The most revealing funding round in payments this year was not for a wallet, an exchange or a token. It was for plumbing. Rain, whose infrastructure converts stablecoin balances into Visa-settled card transactions in near real time, raised $250 million in January at a valuation around $1.95 billion, according to industry research from Spark and Fintech Wrap Up’s enabler directory.
The round put a public price on the layer the whole stablecoin card boom runs on. Every card announcement of the past year, whichever logo it carried, was partly an announcement about the enabler underneath.
For banks, the layer matters twice over. Every one of these programs issues against a sponsor bank’s license, and the population of banks willing to sponsor stablecoin-funded credentials remains small and concentrated, which makes those signed, examined relationships the most defensible asset in the category.
What an enabler is
A stablecoin card looks simple from the outside: a Visa or Mastercard credential funded by a token balance. The apparatus behind it is not. Somebody holds the card-network membership and the sponsor-bank relationship. Somebody runs authorization, checking the token balance, pricing the conversion and answering inside the subsecond window a terminal allows. Somebody settles with the network in fiat while recouping value from a blockchain, absorbing the gap between an irreversible token movement and a reversible card transaction. And somebody carries the compliance load in every market the card is issued: KYC, sanctions screening, disputes, chargebacks, reporting.
The enabler is that somebody, packaged behind an API. An exchange, wallet or fintech can ship a branded card in weeks without becoming a regulated card operation, keeping the customer while the enabler keeps the machinery. Marqeta played this role for the card-issuing wave and Galileo before it; Rain and its peers are the same model rederived for balances that live on-chain.
Tap to settlement, in one coffee
Follow $4 through the stack. The cardholder taps, the terminal routes to Visa and Visa routes the authorization to the enabler acting for the program. In a fraction of a second the enabler identifies the cardholder’s wallet, checks a balance or a preauthorized spending allowance, prices the token-to-fiat conversion, screens the transaction and approves. The merchant sees an ordinary card payment.
The money moves later, in two ledgers. On-chain, tokens leave the spending allowance for the enabler’s treasury, immediately or in batches. In fiat, the sponsor bank settles with Visa on the network’s normal cycle, with the enabler fronting the timing gap and absorbing the spread. Multiply by millions of taps, add refunds and disputes that must reconcile across both ledgers, and the engineering under a $1.95 billion valuation stops looking abstract.
Why the bottleneck sat here
Demand was never the constraint. The constraint was institutions willing to stand between a token balance and a card network: sponsor banks were scarce and cautious, the operational risk of the conversion window sat with whoever ran authorization and network rulebooks were written for deposit accounts, not wallets. Through 2024 and 2025, would-be programs queued behind the few firms that had solved the whole stack in a form sponsors and networks would sign.
Rain’s pitch names the gap directly. “Stablecoins are quickly becoming the way money moves in the 21st century, but adoption by users worldwide requires cards and apps that just work,” said Farooq Malik, the company’s chief executive, in the funding announcement for the round, which was led by ICONIQ.
No sponsor bank has published its approval criteria for token-funded programs, and the scarcity is itself the disclosure: a queue this long behind this few institutions is what an unpublished, demanding checklist looks like from the outside.
That scarcity is what the valuation prices, and the public evidence says the queue is long. The live-program directory has grown from curiosities into a category: self-custodial products like Gnosis Pay, whose cashback pays in GNO and whose balances never leave the user’s control; spend-from-yield products like Kast; exchange staples like the Coinbase Card, paying up to 4% back in a user-chosen asset; and the enterprise programs the Visa-Bridge expansion is carrying toward 100 countries, according to the enabler directory and card-comparison research.
The self-custody frontier
The hardest edge of the business is the one growing fastest. First-generation crypto cards spent from custodial balances, which reduced authorization to a database check. The current frontier keeps custody with the user: the balance sits in the cardholder’s own wallet until the moment of spend, and the enabler must reserve, pull and convert against a signed authorization inside the network’s window.
The trade is a compliance gift bound to an engineering burden. Self-custody keeps client assets off everyone’s balance sheet, which simplifies the regulatory posture in exactly the jurisdictions writing custody rules now. It also moves settlement risk squarely onto the enabler, who fronts fiat against a token movement that cannot be reversed if fraudulent and cannot be compelled if the wallet owner revokes.
Pricing that risk, and building the on-chain permission systems that contain it, is where this generation of enabler engineering actually lives.
The margins and the moat
Enabler revenue stacks three margins: a share of interchange, a conversion spread on the token leg and platform fees. Against them sit the costs that keep amateurs out: sponsor-bank economics, network fees, fraud and dispute losses on the reversible side of an irreversible funding source, and a compliance organization sized for the client list’s jurisdictions.
Scale decides everything. Authorization infrastructure is expensive at one program and nearly free at the fiftieth, sponsor banks price comfort and comfort compounds with track record. So the layer consolidates early: firms that clear the operational bar attract programs, program volume funds the compliance moat, the moat attracts the next sponsor bank. A quarter-billion war chest is a bid to make that loop unaffordable to chase.
The cashback wars ride on the same arithmetic. Rewards are financed from interchange, which regulation caps in Europe and contests everywhere, plus token incentives from ecosystems buying distribution, with Gnosis funding its card’s GNO rewards through a dated program the transparent example of the second source. Headline rates that exceed what the arithmetic supports are burning someone’s budget, and mature markets converge on what interchange actually funds, region by region.
