On 1 July, Cloudflare opened the waitlist for its Monetization Gateway, a tool that lets any site, dataset, API or AI tool sitting behind its network charge for access, with the payment settling in stablecoins.

The mechanism matters more than the launch.

It runs on x402, an open protocol that revives the long-dormant "402 Payment Required" web standard: a machine asks for a resource, the server quotes a price, the machine pays and receives it, all inside a single request.

No checkout page, no account, no API key.

Settlement is sub-second, in stablecoins, with USDC and Open USD both named by Cloudflare as suitable coins. The company handles roughly a fifth of all internet traffic, so this is not a fringe experiment.

The reason it exists is a shift in who is actually using the web.

For thirty years the bargain was content in exchange for human attention, monetised through ads, subscriptions and checkout.

AI agents break that bargain.

An agent does not watch an ad or hold a monthly subscription. It reads a page or calls an API once, takes what it needs and leaves, and it does so far more often than any human visitor.

Per-seat and subscription pricing do not fit a machine. A per-request toll does.

Cloudflare is not alone, and that is the point. Amazon Web Services shipped the same pay-per-request model on CloudFront in June.

And Cloudflare is going a step further than being a neutral rail: alongside the Gateway it has named its own stablecoin, NET Dollar, built for exactly this machine-to-machine commerce.

Why this matters

This is stablecoins arriving as infrastructure, not as a trade.

When the company in front of a fifth of the web builds a stablecoin tollbooth, and builds its own coin to run through it, the demand is coming from the plumbing of the internet rather than from a trading desk.

It is also the clearest sign yet of a fourth kind of stablecoin volume, machine-to-machine payments, that no incumbent bank or card network currently touches. Watch who ends up owning that rail, because the agentic web is being wired for stablecoins now, while the question of which coin settles it is still open.

For the first time, the biggest dollar stablecoin and the most-used are not the same coin.

Visa's onchain analytics dashboard, powered by Allium, recorded a record $1.79 trillion in adjusted stablecoin transaction volume in June, narrowly beating February's prior high. Of that, USDC carried roughly two-thirds, about $1.21 trillion, against USDT's third at $576 billion.

Across the first half of 2026 the gap was wider still: USDC near 70%, USDT around 25%, on $8.82 trillion of volume, already more than the whole of 2024.

Yet by market capitalisation nothing has moved. USDT remains far larger, near $184 billion to USDC's $73 billion. And on raw transaction count USDT still leads comfortably, roughly 145 million transfers to USDC's 57 million.

Read those three measures together and the split is obvious.

USDT moves a great many small payments, the profile of remittances, savings and offshore dollar access.

USDC moves fewer, far larger transfers, the profile of banks, institutions and settlement.

One coin is the store of value and the everyday dollar. The other is the settlement rail. They are drifting into different jobs.

One caveat worth stating: the adjusted figure strips out bot loops and exchange rebalancing to approximate real economic use, and Visa has been a Circle partner since 2020.

The flip is real and has held every month since early 2025 with no rival dataset disputing it, but the number is an estimate, not a raw count.

Why this matters

"Biggest" and "most used" have come apart, and that is the whole game in miniature. The market is not resolving toward a single winner. It is separating into distinct jobs, each with a different leading coin, which is exactly the structure the main story maps in full.

For a treasurer or an allocator, the practical read is simple: stop asking which stablecoin is winning and start asking which one fits the job you actually need done.

On 2 July, Standard Chartered became the first globally systemic bank (G-SIB) to let institutional clients mint and redeem USDC directly through the bank, no separate Circle account required.

The service launched through its Dubai operations, aimed at settlement, treasury and liquidity, with Circle handling the coin on the back end.

Read plainly, a G-SIB has agreed to be the front door to someone else's dollar. The client onboards once, with the bank it already trusts, and the bank quietly routes the mint and redeem through Circle. It is a clean piece of engineering and a genuine milestone: the compliance and custody a treasurer expects, wrapped around a stablecoin the bank did not issue.

It is also the same data point from the volume story, seen from the bank's side. The reason USDC is pulling ahead on settlement volume is precisely this, institutions reaching it through regulated channels like Standard Chartered and, separately, BNY. The bank adoption and the volume flip are one story told twice.

But notice what the bank is not doing. It is not issuing its own coin. It is renting Circle's rail and lending its brand to it, which means the float on those balances sits with Circle, not with the bank.

Standard Chartered has taken the fast path in, the on-ramp, not the ownership.

Why this matters

This is the fork every bank now faces, made concrete.

The quick move is to plug into an existing compliant coin and offer clients access tomorrow, which is what Standard Chartered did. The slower, more valuable move is to issue the rail yourself and keep the economics, which a handful of banks are building toward instead.

