Push-to-card payouts are one of fintech’s fastest-growing corners. But the plastic is the least defensible part of the stack, and the firms building on the card rather than the routing layer beneath it are on the wrong side of the shift.
In 2024, US government agencies loaded $183.2 billion onto general-use prepaid cards across roughly 1,200 separate programmes, according to the Federal Reserve. Every quarter brings another platform issuing another branded card to another pool of gig workers, claimants or benefit recipients.
Read that growth as a card story and you will back the wrong horse. The card is the visible edge of a shift whose value sits somewhere else: in the routing layer that decides how the money actually reaches a person.
The plastic is already a legacy endpoint. The businesses staking themselves on it are building on the part of the stack that gets designed out first. And the tell is not in the US at all. It is in the markets that built the routing layer first, from the UK to the UAE, where the disbursement card barely exists because it was never needed.
Paying out became the prize, and the card took the credit
For two decades the hard problem in payments was pay-in: accepting money faster, cheaper, from more places. Paying out was an afterthought, handled by cheques and next-day bank files.
That has flipped. Getting funds to a named recipient in minutes is now a product people compete on, and for a few years the disbursement card looked like the answer.
The mechanics explain why. A cheque takes days to arrive and clear, invites fraud, and assumes a fixed address.
The legacy bank transfer is cheap but slow and account-bound, whichever market you are in. The US ACH batch settles in one to three business days; the UK’s Bacs runs on a similar multi-day cycle. Both need the recipient’s account details, which people rarely have to hand when they are filing a claim or accepting a gig payout.
Push-to-card removed that friction by anchoring the payout to the card rather than the account: the funds land on a card the recipient already holds. Visa’s rails, Visa Direct, reach more than 195 countries and 160 currencies, and in the US require issuers to make funds available within 30 minutes; Mastercard Send covers a similar footprint.
What created the category was speed to a recipient who has only a card number to hand. That is a property of the rail, not of the plastic.
The card was the format that happened to be lying around when the demand arrived. Build a business on it and you have bet on the wrapper, not the rail inside it.
Four demand pools, and not one of them needs the card
The market is usually told as four segments. Look at each as an engineering problem rather than a product and the card keeps turning out to be incidental.
Disaster relief is the case people reach for first, and it is the card’s strongest. NOAA’s National Centers for Environmental Information recorded an annual average of 3.3 billion-dollar weather and climate disasters in the 1980s, adjusted for inflation; in 2024 the US logged 27 such events costing about $182.7 billion, the fourth-costliest year on record.
More disasters mean more emergency payouts to people whose normal banking has been disrupted, and a prepaid card can be issued without a bank account, activated on the spot and loaded remotely
Yet even here the public rails are moving in: in 2025 the US Treasury’s Digital Payout Program routed its first FEMA disaster-relief instant disbursement over FedNow. The strongest card case is already handing work to an account rail.
The gig economy is the second pool, and its headline number does not survive a close look.
Widely cited forecasts put more than 70 million Americans freelancing today, rising to about 90 million by 2028. Those figures count anyone who did any freelance work during the year, including occasional side income.
The Bureau of Labor Statistics, on a narrower “primary job” test, produces far smaller counts, and MBO Partners, which tracks full-time independents, put that group at 27.7 million in 2024. Treat the 90 million as a participation ceiling, not a payroll.
And the card in this pool is not really a payment method at all. DoorDash ran DasherDirect, a no-fee prepaid Visa with instant payout, before retiring it in April 2025 and moving drivers to a successor programme; Uber offers the Uber Pro Card on the same logic.
These are retention tools that happen to pay out: they hold the worker’s balance and pay cashback on fuel, which makes leaving the platform less convenient. Strip the loyalty mechanics and what the worker actually wanted was the money, fast, wherever they keep it.
Government benefits are the largest single segment, and here the card has its one genuine moat. The Fed’s $183.2 billion splits into $141.0 billion across state and local programmes and $42.2 billion federal.
Fraud is the forcing function: card skimming on magnetic-stripe benefit cards drove a wave of theft, and the USDA has replaced nearly $95 million in stolen SNAP benefits since January 2023, with a chip standard for benefit cards, X9.58-2024, finalised in August 2024.
What keeps the card here is not speed but control. Most spending restriction today runs on the merchant category code, the four-digit tag for what kind of business a merchant is, which is blunt: it can block a liquor store but cannot stop a SNAP recipient buying an ineligible item inside an approved grocery.
Programmes such as SNAP and WIC need eligibility at the item level, which is why basket-level control is the design frontier, and it is where the card actually earns its place.