Read more: Mastercard’s Yellow Card bet awaits its first priced corridor • New payment chains court banks with rails built for dollars
“Stablecoins are quickly becoming the way money moves in the 21st century, but adoption by users worldwide requires cards and apps that just work.”
Farooq Malik, Rain CEO
Where the demand actually is
The growth map does not match the card industry’s traditional one. In the U.S. and Western Europe, a stablecoin card is a convenience competing against excellent incumbents. The volume thesis lives where the card is the banking upgrade: Latin America, where dollar balances are the savings account; Africa and Southeast Asia, where card penetration is thin and wallet penetration is total; and the global freelancer economy, paid in tokens and spending across borders by default.
That is why enabler expansion announcements read like emerging-market licensing tours, and why the Visa-Bridge push toward 100 countries matters to the whole layer. Every market that opens for stablecoin-funded credentials opens for the infrastructure underneath. The addressable market is not crypto-community spending; it is the several hundred million people holding tokenized dollars for whom a card is the missing half of a bank account.
The history this wave answers
This is the third attempt at crypto cards, and the design is a point-by-point response to the first two. The first era died in 2018, when a single European issuer’s failure took down most of the market’s cards in a weekend and taught that borrowed licenses are single points of failure. The second, exchange cards spending volatile assets from custodial balances, taught that the funding asset and the custody model decide the risk, not the plastic.
A market pricing its plumbing at unicorn multiples is a market that has internalized its own history. One legacy constraint survives: sponsor-bank concentration, a softer version of the structure that killed the first era.
The licensing regimes should widen the funnel, since a licensed stablecoin is easier for a sponsor’s examiners to bless than the unclassified assets of 2021. Until sponsor capacity commoditizes, the enablers’ most defensible asset is not code.
The regimes will answer the question themselves within a few examination cycles. A licensed funding asset either widens the population of banks whose boards will sign, or the concentration persists and keeps pricing the layer.
What the round will buy
The company’s own account of the round points at scale first. “In the last year, our active card base has increased 30x and our annualized payment volume has increased 38x, but we’re still in the early innings,” Malik said in the announcement. “This funding lets us bring that infrastructure to new markets and help additional enterprises go live and scale quickly everywhere.”
Follow the money to see which constraint binds next. Sponsor-bank breadth is the likeliest first spend, because every new bank relationship multiplies the programs the platform can sign and dilutes the concentration risk clients price. Market licenses come second, since each jurisdiction added is one the client list can expand into without their own regulatory build. Authorization scale is the third bucket, the unglamorous engineering that keeps approval rates high as volume compounds, and the one clients only notice when it fails.
There is a fourth use no term sheet names: credibility inventory. In a business where sponsor banks price comfort, a balance sheet with a quarter-billion of fresh capital is itself a risk answer. Some of what a mega-round buys in this layer is simply the ability to be the counterparty nobody gets fired for choosing.
The program mix, read as a market map
The live directory splits into recognizable species, and the split predicts where enabler volume comes from. Exchange cards monetize existing custodial balances; they scale fast and churn with markets. Wallet cards, the Phantom and MetaMask pattern running through Bridge, attach spend to balances that already exist for other reasons, which is the healthiest demand in the category. Yield-adjacent cards, Kast and the Ethena Pay launch, use rewards economics as acquisition and inherit the sustainability questions of whatever pays the yield. Self-custodial cards, Gnosis Pay’s lane, trade operational complexity for a regulatory posture that travels well.
An enabler’s client mix across those species is effectively its risk profile. The valuations in this layer will eventually be read the way processor valuations always have been: on the durability of the programs underneath, not the count.
The verticalization threat
Two of the year’s larger stories cut across the layer from above. Stripe owns Bridge, which sells card issuance alongside token infrastructure, and Visa’s new platform handles issuance and lifecycle natively for network clients. When the rails’ owners sell the enablement, the independent must be better, faster or more neutral.
Neutrality is the historically proven path: exchanges and wallets hesitate to build card programs on a competitor’s platform, and independent processors have ridden exactly that hesitation to large outcomes before.
The dispute problem, concretely
One operational detail decides more of the profit-and-loss statement than any headline feature: disputes. Card rules give consumers reversal rights for months, and token movements reverse never. When a fraudulent transaction clears, the enabler has typically already swept tokens it must now make the network whole against, and when a legitimate refund processes, value flows back through fiat rails toward a wallet the enabler cannot credit without its own token outlay.
Every enabler builds machinery to manage that mismatch: provisional holds, delayed sweeps, reserve pools sized to loss history, contractual recourse against program partners. The machinery is invisible until a large program has a fraud event, at which point it is the entire business. Loss rates on the reversibility gap are the number sophisticated investors diligence hardest, and the number no public announcement has ever included.
How the layer will get measured
The category still lacks a public scoreboard, and its maturity will show in which numbers leak first: total card spend processed, the layer’s equivalent of total payment volume; active funded cards, separating shipped plastic from used product; authorization approval rates, where funding-source engineering shows up in basis points; loss rates across fraud and disputes; and sponsor-bank concentration, the systemic number nobody volunteers.
Private markets have plainly seen enough of these to price the round. Public understanding runs on launch announcements and country counts.
The first enabler to publish spend volume the way processors publish total payment volume will simultaneously prove the layer’s scale and start the consolidation clock on everyone who cannot match the disclosure. The last fintech cycle minted its giants in the layer nobody’s customers could see, and this round says the stablecoin cycle has found the same layer.