planning a dollar of their own. Circle has a widely used stablecoin. Tazapay gives it more ways to earn money from the businesses using that coin, and potentially from businesses using other coins too.

Here is how the five groups compare across the same layers: their own dollar stablecoin, payment orchestration, local payouts in more than 100 markets, collections and card acquiring, and a U.S. federal charter. The categories are useful, but the capabilities within them differ.

Circle has been assembling its position in stages. It bought Hashnote in January 2025 for about $100M, adding the USYC collateral business. (CoinDesk) CPN launched in May with Tazapay as a design partner. The June NYSE listing at $31 gave Circle publicly traded shares it could use for acquisitions. Circle Ventures joined Tazapay's Series B in August, then led the March 2026 extension at a reported $122M post-money valuation.

July brought full OCC approval of the trust charter and the acquisition of IBM's blockchain patent portfolio, roughly 1,000 issued patents. The Tazapay agreement followed on September 8. Arc's public mainnet is scheduled for September 16, with 11 founding validators including BlackRock, DTCC, Mastercard, and Visa.

If these plans come together, by early 2027 Circle could issue the dollar, run a blockchain, operate an institutional payment network, hold a federal charter, and collect and pay out in local currencies across more than 100 markets. Stripe, Mastercard, and Ripple are assembling overlapping capabilities from different starting points.

Owning the parts doesn't settle the economic question. Circle still shares much of the income from its core product. Tazapay changes that for a small part of the business. CPN's recurring fees and Arc's economics will have to carry more of the burden. A more complete set of products is only useful if the combined business can earn an attractive return.

What to watch, by seat:

That’s all for now, folks.

– Marc & Team