Insurance is the fourth pool, and it shows how thin the card’s advantage is once speed is the point. Visa’s own material notes policyholders wait six to 10 days for an insurance cheque; push-to-card collapses that to minutes.
The digital insurers built the case: Lemonade says it approves and pays around 40% of claims instantly, some in as little as three seconds.
But a claim paid in seconds is a retention lever whatever carries it. The card is one way to deliver instant payout; it is not the only one, and once one insurer in a line pays instantly the pressure on the rest runs one way.
The rest of the world already skipped the card
Put the four segments together and, inside the US, the card looks like the story. Widen the lens and it stops looking like a story at all, because most of the world’s large economies solved this without the plastic.
The UK is the plainest case. Faster Payments has moved money between any two accounts in seconds since 2008, so a government or gig payout there lands as a real-time bank credit, not a branded card. The EU’s SEPA Instant does the same across the euro area. Neither market grew a disbursement-card industry to speak of, because the rail underneath made one redundant.
The Gulf is building the same thing now, faster. The UAE’s Aani platform, launched in October 2023, moves money to any bank account by mobile number in under 10 seconds. A payout rail built in 2023 has no reason to route through plastic; it goes straight to the account.
The emerging markets make the point at scale. India pushes welfare through its Direct Benefit Transfer scheme straight into Aadhaar-linked bank accounts over the Aadhaar Payment Bridge, and dropped the debit-card requirement for UPI registration in 2022. Brazil’s Pix does the same for hundreds of millions. Where a country has already built a real-time account rail, government and gig payouts route through it, and the disbursement card never gets a look-in.
The US is now doing, late, what these markets already did. Its account-to-account instant rails are growing fast: FedNow processed about $245 billion in the second quarter of 2025, and The Clearing House’s RTP network handled $405 billion in the fourth quarter, up from $80 billion a year earlier. Where a recipient has an account at a participating bank, a real-time credit is faster and cheaper than issuing a card, and it lands in the account the recipient already uses rather than a new balance to spend down.
So the US card boom is not proof the card won. It is proof the US had not yet built what the UK, the EU and now the Gulf already run. The plastic is filling a gap that the rail is closing.
The one payout the account rails still cannot reach cleanly is the cross-border one into a thin banking market, and there the emerging threat is not another card but the stablecoin. The card networks are not fighting it, they are absorbing it. Visa is piloting stablecoin payouts through Visa Direct, announced in November 2025, so gig workers and creators in underbanked regions receive USDC while businesses fund in fiat. Those are the very recipients the disbursement card was built to serve. Stablecoins settled around $28 trillion in real economic activity in 2025, per Chainalysis.
Where the card genuinely survives
The card is not finished, and the evidence shows exactly where it holds: the recipient with no bank account who also needs their spending constrained. That is the relief-and-benefits case.
A real-time credit to an account the recipient does not have is no use, and an RTP payment cannot enforce that benefit money is spent on eligible goods, which is why item-level control is the design frontier. In that corner the card is not a legacy endpoint, it is the right tool, and it will hold.
That corner is narrow and shrinking at the edges as account access widens: a defensible niche, not a market to build a company around. The card keeps the hardest, least commercial slice, the part defined by the recipient having nothing else, and loses the rest to rails that are cheaper, faster and already scaling.
The verdict: own the routing, not the plastic
For anyone building in this space, the call is not close. Do not bet the business on the card.
The durable position is the disbursement platform that treats the card as one endpoint among several and routes each payout to the cheapest rail that reaches the recipient fast enough: a real-time account credit where there is an account, whether that is Faster Payments in the UK, SEPA Instant across the euro area, Aani in the UAE, Pix in Brazil or FedNow and RTP in the US; a stablecoin where the money crosses into a thin banking market; a card where the recipient has neither an account nor unconstrained spending.
Strip out the marketing and every segment is the same engineering problem in different clothing: reach a recipient who may have no bank account, move the money in seconds not days, constrain the spend when the funder requires it, and screen for fraud and sanctions in real time across borders. The plastic is the least interesting item on that list and the least defensible.
Disbursement cards are not a niche prepaid product having a good year. They are the visible edge of institutions being forced to pay people faster than their legacy plumbing allows, and the forces behind that, more disasters, more contract workers, more benefit fraud, more instant-payout competition in insurance, are structural and are not reversing.
The demand is durable. The card is not the way most of it gets served. Whoever owns the routing decision underneath it owns the market this becomes, and the firms that mistake the endpoint for the business will spend the next few years defending a slice that keeps getting smaller.
