Crypto card cashback explained: the honest math behind the rates

A bank card funds its rewards almost entirely from interchange — the small fee merchants pay per transaction, capped at fractions of a percent in many regions. That is why bank cashback rarely climbs far above 1%. A crypto card issuer has more levers: it also earns a spread when you top up or convert crypto, and, crucially, it can pay rewards in its own token rather than in money.

Paying rewards in a native token changes the economics. Tokens the issuer created cost it far less than their market price, every payout puts the token into more hands, and a generous headline rate doubles as an advertising campaign. During a growth push, the rewards program is effectively the marketing budget: a loud cashback rate can be cheaper than the ads it replaces, because cardholders do the promotion themselves.

The same mechanism makes the rewards fragile. When growth slows, the token price falls or the marketing budget moves on, the program shrinks — rates get cut, tiers get repriced, caps and exclusions appear. This has happened across the industry repeatedly, often with only weeks of notice. It is not a scandal; it is how token-subsidised rewards work. Choose a card you would still keep after the next cut.

Most programs use a ladder: hold or lock a set amount of the issuer’s native token and your cashback moves from the base rate to a higher tier. The lock is often measured in token units or in the fiat value at the moment you stake, and typically runs for a fixed period — commonly around six months — before you can unstake without losing the tier. Details differ per program, so read yours.

Three costs hide inside that ladder. First, price risk: the staked tokens can fall in value while locked, and a modest fall can erase years of extra rewards. Second, opportunity cost: the same money could sit in savings, earn yield or simply stay unspent. Third, downgrade rules: if your tier depends on the stake’s current market value, a falling token can silently demote you to a lower rate while your money stays locked.

Run the numbers before staking. Say the base tier pays 1% and a stake worth $400 unlocks 2%. The extra 1% must first pay back the stake itself: you need $40,000 of annual card spend before the upgrade merely equals the $400 you locked — and that is before any change in the token’s price. If the token halves meanwhile, the upgrade started $200 underwater, which is years of extra cashback at typical spend.

Rewards accrue in the issuer’s token at its price on the day you earn them. By the time you claim, convert or spend them, that price has moved — sometimes up, often down in exactly the periods when programs are being trimmed. A 2% reward that loses a third of its value before you sell was a 1.3% reward with extra steps, and the longer rewards sit uncollected, the bigger that drift.

Decide your policy up front: auto-sell or hold. Some cards can convert rewards to a stablecoin or fiat balance automatically; others accrue the token until you act. The honest test: if you would not buy this token deliberately with your own money, do not hold it accidentally through inertia. And watch the dust — small balances below conversion or withdrawal minimums quietly pile up and are never claimed, which suits the issuer fine.

Taxes depend on where you are resident. Many countries treat card rewards as a non-taxable rebate on your own spending; others treat crypto rewards as income at the value when received, and selling the tokens later can separately create a capital gain or loss. Export your reward history regularly — once a quarter takes minutes — and remember the rules can change when your residence does.

Every program carries a list of merchant categories that earn no rewards, and it usually includes the heaviest lines in a real budget: rent, utilities, government payments and taxes, money transfers, wallet and exchange top-ups, sometimes lotteries and — painfully for travelers — sometimes travel agencies and booking platforms. Classification runs on merchant category codes, and the issuer’s reading of them is final.

Monthly caps do the quiet work on the other end. A cap on reward value means your effective rate falls as your spend rises: once a big month hits the ceiling, everything above it earns zero. Heavy spenders routinely discover their real annual rate is half the headline because two or three expensive months were truncated. Before counting rewards, check how much of your budget is even in the game.

Then come the fees that eat rewards from inside. A card that pays 2% but charges a 1–2% foreign-transaction markup nets you close to nothing on travel spend, which for a nomad is most spend. Add top-up spreads, ATM fees and the odd monthly charge, and a nominally generous card can be net negative. The full fee anatomy deserves its own read.

The formula is short: net rate = (earned rewards − fees − FX markup − expected token slippage) ÷ real spend. Earned rewards means your tier’s rate applied only to eligible categories, truncated at the cap. Fees means every recurring and per-transaction charge you will actually hit. Token slippage is a haircut for the reward token’s tendency to drift between earning and selling — pick a conservative discount rather than assuming today’s price.

Run it three times: for the card at your realistic tier, for the same card without staking, and for a plain low-FX card with no rewards. The spread between those three numbers tells you what the stake is actually buying — often surprisingly little once exclusions and caps are applied to your genuine category mix. If the difference is smaller than the risk on the stake, the answer is obvious.

Fees and rewards are not symmetrical. The FX markup applies with certainty to every foreign-currency transaction; the reward is conditional on tier, category, cap and token price. A 2% reward paired with a 2–3% markup is a guaranteed loss dressed as a perk — you pay the markup on everything and collect the reward on less than everything.

The decision rule is blunt: if the card’s FX markup and fee stack exceed your computed net reward rate, a no-cashback card with a near-zero FX markup wins outright. For most travelers this is true most of the time, because travel spending is almost entirely foreign-currency spending — the markup side of the scale carries nearly your whole budget.

The workable compromise many nomads land on: a low-FX card as the default for foreign spend, and a rewards card only for the home-currency categories where it demonstrably earns — provided its fixed fees don’t erase the gain. Two cards from different issuers is also exactly the backup setup you want anyway.

The pattern repeats across the industry: a program launches generous, grows the user base, then trims — lower rates, higher stakes, new exclusions. So manage the risk instead of predicting it. Stake only an amount you would happily hold as a crypto position anyway, so a devaluation of the program never forces a sale at a bad moment. Harvest rewards on a schedule — monthly or quarterly — instead of letting them accumulate in a token you didn’t choose.

Re-check the terms every quarter: reward rate, exclusions, caps, stake thresholds. Issuers publish changes quietly, and community forums usually notice before the email arrives. And keep cashback out of your travel budget — count it when it lands, never in advance. If a rate cut would change your plans, the program was carrying more weight than it should.

Evaluate the card you would actually hold: your realistic tier, your real category mix, your countries. The advertised ladder-top number is somebody else’s card. Confirm every input on the issuer’s official pages, because programs shift — and the comparison page below tracks current terms so you don’t have to re-collect them yourself.

Then repeat the exercise once a year and whenever you change your base country: availability, tiers and even the exclusion list can differ by region, and a card that was the right answer in one residence can be unavailable — or uncompetitive — in the next. Ten minutes of checking is cheaper than a year of the wrong cashback.