The headline arrived as one sentence: MoneyGram has launched a stablecoin-backed Visa card in Colombia.
That is the complete factual payload. One company. One country. One product. No stablecoin named. No blockchain named. No issuer processor disclosed. No licensing structure. No transaction volume. No cardholder count. No settlement latency. No audit report.
Attached to that single fact were three claims, presented not as hypotheses but as conclusions. The card could "accelerate global stablecoin adoption." It may "reshape cross-border transactions." It "advances financial inclusion."
I have watched this industry for twenty-seven years and audited it for nine. In 2017, working out of Tokyo, I ran a fifty-point security checklist against more than forty ICO smart contracts and rejected fifteen of them for failing basic code hygiene. The habit calcified. When I read an announcement now, I count facts and opinions separately and I never let them share a column.
The count here is one to three.
Chaos demands structure before it yields value. The structure of this announcement is a press release wearing an infrastructure costume. That does not make the product wrong. It makes the announcement useless to anyone trying to determine whether the product matters.
So I will do the work the announcement refused to do. I will take the single fact seriously, reconstruct the architecture it almost certainly implies, and tag every inference as an inference. Trust is built through transparency, not promises.
The Corridor, The Incumbent, The Card Network
MoneyGram is not a crypto company. It is a remittance company that has spent a decade trying to decide what to do about crypto.
Its distribution is physical. Hundreds of thousands of agent locations globally — pharmacy counters, check cashers, convenience stores, bank branches — form one of the densest cash-in and cash-out networks in the world. That network predates the smartphone and it still works when the smartphone does not.
Its ownership changed in 2023. Madison Dearborn Partners took the company private, ending its run on the Nasdaq. That matters for how you read this announcement. A private equity owner underwriting a five-to-seven-year hold is optimizing for operating cash flow, not for ecosystem mindshare. When a PE-backed payments company announces a crypto product, the question is never what it does for the industry. The question is what it does for the margin.
MoneyGram's crypto history is instructive and largely abandoned. It partnered with Ripple in 2019 for cross-border settlement, and that relationship ended as Ripple's regulatory exposure mounted. It later built toward Stellar, including a non-custodial wallet and USDC-denominated settlement capability announced around 2022. MoneyGram has also run fiat on-ramp and off-ramp products that let users convert between cash and digital assets.
Confidence: medium. These are documented industry developments, but none of them are confirmed by the announcement under discussion, and MoneyGram has not published which remain live.
Visa's side of the ledger is equally documented. Visa has run USDC settlement pilots with issuers, settling in stablecoins on public blockchains instead of correspondent banking wires. That program exists for one structural reason: if merchant settlement can happen on-chain, Visa wants that settlement to happen through Visa.
Colombia is the third variable and probably the decisive one. Colombia is one of the largest remittance recipients in Latin America, with inbound flows in the low tens of billions of dollars annually, overwhelmingly from the United States. It also ranks consistently high on global crypto adoption indices — among the highest in the Western Hemisphere — and its financial regulator has run a controlled sandbox for crypto asset services since the early 2020s. Regulation there is evolving, not settled.
To understand why the corridor matters, decompose what a remittance actually costs. Three components: a headline transfer fee, a foreign exchange spread, and a last-mile delivery cost for getting pesos into a recipient's hands. In most corridors, the FX spread is the largest of the three and the least visible. Stablecoin rails attack the middle of the transaction — correspondent banking, pre-funded nostro accounts, batch settlement windows. They do not attack the last mile. The last mile is physical, it is expensive, and it is exactly where MoneyGram already owns assets.
Put the variables together and the site selection stops looking random. A dense cash network, an existing Stellar and USDC capability, a card network with stablecoin settlement infrastructure, and a market with both remittance volume and crypto familiarity. Colombia is not the biggest market. It is the cheapest place to find out whether this works.
What A Stablecoin-Backed Card Actually Is
Strip the adjective and you have a five-layer stack. Every layer has an owner. Every owner takes a cut. If you cannot name the owner of each layer, you cannot price the product.
Layer one is funding. How does value enter the account? Either the user tops up with local currency, or an overseas sender pushes value in, or the user deposits existing crypto assets. In the remittance case — and MoneyGram's entire business is the remittance case — the sender pays dollars in the United States and the recipient's account is credited in Colombia.
Layer two is custody. Someone holds the reserves. Either a licensed custodian, or the stablecoin issuer, or MoneyGram itself. The announcement does not say.