Renting the rail gets you in the game. It does not get you the float. Which of those a bank chooses is the question the main story turns on, and it is the one worth putting to your own bank.

On 30 June, 140 companies lined up behind a single new stablecoin: Open USD (OUSD.)

The announcement sent Circle, the only pure-play stablecoin issuer on the NYSE, down 16%.

Visa, Mastercard and American Express in one row. Stripe and Adyen beside them. BlackRock, BNY and Standard Chartered. Google, Shopify and Samsung.

It was described everywhere as a coronation in waiting, the challenger that finally unseats the Tether / Circle duopoly.

That reading is tempting, and I think it is wrong.

Not because OUSD is unserious or because other consortium stablecoin projects (Diem, Global Dollar) have lost traction after early hype.

Because the coronation frame assumes a throne, one winner, one coin that ends up carrying the world's programmable money. Strip away the logos, chains and the governance language, and what those 140 companies are actually fighting over is narrower.

They are fighting over the float.

Every fiat-backed stablecoin sits on a pile of reserves, and those reserves earn the Treasury bill rate. That interest, paid on money the holder parked to make payments, is the entire economic prize.

It is why Tether earns more per employee than almost any bank on earth, and it is what OUSD's 140 partners have arrived to claim. The whole contest reduces to one question.

There are only three answers, and they are the three business models now competing for the market.

Keep it.

The issuer banks all the float and sells the coin to people who do not ask for a cut. This is Tether.

Split it.

The issuer shares the float with the distributor who brings the customers, trading margin for distribution. This is Circle/USDC.

Share it.

The issuer returns almost all the float to whoever holds the balances, and lives on a thin fee. This is what OUSD is really offering, and it is why the partner list looks the way it does.

Keep, split or share.

Three answers to who keeps the interest, three different customers, three different ways to play.

Line them up and the last-coin-standing story starts to fall apart, because they are not fighting for the same seat. They are fighting for different rooms and from different starting positions.

Let me take them in turn.

KEEP (Tether)

Tether keeps the float. Nearly all of it.

In that respect, it has a unique market position.

The reserves behind USDT earn the bill rate (plus gains on reserve assets a GENIUS coin could not hold, chiefly Bitcoin and gold) and that value accrues to Tether, not to the holder.

Why?

Dollar access in high-inflation economies, value transfer through corridors the regulated system does not reach. These users want the dollar, not the yield, and they are the most loyal customer base in the industry because for many of them the alternative is a collapsing local currency, not a rival coin.

You don’t care about foregoing a 4% T-Bill yield when you are trying to escape a 20 or 30% domestic inflation rate.

Tether’s margin is the highest in finance as a result.

Around $13 billion of profit in 2024 on roughly 150 staff, about $86 million per employee against Goldman Sachs at under $300,000. Keep everything, spend almost nothing, and the economics are without peer.

The growth tell is the ceiling on that model.

USDT shed supply in the first quarter of 2026 while USDC gained, and the regulated markets are closing to it: MiCA forced European platforms to delist USDT, and it is not authorised under GENIUS in the United States. The Keep model prints money in the arena it owns.

It simply cannot enter the arenas it does not.

But while dollar hunger and inflation persist in the emerging world, their model has room to run.

SPLIT (Circle/USDC)

Circle splits the float.

It shares the reserve income with the distributor that brings the customers, and the terms are startling. Its agreement with Coinbase hands Coinbase all of the reserve income on USDC held on Coinbase, and roughly half of the rest, a distribution cost worth around $900 million in 2024, more than half of Circle's revenue.

Why?

Circle's customer is the regulated institution, the onshore marketplace.

Payments firms, exchanges, banks and corporates that need a compliant coin with a named issuer and a clean reserve. A regulated venue does not choose USDC over USDT on preference, it uses the only one its rulebook will let through the door. Circle chose that customer years ago and paid for it in reserve composition and audits, which is why it inherited regulated Europe when Tether was delisted.

The margin is deliberately thinner.

This is the price of network effects.

Circle gives away the majority of its economics to buy reach, which is a strange-looking trade until you see what it bought. USDC now carries around two-thirds of adjusted transaction volume against Tether's third, a lead that has held month after month.

Circle sold the float in order to scale.

The growth tell is that the trade is working, but it caps the upside from above and below at once.

Above, the distributors capture the economics Circle gives up. Below, the compliance that won Circle the regulated arena is now the entry ticket every new entrant is buying too, including the banks.

SHARE (Open USD)

OUSD hands the float back.

