What Card Issuing Actually Costs
The real cost structure of a card programme: issuing fees, load fees, FX, the fixed-plus-percentage model, and the costs that never appear on a price list.
In short
Card costs come in four layers: per-card fees, per-load fees, FX, and the operational costs nobody quotes. Understanding the fixed-plus-percentage model is most of what you need to compare providers honestly.
Card pricing is quoted in ways that make comparison hard — sometimes deliberately, sometimes because the underlying structure genuinely is complicated. Four layers cover nearly all of it.
Layer 1: The cost to create a card
Almost every provider charges to issue. The structure is usually a fixed amount plus a percentage of the initial load, and understanding why matters for choosing a provider.
The fixed part covers the operation itself: a network call, a record, a slot in the programme. It does not care whether the card holds $5 or $5,000. The percentage covers the money at risk and the float.
This means the same headline rate produces wildly different effective costs:
| Rate | On a $20 load | On a $500 load |
|---|---|---|
| $5.00 + 1% | $5.20 (26% of load) | $10.00 (2% of load) |
| $1.00 + 3% | $1.60 (8% of load) | $16.00 (3.2% of load) |
| $0.50 + 4% | $1.30 (6.5% of load) | $20.50 (4.1% of load) |
The lesson: know your average load before you compare providers. A programme issuing $20 cards and a programme issuing $500 cards should choose differently, and a provider that looks cheap for one is expensive for the other.
Layer 2: The cost to load
Adding money to an existing card is usually priced the same way — fixed plus percentage — and usually cheaper than issuing, because there is no card to create.
The design decision this drives: is your programme many small cards, or few cards topped up often? If issuing costs $5 and topping up costs $0.50, then a card per purchase is expensive and a card per user topped up per purchase is cheap. That is an architecture decision made by a price list, and it is worth making deliberately.
Layer 3: FX
If your card is denominated in USD and your user spends in euros, someone converts. That conversion has a rate and usually a margin, and the margin is frequently larger than the issuing fee on the same transaction.
Two questions to ask any provider: whose rate is used, and what margin is added. "The network rate" and "the network rate plus 2%" are very different answers, and the second is common.
Layer 4: The costs nobody quotes
These do not appear on a price list and frequently exceed the ones that do.
| Cost | Where it comes from |
|---|---|
| Support | Declines, lost cards, "why did this not work". Scales with cardholders, not volume. |
| Disputes | Chargeback handling, evidence gathering, and the losses you absorb. |
| Compliance | On a sponsored programme, a permanent function. On a network, the network's cost. |
| Engineering | Integration, then maintenance. Reconciliation, retries, provider changes. |
| Float | Money sitting in a prepaid balance is money not doing anything else. |
| Failed loads | Deposits that arrive wrong, on the wrong chain, or below a minimum. |
Support is the one that surprises people most. A programme with 10,000 cardholders generates support volume regardless of how elegant the API was, and it is a recurring operational cost rather than a one-off integration cost.
What about interchange?
Interchange is the fee an acquirer pays an issuer on each transaction. In a sponsored programme it flows partly to you and can be substantial at volume — for some card businesses it funds everything else.
On a card issuing network like ON5, interchange is not part of the model. That is a genuine difference, and if your business case depends on interchange revenue, a network is the wrong shape and you should be talking to sponsors. If your business case is "our users need a way to spend their balance", it is irrelevant.
Comparing providers honestly
- Model your real distribution. Not your average — your distribution. If 80% of cards are $25 and 20% are $2,000, compute both.
- Include the load fees at your real top-up frequency. A cheap card that is topped up weekly may cost more per year than an expensive card topped up once.
- Ask about FX explicitly, with the margin stated as a number.
- Add the operational costs. Support and reconciliation are real headcount.
- Then compare. Very often the headline issuing fee turns out to be a minority of total cost.
A note on fee snapshots
Whatever provider you choose, check that fees are snapshotted onto the transaction rather than computed from a current rate table. If a rate change silently rewrites historical charges, your accounts will not reconcile and your statements will be wrong retroactively. ON5 stores the fee that applied with the transaction itself, so changing a rate never alters what was already charged. It is a small detail that matters enormously at audit time.
Frequently asked questions
How are card issuing fees usually structured?
Typically a fixed amount plus a percentage of the load. The fixed part covers the operation, the percentage covers the money at risk. The same headline rate produces very different effective costs depending on load size.
What is the cheapest card issuing pricing model?
It depends on your average load. A high fixed component with a low percentage suits large loads; a low fixed component with a higher percentage suits small ones. Model your real load distribution before comparing providers.
Does ON5 charge interchange or share it?
No. Interchange is not part of the card issuing network model. Programmes that depend on interchange revenue need a sponsored arrangement instead.
What card programme costs are usually overlooked?
Support, disputes, compliance, ongoing engineering, float, and failed loads. Support in particular scales with cardholder count rather than transaction volume and is a recurring operational cost.
Issue your first card on ON5
Fund an account with USDT or USDC and issue a branded Visa or Mastercard virtual card. The minimum is $5.
Open the dashboard