How to Spend Crypto Profits: 8 Cards Compared

On a $1,000 load with half of it spent in a foreign currency, the cheapest card here costs nothing and the most expensive costs $56. That is the whole article in one sentence, and the gap is almost entirely down to the top-up fee, which is the number providers put in the smallest type.

First decide which of the three things you actually want

Every card here is described as a crypto card, and they do three genuinely different things. Picking the wrong category costs more than picking the wrong product inside a category.

• Custodial debit. You send crypto to the provider, they hold it, and the card spends from that balance. RedotPay, KAST, Bybit, Nexo in debit mode, Kardpay and SolCard all work this way. Simple, widely available, and your balance sits on someone else's book.
• Self-custodial debit. The card spends directly from a wallet you control. Gnosis Pay settles from a Safe; Ether.fi Cash settles against collateral you still own. No top-up step exists because there is nowhere to top up to.
• Credit against collateral. Ether.fi Cash and Nexo's credit mode lend you fiat against your crypto rather than selling it. You avoid the disposal, and you acquire a loan with a liquidation price attached.

That third category is the one people misunderstand. "Spend without selling" sounds like a tax trick with no downside. It is a margin position: if your collateral falls far enough, it is sold for you, at the worst possible moment, and now you have both the tax event and the loss. Borrowing to spend makes sense when you have high conviction and a comfortable buffer. It does not make sense as a default.

Where crypto cards actually charge you

There are five places a card takes money, and marketing pages usually foreground the one that flatters the product. Compare all five or you are not comparing anything.

1. Top-up or conversion fee. Charged when crypto becomes spendable balance. This is the single largest differentiator on this page, ranging from 0% to 5%. It applies to everything you load, whether you spend it or not.
2. FX markup. Charged when you spend in a currency other than the card's base. Typically 0% to 2%. Irrelevant if you only spend at home, decisive if you travel.
3. Per-transaction fee. A flat charge per tap. Rare, but brutal on small purchases: $0.30 on a $3 coffee is 10%. Kardpay charges this on attempts, so a declined transaction still costs you.
4. ATM fee. Usually around 2% after a free monthly allowance, plus whatever the machine's operator adds. Cash is the most expensive way to use any of these cards.
5. The spread. The one nobody quotes. If a provider converts your USDC to USD at a rate a fraction below the market rate, that is a fee whatever the fee schedule says. KAST's explicit 1:1 stablecoin conversion is a real commitment precisely because most providers do not make it.

What $1,000 actually costs

Percentages are not comparable until you run them. Here is the same realistic scenario through every card: load $1,000 of stablecoins, spend all of it across roughly twenty transactions, half at home in the card's base currency and half abroad in another. No ATM withdrawals, no cashback assumed.

Two honest caveats on those figures. KAST's foreign-exchange treatment varies by plan tier and is not published as a single clean number, so its row is a range rather than a point. And Gnosis Pay's €30 is annual, so at low spend it is proportionally worse: at $100 a month it is an effective 2.7%, which is worse than RedotPay. Fixed fees favour heavy users, percentage fees favour light ones.

The full comparison

KAST

The strongest custodial economics on this page. KAST converts stablecoin balances to spendable USD at 1:1 with no spread and no top-up fee, which removes the largest single cost most people pay without noticing. Coverage is quoted at over 170 countries, so it is a realistic option in places where the European-focused cards simply do not work.

The reason to be careful is the reward programme. KAST's headline rates run in seasons, and a seasonal rate is a marketing budget rather than a product feature. Judge it on the 0% top-up and treat any rewards as a bonus that may not survive the next season. The paid membership tiers mostly buy nicer physical cards and higher limits, which is a fair trade if you need either and dead money if you do not.

Gnosis Pay

The only mainstream card where you keep the keys. It spends directly from a Safe smart contract wallet, so there is no top-up step, no intermediary balance, and no withdrawal process to go through if you change your mind. Zero conversion fee and zero FX markup make it the cheapest card here for anyone spending more than a few hundred a month.