Almost all of it goes to the partners who bring the balances, and the issuer lives on a thin management fee. This is not a coin with distributors bolted on: it is a distribution consortium that happens to mint a coin. This is why 140 companies with their own customer bases signed on at once rather than queuing to license someone else's dollar.

Why?

OUSD's customer is the balance holder itself.

Corporates, platforms and merchants who, for the first time, are offered the yield on the cash they were always going to park anyway. A treasurer running a billion dollars a day through the rail, with even a few hours between paying in and the supplier drawing out, is sitting on tens of millions of continuous float.

Today that interest goes to a bank or an issuer.

That is a genuinely new proposition, and it is aimed at the one pool Tether and Circle both leave on the table.

The margin, for the issuer, is barely there by design.

OUSD is not trying to earn the float. It is trying to move it, and to take a small toll on the movement.

The growth tell is the open question, and history is not kind.

Revenue-share consortiums have been tried and have stalled. Much like the immortals of Highlander, the ones that lost their heads early leave a crowded graveyard.

Diem, assembled behind Facebook with a partner list that looked unbeatable, produced nothing, though it launched into a far harsher regulatory climate than today's.

Global Dollar launched with more than a hundred partners and the same pitch, and remains a fraction of Tether’s size. A shared coin with 140 owners is hard to govern and easy to defect from, and almost every marquee partner on the list runs a competing strategy of its own.

OUSD is betting it can reach scale fast enough to prise open the regulated space Circle now leads. But by giving away almost all the reserve economics, it keeps little capital of its own.

The bet is scale.

The risks are consensus and defection.

THREE MODELS

Three models, three customers, three margins.

Tether prints money in a room it cannot leave. Circle bought its way into the regulated room and now watches the door fill with its rivals. OUSD is trying to build a room with a new value proposition and an underserved customer set.

Notice what is not happening: none of the three is beating the others at the same game. Each is playing a different one.

Which means the question the whole market is asking, who ends up the last coin standing, may be the wrong question.

The right one is potentially: how many rooms are there?

USAT

One company has already answered it. It is the largest issuer in the world, and it answered by building a second coin.

In January 2026, Tether launched USAT.

An onshore, GENIUS-compliant dollar coin, issued through a chartered US bank, Anchorage, with Cantor Fitzgerald holding the reserves.

It is everything USDT is not.

Because USDT stays exactly where it has always been. Offshore. Non-compliant. More than $184 billion in circulation, backed by reserves a US payment coin is not permitted to hold.

One company. Two coins. For two different rooms.

USAT is not an upgrade of USDT, and it is not a migration path off it. It is a ring fence. It is the compliant address Tether can show an American regulator, built precisely so that USDT never has to become compliant, and can stay offshore, keeping all of its float, indefinitely.

The rest of the market is playing Highlander, each issuer trying to be the last one standing on a single throne. Tether looked at the same board and refused the game. It did not bet everything on one coin winning every room. It built one coin for the regulated room and kept another for the room it already owns.

The market leader, the player with the most to lose from a fractured market, has quietly structured itself for a world of many rooms.

Be clear about what USAT is not: it is not winning.

USAT is a rounding error, under $200 million against USDT's $184 billion, roughly a thousandth of its own parent. It has grown fast off a tiny base, but its size is not the point.

Tether’s model cannot cross into the regulated arena, so Tether did not try to drag USDT over the line. It minted a separate immortal to hold ground the original never could, and left USDT exactly where it thrives : the world’s developing economies.

Which forces the question : if the market leader needs two coins because there are two rooms, then the rooms are real, and they are the thing worth mapping. Not who wins. Where the walls are.

TWO ROOMS

Start with what builds a wall.

Two things divide the stablecoin market into rooms, and neither of them is the coin.

The first is jurisdiction: whether a coin lives inside a regulated perimeter or outside it.

The second is the customer: whether they demand a share of the float, or do not.

Set those two axes against each other and the picture is not a single throne with challengers massing around it. It is two rooms, with a wall between them that almost nothing crosses.

The offshore room.

This is Tether's, and it owns it by not complying.

No rival coin dislodges Tether in this room, because there is nothing to out-compete on. You cannot undercut a price of zero by offering the customer something they did not come for.

The only force that touches this room is a sovereign. China treats USDT as illegal and drives it into grey over-the-counter channels, and even then cannot stamp it out. That is the tell of how strong the incumbent is here. It is not regulated into submission, it is suppressed, imperfectly, by states.

A competitor cannot take this room. Only a government can even dent it.

The onshore room.

This is the regulated perimeter, and it is the opposite kind of ground. You do not win it by being useful: you win it by being permitted.

