Other People's Money

---

Begin with the trade that defined the stock, because everything that has happened since is a slow argument with it. On the fifth of June, 2025, Circle Internet Group sold shares to the public at thirty-one dollars apiece, a price its bankers had already raised twice and which valued the company at about $6.2 billion. The shares opened at sixty-nine. They closed that first day at $83.23 — up better than 168% — and they did not stop there. Eighteen days later, on the twenty-third of June, Circle touched $298.99, a roughly ten-fold gain on the IPO price in under three weeks, the most violent debut by a company raising five hundred million dollars or more in the entire history of American markets going back to 1980. For a few summer days, a firm that issues a dollar-pegged token was worth more than most of the banks whose dollars it merely holds.

Now set beside that the price today. Circle trades around $79, a fall of roughly three-quarters from its June peak, for a market capitalization in the neighborhood of $20 billion. The stock has, in other words, already told you the most important thing about itself: that it is not a stable instrument, that the crowd which bid it to three hundred dollars was pricing something other than the business, and that the business — when the euphoria drained out — settled at a level the fundamentals could just about be tortured into supporting, and not a dollar more. The question this essay asks is whether even that lower number is a verdict on the company or merely a less hysterical version of the same mistake.

What Circle actually is

Cut through the language and Circle does something a child could understand. It issues a token called USDC. Every time someone wants one, they hand Circle a real U.S. dollar; Circle keeps the dollar and gives back a digital claim on it, redeemable one-for-one, any time. There are roughly $79 billion of these tokens in circulation as of early 2026. Circle takes the $79 billion of real dollars sitting behind them and does the single most boring thing in finance: it parks the money in short-dated U.S. Treasury bills and overnight repurchase agreements — much of it inside a dedicated government money-market fund run by BlackRock — and it pockets the interest.

That is the whole engine. Circle is, functionally, a money-market fund. It gathers cash, buys Treasury bills, and earns the yield. The one twist — and it is the twist that makes the entire business model — is that, unlike an actual money-market fund, Circle keeps all of the interest and passes none of it to the people whose dollars it is. You give Circle a dollar; Circle earns four-and-a-bit cents a year on it; you get nothing. The holder of USDC has handed Circle an interest-free loan, and Circle has lent it back to the U.S. Treasury at interest and kept the spread. It is one of the oldest tricks in banking — earn on the float, pay nothing for the deposits — dressed in the vocabulary of blockchain.

This is not a criticism of the legality; it is the model. But naming it plainly matters, because the market is not valuing Circle like a money-market fund. Vanguard's flagship government money fund earns the same Treasury yield on far more assets and trades — to the extent its sponsor is valued at all — at a thin fraction of its float. Circle, doing the identical thing on $79 billion, is valued at $20 billion: roughly a quarter of the deposits it holds, capitalized as though the float were a software franchise with pricing power and network effects, rather than what it is — a pile of other people's dollars that will sit still only as long as nothing better comes along.

The number that should frighten a Circle shareholder

In the first quarter of 2026, Circle reported total revenue and reserve income of $694.1 million, up about 20% from a year earlier and ahead of Wall Street's $677 million estimate. Strong, on the face of it. But look at the composition, because the composition is the indictment. Of that $694 million, $652.5 million — about 94% — was "reserve income." Interest. The yield on the Treasury bills. For every hundred dollars Circle booked, ninety-four came not from selling a product, not from a network, not from software anyone chose to buy, but from the prevailing level of short-term interest rates as set by the Federal Open Market Committee. The same lopsidedness held across all of 2025: of $2.747 billion in total revenue and reserve income, $2.637 billion was reserve income and a mere $110 million was everything else the company does combined — all the payments rails, the developer platform, the partnerships, the "financial infrastructure of the internet" that the pitch deck foregrounds. Four cents of every dollar.

So the thing the market is calling a technology company derives nineteen-twentieths of its revenue from an input it has zero control over and that is, right now, near a multi-decade high. The Federal Reserve sets that input. When the Fed cuts rates, Circle's revenue falls — not at the margin, but close to one-for-one, because there is almost no other revenue to cushion it. A single quarter-point cut shaves tens of millions off the top line of a company whose entire valuation rests on the top line growing. This is the definition of a business with no pricing power: its single largest revenue driver is a number announced eight times a year by a committee in Washington that has never heard of it and never will. Circle is a leveraged long position on high interest rates wearing the costume of a growth stock — and high interest rates are, by the near-universal expectation of the very market that owns the stock, on their way down.