It is also the card most obviously constrained. Coverage is the EU, EEA, UK, Argentina and Brazil. There is a €30 annual fee for Safe maintenance. Cashback runs from 1% to 5% but is paid in GNO and scales with how much GNO you hold, which is the same token-position trade discussed below. Average transaction sizes on Gnosis Pay run well under $200 while the off-ramp-style cards average several hundred, which tells you what it is genuinely good at: everyday spending, not moving a large balance out.

RedotPay

The default answer for most people, and the one that works when others do not. It has grown into one of the largest programmes by volume, with millions of registered users and billions in annualised payment volume, and that scale shows up as reliability: the card works, in more places, with fewer surprises than smaller programmes.

You pay for it. A 1% conversion fee on load and 1.2% on non-base-currency spend puts it firmly mid-table, and the 2% ATM fee makes cash withdrawal something to avoid. Card issuance costs $10 for virtual and $100 for physical, which is a real amount for a piece of plastic. The average transaction size across the programme is over $800, which tells you how it is really used: not as a daily spending card but as an off-ramp that happens to work at a payment terminal.

Bybit Card

If your money already sits on Bybit, this removes a withdrawal step and a network fee, which is worth something real. Fees are unremarkable but not predatory: 0.9% crypto conversion, a region-dependent FX fee around 2%, and ATM withdrawals free up to 100 USD or EUR a month before 2% applies. Coverage is the narrow part, limited to the EEA, Switzerland, Australia and a handful of regional programmes.

The 10% cashback headline is doing a lot of work in Bybit's marketing and almost none in practice, because reaching it requires holding a substantial BIT or MNT position. See the section below on why that is a worse deal than it looks.

Ether.fi Cash

The most interesting product here and the one most likely to be misused. It is self-custodial, charges nothing on USD transactions and 1% on non-USD, pays up to 3% cashback in wETH, and works in a broad set of markets including the US. On the published fees alone it is close to the cheapest option on the page.

What it actually does is lend against your collateral. You are not spending your ETH, you are borrowing dollars secured on it, which means there is a loan accruing interest and a collateral ratio that can be breached. In a calm market this is elegant. In a 40% drawdown it is how people get liquidated while buying groceries. If you use it, keep the borrowed amount far below the limit, and read how liquidation works first if any of that is unfamiliar.

Nexo Card

A dual card: it can spend your balance directly or borrow against it, switchable in the app. That flexibility is the product. Fees are competitive in debit mode with no top-up charge, FX and ATM are free up to monthly allowances, and cashback reaches 2% depending on your loyalty tier, which is itself a function of how much NEXO you hold.

Coverage is the EEA and selected markets. In credit mode the same warning as Ether.fi Cash applies with the added consideration that Nexo is a centralised lender, so you are taking counterparty risk on the company as well as market risk on the collateral.

Kardpay

A virtual-only Mastercard funded with USDC, with light onboarding that asks for little more than a phone number and country. That low friction is genuinely useful if you need a working card quickly, and the gift-card and eSIM features have real utility for travel.

The economics are poor and the disclosure is worse. The top-up fee is 4%. There is a $0.25 charge per transaction attempt, which means a declined payment costs you money, and subscription services that retry failed charges can cost you repeatedly for nothing. The advertised 8% cashback requires staking the platform's own KDY token. Most seriously, the terms do not name the licensed issuer or the payment processors behind the programme, so if it stops operating it is not clear which regulated entity, if any, holds your balance.

SolCard

Solana-native funding and a fast signup, and that is the case for it. The 5% top-up fee is the highest on this page by a wide margin, and it is joined by a $0.30 charge on every transaction. Loading $1,000 costs $50 before you have bought anything. At twenty transactions a month you add another $6.

There is no cashback rate that repairs this. A 5% up-front cost against tiered cashback in low single digits is negative on arrival, and it stays negative however long you hold the card. If the appeal is spending from Solana specifically, funding a KAST or RedotPay balance with USDC bridged from Solana costs a fraction of the difference.

