A weekly email on crypto card fee changes, new cards and cashback cuts.
The Sweepbase crypto cards index 2026: 178 cards analyzed
We keep a single CSV with every crypto debit and credit card we have ever tracked. There are 178 rows in it right now: 141 cards you can get today, 34 that no longer accept new applicants (kept with shutdown dates and sources), and 3 announced products nobody holds yet. This page is a straight read of that file: where the money in the market sits, how many cards charge real FX fees, how many are self-custody in practice (not just in marketing), and which few cards manage all of it at once. If you have ever wondered how many crypto cards there really are or what an average one looks like, this is the distribution.
Everything below is computed from the same data that drives individual card pages. The numbers refresh whenever the site is rebuilt. If a value looks off, the card detail page is the source of truth, and a confirmed error gets fixed in the data file itself.
178
Cards tracked
141 obtainable, 34 discontinued
18%
Self-custody
25 cards keep keys with the user
21%
Charge 0% FX
30 cards
53%
Pay no cashback
75 cards
43%
Fee-free
No issuance, no annual fee
3
Rare gems
Match every box at once
Shares are over the 141 cards you can get today; closed and pre-launch rows are left out. Hover or read the table sections below for the full breakdown.
Four numbers from the file
- 18% of cards are self-custody (25 of 141). The rest hold user funds on an exchange or e-money account until the moment of purchase.
- 53% of cards pay no cashback at any tier (75 of 141). At the other end, 21 advertise 5% or more somewhere on their ladder, usually behind a token stake, a paid tier or a monthly cap.
- 21% of cards charge 0% FX on foreign-currency transactions (30 of 141). 39 charge 2% or more, and 20 publish no clear FX fee at all.
- 3 cards combine self-custody, 0% FX, no annual fee, and a non-zero cashback. These are the “rare gems” named below.
How I built the dataset behind this index
The 178-card sample is everything Sweepbase has ever tracked as a real crypto debit or credit card: an issuer with at least a virtual card in production, or one announced on the issuer's own site, and public documentation I can read. Of those rows, 141 are obtainable today. 34 are discontinued, and 3 are announced products still waiting on their first cardholder; those carry a pre-launch flag and stay out of every obtainable count. What I exclude entirely is vapor: cards that show up in press releases but never in the issuer's own product navigation. When a card dies, I do not delete the row. It keeps its place in the database with a discontinued flag, the date applications closed, the reason, and a source link, then drops out of the active listings and rankings. The shutdown history is half the point of running an index like this.
Every column in the database maps to a question I have asked at least once when deciding whether to use a card: issuance fee, annual fee, FX rate, cashback rate, top-up methods, supported chains, KYC region, custody model. One check runs on top of the columns: it counts how many of five key fields (network, custody, regions, description, cashback) a card leaves empty or marked not applicable, and that count is how I keep the index honest about which products have enough public information to be reviewed responsibly. Cards with three or more missing fields are flagged thin because there is not enough out there to write a usable comparison entry.
Custody split
Only a minority of cards in the Sweepbase database are true self-custody, meaning the private keys stay in the user's wallet until a purchase happens on-chain or through a signed intent. The rest are custodial: the issuer or its banking partner holds the balance in an omnibus account, and the card debits that balance at the point of sale. The file puts the self-custody share at 18% today (25 of 141). Cards that pair a self-custody wallet with a card balance the issuer holds, such as Ledger and Jupiter, count as custodial in this split.
| Model | Cards | Share |
|---|---|---|
| Self-custody | 25 | 18% |
| Custodial | 116 | 82% |
Network split
Visa sits on 113 of the 141 obtainable cards and Mastercard on 50. The two add up to more than 141 because some programmes ship on both networks, or switch network by region: Bybit is Mastercard in the EEA and Visa in Asia-Pacific. 1 card names neither network or no network at all.
| Network | Cards | Share |
|---|---|---|
| Visa | 113 | 80% |
| Mastercard | 50 | 35% |
| Other / unspecified | 1 | 1% |
Cashback distribution
The distribution splits in two. 93 cards top out under 2% at their highest advertised rate, most of them at zero, while 21 advertise 5% or more somewhere on the ladder. Those high rates usually sit behind token staking, a paid tier or a monthly cap, and the bucket reads the top rate in the field, not the one a new applicant gets.
