MothYou have money on the card. The card is active. You enter the details on an AI service's checkout page, and the payment fails anyway — often with nothing more useful than "your card was declined."
Most people assume the card is broken, top it up again, and try a second card. Then the second card fails the same way. The reason is usually not your balance and not your card being faulty. It is where your card was issued.
The first six digits decide more than you think
Every card number starts with a BIN — Bank Identification Number, the first six to eight digits. The BIN tells the merchant two things before a single cent moves: which institution issued the card, and which country or region it was issued in.
That second part is what trips people up. Merchants can and do write rules against it. A checkout can be configured to accept cards issued in some regions and refuse cards issued in others, regardless of how much money sits behind them.
AI services are among the stricter ones. They sell a product that is easy to abuse at scale — free trials, API credits, bulk account creation — so they lean hard on payment-side risk controls. Filtering by issuing region is one of the bluntest and cheapest of those controls.
This is not a theory about how these systems might work. OpenAI's own troubleshooting documentation has a section headed "Verify Your Country or Region", which asks you to confirm that both you and your card issuer sit in a supported country, and states plainly: "Payments may be declined if either party is outside of the supported regions."
Read that as two separate conditions. Where you are, and where your card was issued. Either one failing is enough, and the second one is not something you can change by topping up, retrying, or switching browsers.
The practical consequence: a card issued in a region that a given AI service does not accept will be refused at authorization, before your balance is ever consulted. Topping up more money cannot fix a rule that never looked at your money.
Hong Kong-issued cards run into this most often
If you bought a virtual card from a crypto-friendly provider, there is a good chance it was issued on a Hong Kong BIN. Hong Kong is a common issuing jurisdiction for this category of card, which means a lot of cards in circulation share the same handful of BIN ranges.
That concentration is exactly what makes them easy to block. When a merchant's risk team sees abuse clustered on a small set of BIN ranges, blocking the range is a one-line change on their side — and every legitimate user on that range gets caught along with the abuse.
This is worth stating plainly: a Hong Kong BIN being refused is not evidence that your card is low quality. It is evidence that the merchant made a geographic decision. The same card will often work perfectly on cloud bills, streaming, and general online shopping, and fail only on the handful of merchants that filter this way.
Telling BIN blocking apart from the other four causes
Before you buy another card, it is worth knowing which problem you actually have. Five things cause most declines, and they look different:
| What's happening | How it usually looks | Does adding money help? |
|---|---|---|
| Issuing-region (BIN) rule | Fails instantly, every attempt, same merchant. Other merchants work fine. | No |
| Not enough balance | Fails only above a certain amount; smaller charges go through | Yes |
| 3-D Secure / verification step | You get bounced to a verification screen, or it never appears and the payment times out | No |
| Address or postcode mismatch | Fails at checkout after you enter billing details | No — fix the billing address |
| Merchant's own risk score | Inconsistent: sometimes works, fails on retry, may clear on a different device or after a day | Sometimes |
The signature of a BIN rule is consistency plus specificity: it fails every single time on that one merchant, immediately, while the same card happily pays elsewhere. If that is your pattern, no amount of retrying, topping up, or clearing cookies will change the outcome.
What actually fixes it
Only one thing does: paying with a card issued in a region the merchant accepts.
That means the question to ask before buying a card is not "is this card good" but "where is this card issued, and is that region accepted where I need to pay". Providers rarely lead with this. It is usually buried, or described only as a country flag on the card art.
Three things worth checking before you pay for any card:
- The issuing region, stated explicitly. Not the company's headquarters — the card's issuing jurisdiction. These are often different.
- Whether the provider has actually tested the merchants you care about. "Works everywhere Visa is accepted" is a statement about the network, not about the merchant's risk rules. Those are separate layers, and the second one is where AI services block you.
- What happens when a payment fails. A refused authorization should cost you nothing. Find out whether failed attempts, refunds, or disputes carry fees before you need the answer.
Where MothCard fits
We issue cards, so treat this section as what it is — but the mechanics above hold whoever you buy from.
MothCard runs cards on several issuing regions rather than one, specifically because a single BIN range cannot cover every merchant. Two of them have been tested against AI-service checkouts and clear them: a UK-issued card and a Singapore-issued card. Cards on our Hong Kong BINs are kept for the many merchants where they work well, and we do not position them for AI subscriptions, because that would be setting you up to fail.
A few things we would rather you know before signing up than after:
- Cards are funded from your wallet balance, and the wallet is topped up with USDT. Fiat top-up is not available yet.
- Opening a card requires a one-time identity check. There is no way around it, and a provider that offers you one is telling you something about itself.
- Cards settle in USD. Paying in another currency adds a conversion fee, so USD-priced subscriptions are the cheapest case.
- We do not support withdrawing funds back out. Money you move in is for spending.
If your AI subscription is failing on a card you already own, the diagnostic table above will tell you in about a minute whether you are looking at a BIN rule or something else. That is worth doing before you spend money on a replacement — including ours.
Related reading
- What you're actually trading away when someone else pays for your subscription — the usual next step after a card keeps failing, and why it costs more than it looks.
- Is a "no KYC" crypto card actually safe? — why "no identity check" is a warning sign rather than a feature.