What I learned from putting 140 crypto cards into one spreadsheet

Six months ago I started a spreadsheet to track crypto debit and credit cards. The first row was Crypto.com Visa. The second was Coinbase. By the end of that first week I had twenty cards. Now there are around 140. Some things only become obvious when you put them all next to each other.

Here are the patterns I did not expect.

Most "global" cards are actually five to ten countries

Marketing pages like to say "available worldwide" or "EEA, UK, USA, and more." When you click into the actual eligibility list, the long tail is rarely longer than ten markets. Truly global cards exist but they are a minority. The honest way to filter is by your specific country, not by the global label.

Stablecoin support is more concentrated than you would think

USDC works on basically every card that supports stablecoins. USDT is on maybe seven out of ten. DAI is on three out of ten and shrinking. If your funds are mostly in something else, the card list narrows fast.

Visa and Mastercard split cleanly by region

In USA listings, Visa dominates. In Europe, Mastercard is more common. There are exceptions but the pattern holds. Probably an artifact of which network has easier issuer requirements in which jurisdiction.

Cashback rates cluster around two percent

When you ignore the marketing tier (like "8 percent if you stake $40,000 for six months") and look at the realistic tier most users hit, the distribution lands close to 2 percent. Outliers above 3 percent usually have heavy conditions. Outliers below 1 percent are often paired with no monthly fees, which is its own kind of return.

Self-custody is rarer than the online noise suggests

Most discussion online treats self-custody crypto cards like a mainstream option. In the data, custodial cards outnumber self-custody by about nine to one. The self-custody options are mostly Gnosis Pay wrappers (Rebind, Zeal, Picnic, Gnosis Pay itself) plus a handful of independent ones. If you want non-custodial, your real choice set is small.

None of this is a recommendation. Each pattern is just what 140 rows reveal once you stop reading them one at a time. If you want to filter the same data by your own constraints, the open version lives at sweepbase.net. The fee calculator at sweepbase.net/calculator lets you test what each card would actually cost for your spending pattern, after FX margin and the cashback you qualify for.

The full card database stays free to browse. I update it on most weeks. When readers spot errors and send a report, those get fixed within a day or two.