| Cashback (top advertised rate) | Cards | Share |
|---|---|---|
| None (no cashback) | 75 | 53% |
| Below 1% | 4 | 3% |
| 1% to under 2% | 14 | 10% |
| 2% to under 5% | 27 | 19% |
| 5% and above | 21 | 15% |
| Card | Value |
|---|---|
| No cashback | 75 cards |
| Below 1% | 4 cards |
| 1-2% | 14 cards |
| 2-5% | 27 cards |
| 5%+ | 21 cards |
The shape that surprises most readers: about half the obtainable cards pay no cashback at any tier (buckets use each card's top advertised rate), and the 5%+ rates usually sit behind staking, a paid tier or a monthly cap.
Only 53% of cards pay cashback in any real sense once fields that describe only fee discounts, rewards on someone else's spending or programmes not yet live are set aside. A few high-headline cards skew perception. See our Real Fees Top 10 for the cards that keep their rate after fees.
FX fee distribution
Foreign-exchange markups are the quiet fee that eats more money than any annual card charge. A card that advertises 2% cashback but charges 3% FX is net-negative on every overseas purchase. The mechanics (explicit FX line vs. embedded conversion spread) are unpacked in our crypto card fees explained primer.
| FX fee | Cards | Share |
|---|---|---|
| 0% FX | 30 | 21% |
| Below 1% | 5 | 4% |
| 1% to under 2% | 47 | 33% |
| 2% and above | 39 | 28% |
| Not disclosed | 20 | 14% |
| Card | Value |
|---|---|
| 0% FX | 30 cards |
| Below 1% | 5 cards |
| 1-2% | 47 cards |
| 2%+ | 39 cards |
| Not disclosed | 20 cards |
The 'not disclosed' column is the one that should worry buyers most, it captures cards where the FX cost is buried in the spread or only revealed after KYC. Each of those card pages prints the FX line as the issuer wrote it, gap included.
Feature shares
How many cards in the database carry each of the common feature tags:
| Feature | Cards | Share |
|---|---|---|
| Zero issuance and annual fees | 61 | 43% |
| No mandatory KYC | 4 | 3% |
| Apple Pay or Google Pay | 113 | 80% |
| A yield product, on the card balance or beside it (cards that earn on the balance you spend) | 87 | 62% |
Rare gems that tick every box
Each of these cards is self-custody, charges 0% FX, has a free entry card with no annual fee, and advertises a non-zero cashback. Very few products combine all four because each one costs the issuer margin somewhere, and several pay the cashback in their own token, which shifts part of that cost onto you.
| Card | Cashback (advertised) | Network |
|---|---|---|
| Bleap Card | 1% | Mastercard |
| Ether.fi Cash Card | Up to 3% (Tiers) | Visa Signature |
| MiniPay Card | Up to 5% (in XAUt, capped) | Visa Debit |
Region availability, top-up methods and spending caps still differ between these cards, so check each one's detail page before you apply.
How we counted
Each card in the Sweepbase database is a CSV row with 46 columns pulled from issuer documentation, regulatory disclosures, and direct testing (the public card model exposes 25 of them after validation). Every distribution on this page is computed over the 141 cards you can order today. The 37 closed and pre-launch rows stay in the file and out of the shares. For this report we group the fields into simple buckets:
- Cashback tier: we read the headline rate each card page shows (the highest figure in it) and bucket it (none, below 1%, 1-2%, 2-5%, 5%+). Fee discounts and rebates count as no cashback. The rate is the advertised top tier, and most applicants get less.
- FX tier: the rate the “FX Fee” field says the card charges on a purchase, read clause by clause. A range counts at its top, a field that says the rate is not published counts as not disclosed, and a 0% that applies only to a top-up or a single currency does not count as 0% FX.
- Self-custody: a card counts if its Custody field opens with self-custody, self-custodial or non-custodial. A field that opens with hybrid (your own wallet, but a card balance someone else holds) counts as custodial.
- No-fee: both the issuance and the annual fee field must open with a free card, not just one of them. A paid metal card or replacement further down the field does not disqualify a free entry card.
The full rating formula is on the Methodology page, and the raw per-card data is published on every card detail page. If any number on this page is wrong, the card page is the source of truth and we will fix the number that drives the chart, not the chart itself.