Cashback is not a discount, it is a yield on a position you were talked into

Every high cashback rate on this page is gated behind holding the provider's own token. Bybit's 10% needs BIT or MNT, Gnosis Pay's top tier needs GNO, Kardpay's 8% needs KDY, Nexo's tiers need NEXO. This is not cashback in the sense a bank means it. It is a rebate paid for taking an unhedged position in an illiquid asset, and it should be priced that way.

Work it through. Suppose the top tier needs $10,000 of the token and you spend $2,000 a month. At 10% you earn $200 a month, or $2,400 a year, which looks excellent. A 30% drawdown on that $10,000 position, entirely ordinary for an exchange token over a year, costs you $3,000. You have paid $600 for the privilege of earning cashback, and you are still holding the position.

The correct test is simple: would you buy and hold $10,000 of that token if the card did not exist? If yes, take the cashback, it is free. If no, the cashback is not compensating you for the risk, and you should compare cards on their base rates instead. That is why the verdict at the top of this page ranks on fees and ignores reward rates entirely.

Every tap is probably a taxable disposal

In most jurisdictions, converting crypto to fiat is a disposal, and the tax authority does not care whether that happened on an exchange or at a supermarket till. A debit-style crypto card can therefore generate a separate capital gain or loss calculation for every single transaction, which is how people end up with four hundred taxable events and a coffee habit.

There is a clean practical answer. Convert to stablecoins deliberately, in one transaction, and recognise the gain then. Fund the card from that stablecoin balance. Each subsequent card transaction now disposes of a stablecoin whose value has barely moved, so the gain is approximately zero and the record-keeping is trivial. This is a better reason to prefer stablecoin-funded cards than any fee on this page.

Credit-mode cards avoid the disposal entirely, because borrowing is not a sale. That is a real advantage and it is also exactly why the leverage warning above matters: the tax treatment is the incentive that persuades people to take a position they would otherwise decline. For the general principles see crypto taxes 101, and for the software that reconciles it, our comparison of crypto tax tools. None of this is tax advice and rules vary enormously by country.

What happens if the provider disappears

Custodial cards hold your money. That is the deal, and it is usually fine, but it is worth being explicit about what "usually" is carrying. If the programme is suspended, your balance is a claim against a company, and how good that claim is depends on who actually issues the card and under which regulator.

The larger programmes name their issuing partners and operate under recognisable licences. The light-KYC end of the market frequently does not, which is the specific reason Kardpay is flagged above rather than merely priced. If a provider's terms do not tell you which entity holds the funds, treat the balance as a spending float sized to what you would tolerate losing, and top it up weekly rather than parking a quarter's expenses on it.

Self-custodial cards sidestep this entirely. If Gnosis Pay stopped operating tomorrow, the money is in your Safe and you still control it. That property is worth more than a percentage point of cashback, and it is the strongest argument for the category.

What we could not verify

Three gaps, stated rather than papered over.

• KAST's FX treatment by plan tier. The 0% top-up and 1:1 stablecoin conversion are clearly published. The foreign-exchange handling on the entry plan is not stated as a single figure, which is why its row above is a range.
• Regional fee variation. Bybit publishes different FX terms per regional programme, and several providers vary ATM allowances by country. The figures here reflect the commonly published schedules, not every jurisdiction.
• Who issues the light-KYC cards. For Kardpay in particular, the issuing bank and processors are not identified in the public terms. We could not establish them from public sources, and that absence is itself the finding.

Choosing, in one paragraph

If you are in the EEA or UK and spend more than roughly $150 a month, use Gnosis Pay: it is the cheapest and you keep the keys. Outside that footprint, use KAST for the 0% top-up, and RedotPay if KAST does not work where you are or you want the most-tested option. Use Bybit Card only if your balance already lives on Bybit. Consider Ether.fi Cash only if you genuinely want a collateralised loan and understand what that means. Skip SolCard and Kardpay unless the specific convenience outweighs a 4% to 5% entry cost, which for most people it will not.