Two years ago, "self-custodial crypto card" was a category with one serious entrant and several theoretical ones. Today it is a genuine market with at least six products worth evaluating, meaningfully different architectures, and enough real usage data to make informed comparisons. This guide maps the current field.
We define self-custodial strictly: the card issuer does not hold your private keys, and your assets do not sit in a custodial account between top-ups. When you spend, the card draws against a wallet you control — either by converting from your on-chain balance at the point of sale, or by drawing against a collateral position you have provisioned. If the card company shuts down tomorrow, your assets remain on-chain.
This definition excludes a large number of products that market themselves with self-custody language but are functionally custodial: cards where you load a balance onto an exchange-managed wallet, cards where the issuer holds funds in a pooled account, and cards where "self-custody" refers only to the wallet app interface rather than actual key ownership. We name those products where relevant but do not include them in the main evaluations.
Why custody model is the only filter that matters first
The instinct when comparing cards is to look at cashback rates first. This is the wrong order of operations. The cashback rate is irrelevant if the card company fails and your balance disappears with it. The custody question should be answered before any other.
The recent history of crypto lending and card products is instructive. BlockFi, Celsius, Voyager, and FTX all offered attractive yield and reward rates on custodial products. Users who treated those rates as the primary decision variable lost capital. Users who refused to hold assets in custodial accounts, even at lower yield, did not.
The lesson is not that custodial products are always bad. It is that the yield or cashback premium offered by custodial products is compensation for counterparty risk — and you should decide whether you want that trade-off deliberately, not by default.
With that framing established, here is the current field of genuinely self-custodial cards.
The cards
The only card in this field that earns yield on your balance while you are not spending. Funds sit in ether.fi's Liquid Vaults earning restaking or lending yield until drawn at the point of sale. Two spending modes: Direct (stablecoin to fiat, disposal event) and Borrow (ETH collateral, not a disposal in most jurisdictions). Cashback in wETH — a real asset, not a platform token. Visa Signature benefits including airport lounge access on all tiers. Geographic gaps in Netherlands, India, most US states, UK until September 2026. The strongest all-round product in this category for ETH holders in supported jurisdictions.
The most architecturally rigorous card in the field. Every transaction settles on Gnosis Chain — a public blockchain you can verify independently. Your funds sit in a Safe smart contract wallet under your sole control. No other card currently delivers on-chain settlement as a verifiable technical fact at scale. The trade-offs are real: EEA-only coverage, no Apple Pay, thin asset list (EURe, USDCe, xDAI, GNO), and cashback paid in GNO rather than a more liquid asset. The interim cashback programme (running until June 2026) changes tier structures; check current rates before applying. Right for EEA users who hold GNO and want on-chain settlement as a principle, not just a marketing claim.
The lowest-friction entry point into self-custodial card spending. If you already use MetaMask, setup requires only a KYC check — no new wallet, no separate account, no bridge step. Operates on Linea (a zkEVM L2), which keeps gas fees low and settlement fast. Supported tokens include USDC, USDT, wETH, EURe, and GBPe. The 1% cashback rate is the weakest in this comparison, and the 1% cross-border fee means international spending erodes rewards entirely. The geographic coverage — US, EEA, UK, and Switzerland — is the broadest of any self-custodial card currently available. The right choice for MetaMask users who want to start spending from their existing wallet without changing their setup, and for US-based readers who find ether.fi Cash unavailable in their state.
The newest entrant in this space worth taking seriously. Bleap uses MPC (multi-party computation) key management rather than a smart contract wallet — a different self-custody architecture that removes the single-point-of-failure risk of seed phrases while maintaining user key control. The headline cashback figures (up to 20% on gaming, streaming, and AI subscriptions; 3% on rides and delivery; 2% on groceries) are category-leading if the merchant coverage matches your spending patterns. The 0% FX fee is a genuine advantage. Currently EEA and Switzerland only, which limits its audience. Track record is short — the product launched in late 2025 — and we would recommend waiting for a longer operational history before treating it as a primary card. Keep it on the watchlist.
COCA offers self-custody with tiered cashback up to 8% at the Elite level (which requires staking 30,000 COCA tokens). The architecture is non-custodial and the 0% FX fee is a genuine differentiator. The core issue is the same as Crypto.com's: cashback paid in the platform's own native token creates circular demand that makes the headline rate misleading. 8% cashback in COCA is worth 8% cashback only if COCA holds its value. The token launched in 2024 and has a short price history. Approach with the same caution you would apply to any product where the reward asset is the same entity issuing the card. The Starter tier at 1% with no staking requirement is reasonable; the higher tiers require real conviction in the COCA token.
Products we evaluated and excluded
Several products market self-custody credentials that do not hold up to scrutiny.
Crypto.com Visa is custodial — Crypto.com holds your funds. Covered in our previous comparison piece.
Coinbase Card is a custodial exchange card. Coinbase holds your assets.
Nexo Card is a hybrid: Nexo holds custody while you borrow against crypto collateral. The custody risk is Nexo's solvency, not your key management.
Wirex and Brighty are fiat-crypto hybrid apps with card features. Neither is self-custodial in the strict sense used here.
KAST is multi-chain and Solana-native, with self-custody credentials that appear genuine — but the track record is very short and the primary user base is Solana-native, which puts it outside the scope of this publication's ETH-focused readership.
How to choose
Three questions resolve most situations.
Where do you live? This eliminates most options immediately. US users outside the blocked states have MetaMask Card and ether.fi Cash (check your state). EEA users have all five options evaluated above. UK users currently have MetaMask Card; ether.fi Cash from September 2026.
What do you hold? ETH-heavy portfolios point toward ether.fi Cash — the Borrow Mode and wETH cashback are built for this. Stablecoin-heavy portfolios work well with any of the five. GNO holders have a specific reason to consider Gnosis Pay for the tiered cashback. Existing MetaMask users who hold USDC or wETH have the shortest path to the MetaMask Card.
How do you spend? High monthly spend in EUR with no need for cash → ether.fi Cash or Gnosis Pay. Heavy spending on streaming and subscription services → Bleap's merchant-specific cashback is worth evaluating. Multi-currency travel spending → MetaMask Card's geographic breadth or Gnosis Pay's 0% FX rate.
For the majority of readers of this publication — ETH holders in the EU — the answer is ether.fi Cash. The yield-while-holding feature, wETH cashback, and Borrow Mode tax efficiency are a package the rest of the field does not match. MetaMask Card is the right fallback for US users and the right starter card for people who want the lowest possible setup friction. Bleap is the one to watch for European users if it builds the operational track record its architecture deserves.
Get ether.fi Cash →
What changes this year
The self-custodial card space is moving faster than any other part of crypto consumer products. Three developments worth tracking over the next twelve months:
ether.fi Cash's UK launch in September 2026 opens the product to one of the largest crypto-holding populations in Europe. The FCA regulatory framework will also set a precedent for how other non-custodial card issuers approach UK authorisation.
Gnosis Pay's post-June 2026 cashback programme will likely change the GNO staking tier structure significantly. The current interim programme was introduced as a bridge between the original design and the next iteration. Watch the updated terms before committing to a large GNO position for cashback purposes.
MetaMask Card's geographic expansion is ongoing. Consensys has indicated it intends to extend coverage to Latin America and parts of Asia through 2026; the current US and EEA footprint is the starting point, not the ceiling.
We will update this guide quarterly as the landscape changes. If a product materially improves or a new entrant earns inclusion, it will appear here.