Editor's notes on this cut
Editorial commentary from Mihail B., who runs the database refresh.
Self-custody covers 18% of the index
The self-custody cards worth knowing by name are spread across regions. Bleap serves the EEA and Switzerland. Solid counts 105 countries, the UK and every EU and EEA state among them. Ether.fi lists 91 countries and territories, and the MetaMask Card 47. Gnosis Pay went the other way: it closed to new sign-ups on 30 September 2026, and its card and web app stop on 20 December 2026. The balance still tips heavily toward custodial cards.
The “rare gems” filter is unforgiving for a reason
3 cards in the database satisfy all four constraints (self-custody, 0% FX, no annual fee, non-zero cashback). It is a hard set to populate, since each property costs the issuer margin in a different place. Cards that reach it usually do so by subsidising one dimension with token economics. Bleap pays cashback in USDC (no token risk, but the higher tiers are gated by category); Ether.fi pays in ETHFI, which transfers volatility risk from the issuer to the user. Keep that in mind before treating a “rare gem” listing as unconditional.
FX disclosure is still the worst part of the market
20 cards in the index do not publish a clear FX fee, which is roughly 14% of the catalog. Some write “standard network rates” and leave it there, others say nothing at all. The rating formula withholds the low-FX bonus from a card that prints no rate, and that is all it does. If you arrived here from a card review and that card sits in the “not disclosed” bucket, the absence of a number is itself the number.
How this page stays current
Every chart above regenerates on each site rebuild from the live database, so the distributions move only when rows are re-read. What changed in each row, and when, is logged in the data changelog.
Sources & references
This is what we read to write the guide: regulators, issuer fee schedules and archived snapshots. Not every sentence above has its own link here. If a number looks wrong, start with these.
- Sweepbase Crypto Cards Dataset (CSV, CC-BY 4.0), SweepbaseSource data for every chart and statistic in this index, released under CC-BY 4.0 with provenance per row.
- Visa Foreign Exchange Rate Calculator, VisaVisa's own conversion rate, the baseline that each Visa card's published FX markup is added to.
- Currency Conversion Calculator, MastercardThe same baseline for Mastercard cards.
- Markets in Crypto-Assets (MiCA) Regulation, European Securities and Markets AuthorityEU rulebook that applies in full from 30 December 2024; it shapes the regional-availability column for every card sold in the EEA.
- FATF VASP guidance, Financial Action Task ForceInternational AML standard that explains why every card in the index either does KYC or limits balances.
FAQ
Most crypto card "best of" posts rank 5 to 10 cards. This is an index of the full market that Sweepbase tracks: 178 cards, 141 of them obtainable today, with the distribution of fees, cashback, custody, and FX markup disclosed in aggregate. The raw rows come from our own database (linked on every card page), so you can reproduce any number in this report.
A card that combines self-custody, 0% FX fees, no annual fee, and a non-zero cashback. Very few cards meet all four because each property costs the issuer margin in a different way. The list changes as new cards launch and existing ones change their fee schedules.
Because the chart buckets each card by the highest rate it advertises, and many crypto cards advertise a headline rate that only applies to the highest tier. The rate a new user with no staked tokens gets is usually far lower: Crypto.com's Basic tier pays 0%, and Bybit's EEA card starts at 1%. High rates (5%+) usually sit behind a token lockup, a paid membership or a tight monthly cap.
The catalogue is re-reviewed in rolling passes rather than end to end on a fixed day, so review dates differ from card to card and each card page prints the month of its own last review. Confirmed fee changes are applied ahead of the scheduled pass. The numbers on this page are regenerated whenever the site is rebuilt, so they match the current card pages exactly.
Yes. The full dataset ships as a CSV under CC-BY 4.0 on our dataset page, the same file that powers every chart in this report. If you only need one row, each card detail page publishes its data too.
Headline rates rarely survive the fee print
Three numbers from the charts above are the ones to carry away. The 5%-plus cashback bucket holds 21 cards, and the rate in it is usually a top-tier rate that a new applicant will not get. 18% of cards are true self-custody. And the FX column to worry about is not the 2% one but the one marked not disclosed, because a fee that is not printed is a fee you cannot budget for.
Browse the full card database, see what changed in our methodology, or jump to the Real Fees Top 10 for a ranked net-value view.