Named issuer, clean reserve, licence, audit. And it is the only room that is genuinely contested right now: Circle the incumbent, USAT the entrant, OUSD the challenger, all fighting for the same institutional customer who, by law, cannot touch the coin that rules the room next door.

Now watch what happens inside a single room, because this is where the last-coin instinct is actually right.

Europe was the first corner of this room to be sorted.

MiCA set reserve terms Circle accepted and Tether refused, and the consequence rolled through the market as an eviction. Platform after platform delisted USDT for European customers, and USDC inherited the regulated continent almost by walkover.

In Highlander terms, this is the Quickening.

One coin falls in a room, and its strength, the balances, the flows, the liquidity all transfers to the survivor. Inside a room, network effects are brutal and they compound. Inside a room, there can very nearly be only one.

But here is the wall.

When USDT was evicted from Europe, its float did not cross over and crown a single global winner. It retreated to the room Tether owns, where it is still king. The Quickening runs to the edge of a room and stops. It does not cross the wall, because the customer on the other side cannot follow.

Which lets us finally answer the question this whole essay has been carrying.

Can there be only One?

Inside a room, almost. Across rooms, no.

The number of coins the market supports is not one. It is roughly the number of rooms, multiplied by the one or two that come to dominate each.

The duel was never the right image, because a duel needs one arena and a single prize. This is not a duel, it is a partition. The winners do not defeat each other. They divide the ground and stop meeting.

And the largest issuer in the world already told you so when it built a second coin rather than bet everything on one.

So the real work is not picking the survivor. It is mapping the rooms and knowing which one you stand in. We have found two.

There is a third, and it is the most valuable of all, because it is the one room every coin we have discussed is locked out of by design.

THE ROOM THAT PAYS

Every coin in this essay lives outside the banking perimeter, and every one of them gives something away to reach its room.

Which is the point. There is a room on the other side of that wall, and it is the largest of the three.

Think about what the outside coins are actually doing. They take a dollar of deposit, move it off the bank's balance sheet, and hand the float to an issuer, a distributor or a holder. Useful, but subtractive.

Now picture the same dollar tokenised and put on the same rails, moving just as fast, settling just as cleanly, but never leaving the bank's balance sheet at all. The deposit stays a deposit. The credit relationship it funds stays intact. The insurance stays. The float stays inside the perimeter, with the institution the customer already trusts.

That is the tokenised deposit, and it is the one instrument that keeps the float and the bundle together.

Keep, split and share are three ways to distribute float you cannot also lend against. This is the only model that keeps the float and lends against it too.

This is holy ground.

In the Highlander mythology, it is the one place the immortals cannot fight.

Here it is the perimeter itself. Tether cannot enter it without becoming something it has structured itself never to become. Circle can approach the door but not hold the credit relationship behind it. OUSD can offer the yield but not the loan, the overdraft, the insured deposit, the whole apparatus a business actually banks with.

The outside coins can price the float. Only a bank can keep the float and keep the customer.

And this is not the speculative room. It is already the biggest one.

Let’s set the scale out plainly.

Stablecoins, all of them, are around $312 billion in circulation, and their payment activity is still small against the river of global money. Tokenised bank deposits are already moving more than $4 trillion a year, and a single bank's network, JP Morgan's, reportedly clears more than $1 trillion of that on its own.

Read that and the hierarchy inverts. The coins we have spent this whole essay ranking are the top layer, the visible one, the one in the headlines. The layer with the real weight sits underneath, on the banks' own books, and it is already larger than the entire stablecoin market it is supposed to be threatened by.

Which reframes what the OUSD launch actually was. Not a coronation: a knock on a door.

When 140 companies assemble to offer treasurers the interest on their operating cash, they are not attacking Tether or Circle. They are pricing, in public, the float that banks have kept for nothing for fifty years.

The moment that number is quoted out loud, every treasurer can see what their idle balances are worth to someone else, and the operating-account conversation stops being about fees and starts being about rate.

So the question for a bank is not whether it survives the stablecoin era. On the numbers, the bank room is already winning.

The float subsidy is being repriced whether they act or not, and the instrument that answers it, the tokenised deposit, is theirs alone to build.

The majors that build it hold the most valuable room in the market on ground no coin can contest. The ones that wait will keep the room, but start paying rent on what used to be free.

There can be more than one coin. There is only one holy ground.

And it belongs to whoever builds on it first.

-Thanks for reading.

Mark McKendry, Stablecoin State

P.S. If you would like to contact our team just reply to this email - we read every response.

If this issue made you think differently about stablecoins, forward it to a CFO, treasurer, or finance leader who should be reading it.

The conversation is just getting started.