The partner who eats half the meal

Here is the part that turns an aggressive valuation into something closer to a trap. Circle does not even keep the interest it earns. It shares it — and the terms of the share are extraordinary.

Circle's distribution arrangement with Coinbase, the exchange that helped birth USDC, works like this: Coinbase keeps 100% of the interest income on any USDC sitting on Coinbase's own platform, and 50% of the residual reserve income on USDC held everywhere else in the world. Read that again. On the tokens parked at Coinbase, Circle — the issuer, the manager of the reserves, the entity whose name is on the regulatory filings — earns nothing. On the tokens held anywhere else, it splits the take down the middle with a company that neither issues the coin nor manages a dollar of its reserves.

The figures this produces are staggering. In 2024, Circle's total distribution costs were about $1.01 billion, of which $907.9 million went to Coinbase — meaning that of every dollar of reserve income Circle generated, roughly fifty-four cents flowed straight to its distribution partner. Circle does the work — holds the reserves, bears the regulatory burden, takes the headline risk — and Coinbase takes more than half the profit for the service of letting USDC exist on its app and in the broader ecosystem it seeded. It is one of the most lopsided revenue-share agreements in modern finance, and it sits, immovable, at the center of Circle's income statement.

Worse, from a shareholder's vantage, is that this cost is outside Circle's control. Circle's own disclosures concede that the payments to Coinbase are driven by Coinbase's business strategies and policies — which Circle "neither controls nor oversees." If Coinbase succeeds in driving more USDC onto its own platform, the share of coins on which Circle earns nothing rises, and Circle's economics get worse even as USDC grows. Growth in the wrong place is a cost, not a benefit. And the agreement comes up for its periodic review in 2026 — a negotiation in which Coinbase, holding the distribution and the relationship, has every incentive to press, and Circle, dependent on it, has little leverage to resist. A company whose largest revenue line is dictated by the Fed and whose largest cost line is dictated by Coinbase does not, in any meaningful sense, control its own profit-and-loss statement. It is a passenger in its own business.

A loss-making company at seven times sales

Put the two facts together — revenue at the mercy of rates, costs at the mercy of a partner — and the bottom line is what you would expect. For all of 2025, against $2.7 billion of revenue and reserve income, Circle reported a net loss attributable to common shareholders of about $70 million. It lost money. After taking in the entire interest yield on $75-billion-plus of float at the highest rates in a generation, after the IPO, after the fanfare — a loss. Compensation expense alone ran to roughly $845 million, swollen by the stock-based pay that public-company life and a soaring share price make so easy to hand out and so expensive to account for.

And for this — a money-market-fund-shaped business that loses money, surrenders half its yield to Coinbase, and depends on the Fed for 94% of its revenue — the market pays a forward price-to-sales multiple of roughly 7, against an industry average closer to 2.5, and a price-to-earnings ratio that cannot be calculated because there are no earnings to divide by. Seven times sales is a multiple the market reserves for software companies with 80% gross margins, recurring contracted revenue, and pricing power — businesses whose dollar of sales is durable and defensible. Circle's dollar of sales is neither: it is a dollar of interest income that halves if rates halve and that it shares away besides. To pay a software multiple for a money-market fund's revenue is to make a category error and call it a thesis.

The bulls have an answer, and it deserves a fair hearing. The answer is growth: USDC in circulation has been compounding fast — up better than 70% year-over-year at points in 2025, with management guiding to something like a 40% annual growth rate — and if the float keeps swelling, the argument goes, then even at lower rates the sheer scale of deposits will drive reserve income higher. There is something to this. A money-market fund that triples its assets earns more even as yields fall. But the argument quietly assumes three things at once: that USDC keeps growing at a torrid pace; that it grows in the places where Circle keeps the economics rather than on Coinbase where it keeps none; and that rates do not fall faster than the float rises. Each assumption is contestable. Stacked on top of one another and multiplied out to justify a seven-times-sales valuation, they require a near-flawless execution of a plan over which the company controls, as we have seen, almost none of the variables.

The law that hangs over the whole model

There is a final exposure, and it is the one the blockchain framing is least eager to discuss: the law. The United States passed a stablecoin statute — the GENIUS Act — that, among its provisions, prohibits stablecoin issuers from paying interest to holders. On its surface this looks like a gift to Circle: it forbids exactly the competition Circle most fears, a rival that would attract deposits by sharing the yield. Circle's whole model, remember, depends on holders accepting zero interest while Circle keeps it all; a law banning interest payments protects that arrangement.