Layer three is settlement. A stablecoin is issued on a chain and transferred. The likely candidates given MoneyGram's history are USDC on Stellar, or USDC on a Visa settlement chain such as Ethereum or Solana. The announcement does not say.
Layer four is acceptance. Visa supplies the BIN range. An issuer processor connects the program to the network. Someone is the licensed issuer of record in Colombia, because Visa does not issue cards to consumers.
Layer five is cash-out. This is the layer everyone skips and the layer that determines whether the product is real. The user needs pesos — at an ATM, at an agent, or at a merchant.
Five layers. Five sets of counterparties. Five sets of contracts. The announcement named one participant. It was not a technical disclosure. It was a distribution claim.
Now the economics, which is where the actual story sits.
When a user holds a dollar-denominated stablecoin, that dollar does not sit inert. It sits in a reserve account earning short-term government yields. In a high-rate environment, a dollar of stablecoin float can earn roughly four to five percent annually before fees. That yield accrues to whoever holds the reserve, normally the stablecoin issuer, sometimes shared with the distributor.
This is the part the press release buries. The card is not the innovation. The float is. A payments company that holds a working balance in stablecoin instead of pre-funding nostro accounts at correspondent banks captures a yield that did not exist on its balance sheet five years ago. The card is the interface that makes the float legible to the customer.
Which raises a question the announcement did not answer. Who earns the reserve yield on this float, and does any of it return to the cardholder? If the answer is MoneyGram and its stablecoin partner, then the cardholder holds a non-dividend instrument with a plastic wrapper. That structure is familiar in this industry. I have written about it in the context of governance tokens, and the shape is identical: you hold the instrument, someone else collects the cash flow, and your only exit is a later buyer.
Second economic layer: interchange. Card programs earn a fee on each transaction, shared among the issuing bank, the processor, the program manager, and the network. Latin American interchange is thinner than United States interchange, and several jurisdictions in the region have moved to cap it. If interchange is compressed, the program cannot fund rewards, and the card becomes a utility vehicle rather than a rewards vehicle. That distinction decides whether the card attracts volume or merely accounts.
Third: foreign exchange. Every cross-currency transaction carries a spread. A stablecoin rail removes correspondent banking cost from the middle of the transaction, which means the FX spread becomes the primary revenue line instead of a residual. The stablecoin does not eliminate cost. It relocates it.
Now the verification protocol. This is the same discipline I applied to ICO contracts in 2017, translated to a payments product. Eight conditions before I will call any card program functional rather than promotional.
One: a named stablecoin and a named chain. Without this you cannot verify reserve quality, redemption rights, or failure modes. USDC and USDT are not interchangeable risks.
Two: a named issuer of record and a named BIN sponsor. Without this you cannot assess licensing, capital requirements, or dispute resolution.
Three: a published reserve attestation cadence for the stablecoin. Monthly attestation is the floor. Real-time proof of reserves is the ceiling and almost nobody clears it.
Four: published settlement metrics — cost per transaction and time to finality, measured rather than described.
Five: a published cash-out map for Colombia. Which banks. Which ATM networks. Which agent locations. At what fee.
Six: a published AML and KYC architecture, including how the program handles sanctions screening on an on-chain settlement leg that can move funds in seconds.
Seven: counterparty concentration — how many stablecoin issuers, custodians, and processors sit behind the card, and what happens if one fails.
Eight: reserve income disclosure. Who earns the yield on customer balances, and is any of it passed through.
This announcement published zero of the eight. That is not a scandal. It is the normal condition of launch announcements, and it is exactly why launch announcements should never be treated as evidence of adoption.
What is defensible here, and where I give MoneyGram credit, is the combination of acceptance layer and agent network. Crypto-native card programs have died repeatedly because they had rails but no cash. The pattern is consistent. A token-funded card launches, the token price falls, rewards become unfundable, and the program is discontinued — which is precisely what happened to multiple high-profile crypto card programs in Latin America and the United States between 2022 and 2024. The cards that died all had one thing in common. They rented their cash access.
MoneyGram owns cash access. That is the actual asset. Utility is the only bridge over hype, and in remittances utility is measured in how many places a person can turn digital value into rent money.
One forward-looking architectural observation, clearly labeled as speculation. If this program runs on a programmable settlement rail — a chain with smart contract capability, a stablecoin with transfer restrictions enforceable on-chain, and a processor that accepts programmatic authorization — then the same stack extends beyond human cardholders. Autonomous software agents transacting on behalf of users need three things: a funding source, a spending limit, and a verifiable identity. Card rails supply the first two. The third requires a credential standard, and no standard for machine identity on card rails exists today outside a handful of pilots. Identity without utility is just noise, and machine identity with no spending authority is the same noise. I raise it because the architecture would support it. I am not claiming MoneyGram is building it.
