Stablecoin cards have now crossed $10 billion in transaction volume. Yet the infrastructure powering them is still taking shape. 💳 In our latest F-Prime deep dive, I map the emerging stablecoin card market and examine where the next $100 billion of volume could come from. Full-stack providers such as Rain and Reap dominate today, but a modular model is emerging. Lightspark’s Visa stablecoin debit card program brings together Lithic and Lead Bank, while Marqeta’s partnership with zerohash shows how existing fintechs can add stablecoin capabilities without rebuilding their card programs from scratch. We believe the addressable market for stablecoin cards is an order of magnitude larger than the one being contested today. Fintech, neobank, and commercial card programs represent ~$5.5 trillion of annual volume. Moving just 1% of that volume would create a $55 billion market. There's still plenty of room to build. Hercle is working on 24/7 settlement and conversion between stablecoins and fiat currencies. Pave Bank is building a regulated bank designed around stablecoin, while Karta is bringing stablecoin-powered card into private wealth. If you are a payments veteran building the next issuer processor or other part of the infrastructure, I want to hear from you. Full article in the comments 👇. #stablecoins #payments #fintech
The layer this map understates is authorization timing. A card auth needs guaranteed funds in about 2 seconds. Onchain settlement finality doesn't move at that speed, so someone in the modular stack is always fronting the float between the swipe and the chain confirming. Building agent payment rails on x402 I hit the same wall. The 402 gate authorizes instantly, the USDC leg settles later. That gap is where the real infra work lives, not in card issuance. Whoever owns that reconciliation layer owns the economics.
OZAV built the whole vertical stack onchain, but our view this infrastructure will be commodity, but the date will be where infrastructure value compounds
The problem with stablecoin cards is that the money is actually made by the same tradfi companies as always. The infrastructure layers are still dependent on them. The crypto cards issuers are simply monetizing their distribution channels. So there’s no reason for all fintechs to offer stablecoin cards, even if they offer stablecoin remittances. By that time there’ll be more convenient and cheap off-ramp to offer
Jingwei Li Kulipa and Rain are the two cases the modular vs. full stack debate keeps stepping over. The fact that all 20 programs had to be put on "hold" when one processor went out of business was an issue with portability-not operational. What will happen to BINs, Tokenized Credentials, Card Holder Data; etc. in three or four months time when the processor goes out of business? All of these issues result in issuances of new cards and customer turnover-the true cost of a simply "take rate" comparison. Entering stable coin rails without changes to BINs is much less important than having a well-defined plan for how you can get out of stablecoin rails. As such, I believe that having a processor who has clearly documented an exit strategy and has escrowed tokens/credentials (and provides a defined process for transferring BINs) provides something that most processors cannot because nobody worried Fiserv would disappear. In my view, that converts more of the $5.5T.
Don't forget Pomelo which is actually the closest to being full-stack out there.
From a card acquiring and issuing standpoint, the modular model makes sense, but the real bottleneck is who owns settlement, conversion, and reconciliation across providers. That operational layer will determine whether stablecoin cards scale beyond early adopters. 💳
The part I keep coming back to is where that $10B ends up. The card spends a stablecoin, the merchant still receives fiat. Same rails, nicer front door. We (LFG) went the other direction and built for acceptance. Money lands in the merchant's own wallet in seconds, from anywhere, screened before it arrives, and we never touch it. No middleman picking the currency or taking a slice on the way. Cards are the half everyone can see. Acceptance is where I think the next chunk of that $5.5 trillion shows up. Check us out www.getlfg.app
Jingwei Li, the discussion around modular infrastructure raises an important question for agent-driven payments: who can reconstruct the complete decision when several providers each hold only part of the record? If an agent initiates a purchase, we should be able to connect its identity, delegated authority, decision context, and final outcome across those handoffs. That includes failed attempts and retries, where the history can matter just as much as the completed payment. This is the infrastructure I’m focused on building: persistent identity, contextual memory, and verifiable accountability for agents operating across systems. As the payment stack becomes more modular, we should design for accountability to remain continuous.
https://fprimecapital.com/blog/stablecoin-cards-the-infrastructure-race-for-the-next-100-billion/