But read it the other way and the same law is a flashing warning. It is a public, statutory acknowledgment that the natural, competitive, fair outcome — the one the market would produce if left alone — is for stablecoin yields to flow to the holders of the coins, not the issuers. The only reason Circle gets to keep the float's interest is that the law currently forbids anyone from competing it away by offering to share it. Regulatory protection is not a moat; it is a moat the government dug and the government can fill. Laws change. The political coalition that bars interest-bearing stablecoins today can be lobbied, litigated, or legislated into something else tomorrow, and the moment holders can earn yield on their digital dollars — through a competing structure, an offshore venue, a clever legal wrapper, or a future Congress — the entire premise of Circle's profitability evaporates. A business whose margins exist only because a statute forbids the obvious form of competition is a business renting its economics from the political process. And the rent can be raised, or the lease cancelled, at any session.

Even within the current legal frame, competition is arriving. The GENIUS Act was designed to invite regulated entrants, and banks, fintechs, and rivals are circling a float that throws off billions in interest. Circle remains the second-largest stablecoin behind Tether's USDT — roughly $140 billion to Circle's $79 billion — and while Circle leads in regulated, U.S.-facing, institutional venues, "second place in a market about to get more crowded" is not the profile of a company that should trade at seven times sales. It is the profile of a commodity issuer in a business where the product is, definitionally, a dollar — the most undifferentiated good imaginable. You cannot build a durable premium selling something whose entire value proposition is that it is exactly worth one dollar, no more, no less, identical to every competitor's.

The interlocking dependencies

Step back and look at the structure whole, because the danger is not any single exposure but the way they interlock. Circle's revenue depends on interest rates, which it does not set and which are expected to fall. Its largest cost depends on Coinbase, which it does not control and which renegotiates in 2026. Its profitability depends on a statute banning the natural form of competition, which can change. Its growth depends on a float that earns Circle the most precisely where Circle keeps the least. And its valuation depends on the market continuing to file it under "fintech disruptor" rather than "leveraged money-market fund" — a classification that the first sharp move in any of the other four variables could overturn in an afternoon.

These are not independent risks that might average out. They are correlated. A Fed easing cycle that cuts rates also tends to accompany the kind of risk-on environment in which capital chases yield — and a market full of yield-seekers is precisely the constituency most likely to push for, and most likely to defect to, an interest-bearing alternative the moment one becomes legal. The same macro turn that compresses Circle's reserve income strengthens the competitive and political case against its zero-interest model. The exposures do not hedge each other; they reinforce. Circle is most fragile in exactly the scenario — falling rates, abundant liquidity, hungry capital — that the market currently expects.

What the price already knows, and what it doesn't

The collapse from $299 to $79 means the market has already digested something. The pure mania — the notion that a stablecoin issuer could be worth what it was worth on the twenty-third of June, 2025 — has been purged. To that extent the stock is more honest now than it was. But honesty is relative, and a seven-times-sales valuation on a loss-making, rate-dependent, Coinbase-shackled, statute-protected money-market fund is not honesty; it is a quieter denial. The crowd has stopped believing Circle is a miracle. It has not yet started believing Circle is a money-market fund. The remaining distance between those two beliefs is the remaining distance in the stock.

None of this requires USDC to fail, or Circle to be a fraud, or stablecoins to be a bad idea. Stablecoins are a genuinely useful technology; USDC is a well-run, regulated, transparent instrument; the float is real and the Treasury bills behind it are real. The business is not fake. The valuation is the problem — the decision to capitalize a money-market fund's rate-dependent, profit-shared, loss-making revenue at a multiple built for software franchises. That decision rests on calling Circle something it is not, and the income statement, read honestly, calls it what it is on every line: ninety-four percent interest income, fifty-four cents of every dollar to a partner, a net loss at the cyclical peak of the only input that matters.

The deepest irony sits in the name of the asset itself. USDC is engineered to be the most stable thing in crypto — a dollar that is always worth a dollar, by design, by construction, by the full weight of its reserves. And the company built on top of that perfect stability is one of the least stable equities in the market: down three-quarters in a year, levered to a falling rate, hostage to a partner's strategy, protected only by a law that can change. The token never moves. The stock cannot stop. That divergence — between the thing it sells and the way it is priced — is the whole story, and it is the kind of story that does not resolve in the issuer's favor. The dollars in the reserve belong to other people. Sooner or later, so does the yield. And when it goes, there is nothing else on the income statement to take its place.