The Counter-Intuitive Read
Now the part most coverage will miss.
This is not a stablecoin adoption event. It is a moat defense event.
Consider the threat model from MoneyGram's seat. A Colombian worker in the United States today sends money home through MoneyGram, paying a fee and an FX spread, and the recipient collects cash at an agent. That is the legacy flow, and MoneyGram takes a cut of both ends.
Now the alternative stablecoins make possible. The same worker opens a non-custodial wallet, buys USDC, sends it on Stellar for a fraction of a cent, and the recipient converts to pesos through a local exchange or a peer. MoneyGram is not in that transaction. Neither is Visa. The cost to the sender collapses, and two intermediaries lose their seat.
That is the real competitive dynamic. The stablecoin is not competing with the dollar. It is competing with the intermediary.
Read the card again with that lens. MoneyGram has inserted itself back into a flow from which stablecoins threatened to remove it. Visa has placed its acceptance mark on a stablecoin-funded instrument rather than watching stablecoin-funded transfers bypass card rails entirely. Both incumbents are absorbing the technology that was going to route around them. That is a rational, well-executed defensive move. It is not a revolution, and it should not be reported as one.
Second contrarian point: adoption is not what is being accelerated. Monetization is.
Stablecoin usage in Latin American remittance corridors is already happening and has been for years. What has not happened is incumbents figuring out how to charge for it. The unsolved problem in stablecoin payments was never how to get people to use it. It was how to make money on a rail whose entire value proposition is that it removes cost. This card is an answer to the second question. That makes it commercially significant and narratively banal, the reverse of how it will be covered.
Third point, and it concerns the inclusion claim specifically. Financial inclusion in Colombia is not limited by card acceptance at large merchants. It is limited by informality. A very large share of Colombian workers operate outside formal employment, which means income that arrives irregularly, in cash, often to people without conventional banking relationships. A Visa card solves a problem for the formally banked and the formally employed. It does not solve the last mile for the informal sector, because the informal sector's counterparties — the corner shop, the landlord, the relative — do not run card terminals.
So the testable question is not how many cards were issued. It is how many cardholders were previously unbanked, and what share of their monthly spend moved off cash. Those are measurable. They were not measured. In their absence, financial inclusion is a word doing the work of a number.
Fourth, and this will irritate people: the float structure deserves the same scrutiny I give governance tokens. Hold a governance token in a protocol with no fee distribution and your only return path is a later buyer. Hold a stablecoin balance in a program where reserve yield accrues entirely to the issuer and distributor, and your only return path is convenience. Convenience is real. It is also not a yield. The question worth asking is whether any participant in this stack committed to sharing reserve income with the people whose balances generate it. The announcement does not suggest otherwise. The product's economics are extractive by design. That is fine, but it should be described that way.
There is one more angle worth naming, drawn from how I handled the 2022 unwind. When liquidity vanished, the projects that survived were the ones with a defined counterparty map and a pre-written exit path. Twelve programs I audited that cycle had no documented cold-storage procedure and no named custodian in their own materials. Colombia is not a contagion scenario, but the principle holds. A card program with five layers and one named participant has no counterparty map. You cannot run an exit path on an architecture you cannot see.
What To Watch
Strip the announcement down and this is what remains: a licensed payments incumbent with a physical cash network has connected a stablecoin settlement rail to a card acceptance rail in a market chosen for its remittance density and crypto familiarity. The architecture is defensible. The economics are probably real. The disclosure is nonexistent.
Four signals will settle whether this matters.
Watch for the named stablecoin and chain. That identifies who actually captures the reserve yield.
Watch for the licensed issuer of record, and watch whether Colombian regulators publish anything about the program. A regulator who says nothing is not endorsing anything.
Watch for transaction volume and active card counts in MoneyGram's subsequent disclosures. A private company under PE ownership reports metrics when they help and stays quiet when they do not.
And watch the competition. If Western Union or Remitly announces the same structure within twelve months, you have learned that the model is copyable and that first-mover advantage was never the point.
We do not speculate; we engineer certainty. The engineering here is legible. The certainty is not yet available.
One question for the people who will celebrate this as a milestone. When the float is the product, who is the customer — the person holding the card, or the person holding the reserve?
The crypto card graveyard is full of programs that had rails and rented their cash. The survivors own the last mile. MoneyGram owns the last mile. Everything else about this announcement is still an inference.