The two fees that decide whether a crypto card is worth using

Most crypto card reviews open with the cashback number. Eight percent on Crypto.com Obsidian. Five percent on Nexo Platinum. Two percent on Coinbase. Those numbers are true, but they are also the part that matters least once you spend a few months with an actual card.

The two fees that decide whether a card is worth holding are the FX spread and the stablecoin conversion markup. Neither gets much attention on issuer marketing pages. Both are larger than the cashback tier on almost every card I have compared at sweepbase.net.

The FX spread is not what the issuer advertises

Most issuers advertise a 0 to 1 percent foreign transaction fee. That is the fee you pay for using the card in a currency different from the account currency. For a card denominated in USD used in EUR, the advertised fee might be 0.5 percent.

What the issuer does not advertise is the FX rate they use. Visa and Mastercard network rates are public and close to the interbank rate. The issuer rate is usually 0.5 to 1 percent worse than the network rate. Add the advertised fee on top, and the real cost of a non-USD purchase becomes 1 to 2 percent for most custodial cards.

Self-custody cards like Gnosis Pay and 1inch Card have a different structure. They charge a crypto-to-fiat conversion at the point of sale, and the spread there is visible on chain if you want to trace it. Typical spreads I have measured run 0.6 to 1.2 percent depending on the backend liquidity source.

The stablecoin conversion markup nobody talks about

If the card is funded with USDC, USDT, or a crypto asset, there is a second fee hidden inside the transaction. The card issuer converts the funding asset to the settlement currency at a rate that includes a spread.

USDC to USD spreads are tight, typically 0.1 percent. USDT to USD is slightly wider, around 0.2 to 0.4 percent. USDC or USDT to EUR runs 0.5 to 1 percent on most cards. Converting ETH or BTC to fiat at point of sale is 1 to 2 percent on the cards I have benchmarked.

Stack that on top of the FX spread and you start seeing why an advertised 2 percent cashback on stablecoin spend does not always beat a 1 percent cashback card with no conversion step. The math is not subtle once you lay it out.

Break-even math at $1500 per month

Here is a rough calculation for a person spending $1500 per month on a crypto card, half of it in a non-account currency.

Card A offers 4 percent BTC cashback and charges 1 percent FX. Monthly fees at $750 non-account spend: $7.50 FX. Monthly cashback: $60 assumed paid in BTC at spot. Net benefit: $52.50 before BTC volatility.

Card B offers 1 percent USDC cashback, no FX fee, and uses a 0.2 percent stablecoin spread. Monthly fees: $1.50 spread on the full $1500. Monthly cashback: $15 in USDC. Net benefit: $13.50 locked in at spot.

Card A looks better on paper. It often is not. Most cards in Card A tier require a staking deposit of $2000 to $10000 to unlock the cashback rate, and cashback is commonly capped at $25 to $50 per month once you do qualify. Layer on BTC drawdown risk between when cashback lands and when you sell it, and the headline 4 percent shrinks fast.

Why I built a calculator

I wanted to stop guessing. The tool at sweepbase.net/calculator takes a monthly spend, a split between account and non-account currency, and an optional staking stipulation, then returns a break-even month count for any pair of cards in the database. It is the first thing I reach for when a reader emails me asking which card fits their profile.

The calculator does not try to guess future BTC prices. The cashback side assumes spot value at earning. Readers who think BTC is going to 200k should add that expectation separately. Readers who want to realize cashback in stablecoins to avoid drawdown risk should read the card Terms of Service carefully, because not every BTC-rewarding card lets you convert inside the app without a spread.

What the data says

Across the 139 cards I have catalogued, the median advertised cashback is 2 percent and the median real net benefit at $1500 monthly spend is 0.4 percent. The worst cards have negative real benefit once you account for annual fees and staking opportunity cost. The best performer for a typical European spender is not the card with the biggest headline number.

If you want to check the numbers against a specific card, the full comparison view is at sweepbase.net. Every card has a Real Fees section that separates the advertised tier from the realistic monthly cost, so you can run the comparison without the marketing framing getting in the way.

The shortcut readers keep asking for does not exist. You have to do the arithmetic for your own spending profile.