PayPal Stopped Selling Its Dollar and Started Selling the Factory

PYUSD has shrunk by about a third this year. PayPal's answer, delivered Wednesday, was to help other companies launch stablecoins instead.

The company launched PYUSDx on September 9 alongside M0 and MoonPay — a platform that lets a business issue a branded stablecoin backed one-for-one by PayPal USD. M0 supplies the programmable token plumbing. MoonPay handles issuance, onboarding and distribution. PayPal supplies the reserve asset and, more usefully, the name that makes a treasurer relax. Time from decision to live token: days, per the announcement, instead of the months it normally takes to assemble reserves, custodians, attestations and legal opinions.

Three early users — Saturn, Concrete and Cap — have pushed more than $100 million of combined volume through the system. A real number. Also a small one. That combination is the whole story.

The product stalled, so PayPal started selling the machine

PYUSD hit an all-time high near $4.2 billion in March. By mid-June it had fallen to somewhere around $2.7 billion, a contraction of roughly 31% in a single quarter. It sits third by market value behind Tether and Circle, holding something like 1.4% of a market worth about $300 billion. Third place sounds respectable until you see the concentration underneath it: USDT and USDC together account for roughly 82% of all stablecoin supply.

Three years of distribution through one of the largest consumer payment brands on earth bought PayPal a rounding error. Not for lack of trying, either — PYUSD launched with more regulatory groundwork and more retail reach than any dollar token before it, and it still could not pry meaningful share away from two incumbents whose main advantage is that they got there first and never broke.

So the strategy changed. If nobody wants a PayPal-branded dollar in particular, sell the ability to make a dollar to everyone who wants their own brand on one. The margin moves from float to infrastructure. The distribution problem becomes somebody else's.

A dollar that only works in one app is a gift card with better plumbing

There are already 213 distinct stablecoins in circulation. Almost none of them matter, and the reason is not technical. Money gets valuable through fungibility. Every additional branded dollar fragments liquidity, adds a redemption counterparty, and asks a user to hold something that works in fewer places than the thing it was minted from.

PYUSDx does soften that. Because every issued token is backed one-for-one by PYUSD and moves across chains on M0's rails, the branded dollars are meant to stay convertible back to a common base. That is a genuine improvement over the last generation of app-specific tokens, which were islands by design.

But convertibility is not the same as acceptance. A merchant who takes USDC has no obvious reason to also learn the redemption mechanics of a fintech's house token, and a user holding that token is one integration away from being stuck.

The float is the actual pitch, and nobody says it out loud

Here is the part that explains the demand. An issuer that puts $200 million of customer balances into a branded stablecoin earns the yield on the reserves behind it. At current short-term rates, that is real money for doing nothing except convincing users to hold your dollar instead of a bank's.

This is why the early adopters are payments and credit companies rather than consumer brands. They are not buying a marketing asset. They are buying a balance sheet.

Which raises the question the launch materials skip: whose balance sheet? PayPal holds the reserves behind PYUSD. The branded issuer holds a claim on PYUSD. The end user holds a claim on the branded issuer. Each layer is probably fine. Three layers of probably-fine is how 2022 happened.

Issuing a token in days means answering the hard questions in days too

There is an operational reality that gets lost in the speed pitch. A company that spins up a token in a week now owns a smart contract with mint authority, freeze authority, and usually an upgrade path. Who holds those keys. What can be changed after launch. Whether the treasury allocation is actually locked or merely described as locked.

Token teams have been answering those questions for years, and the tooling for it — Team Finance and similar contract-level lock and vesting infrastructure — exists precisely because "trust the issuer" stopped being an acceptable answer around the third exit scam. Fintechs arriving fresh to token issuance inherit that history whether they read it or not. The compliance file is not the audit trail. The chain is.

Washington has not decided whether these are stablecoins

The legal status of a PYUSD-backed brand token is not settled, and pretending otherwise would be generous. Treasury Secretary Scott Bessent spent this week pressing the Senate to move the Clarity Act, the market-structure bill that has been stuck for months. Until something passes, the question of whether a wrapped, rebranded claim on a licensed payment stablecoin is itself a payment stablecoin — requiring its own issuer license, its own reserve rules, its own attestations — has no clean answer.

Issuers will argue they are distributors, not issuers, because the reserve sits with PayPal. Regulators may look at whose logo is on the token a consumer holds and whose customer service line rings when redemption fails. Both readings are defensible. Only one of them is cheap.

The redemption test is the only test

Everything about PYUSDx works until it doesn't, and the failure mode is easy to picture. A branded token depegs on a weekend. Users try to redeem. The brand's app says the token is fully backed by PYUSD, which is true. PayPal says PYUSD is fully backed by Treasuries and cash, which is also true. Nobody is lying and nobody can process the withdrawal until Monday.

Stablecoins have never actually been tested on the layer-of-abstraction problem. Tether and Circle failures would be issuer failures — bad reserves, frozen banks, the things people already model. A PYUSDx failure would be a plumbing failure between two solvent parties, which is a category the market has no pricing memory for.

PayPal has made a reasonable bet: it lost the branded-dollar race and pivoted to being the thing branded dollars are made of. The same drift is visible everywhere in the sector — 21 institutions including Bank of America, Citi, Goldman Sachs and Fidelity have committed to a jointly owned dollar token rather than each launching one alone. The money in stablecoins is moving away from the tokens and toward the rails underneath them.

The uncomfortable version of that thesis is that it works best if hundreds of branded dollars launch, most of them are badly run, and PayPal collects infrastructure revenue either way. Watch what the second cohort of issuers looks like. If it is fintechs with treasury desks, this is boring financial engineering. If it is consumer apps chasing float income they don't have the operations to support, somebody is going to find out what a three-layer dollar does under stress — and they will find out on a Saturday.