There is a quiet shift happening at the edges of crypto. It is not the next narrative cycle, and it does not show up in the price charts. It is the slow construction of a parallel financial layer — vaults that yield, stablecoins that move, cards that spend — that is finally complete enough to actually live inside.
For most of the last decade, people who held crypto did one of two things with it. They bought and waited, treating their wallet as a savings account they could not really touch. Or they traded, treating it as a casino. The third option — using their holdings as money, the way people use bank balances — was technically possible but economically absurd. The friction was too high, the rails too few, and the experience too brittle. You did not pay rent in ETH. You sold ETH on an exchange, withdrew dollars to your bank, and paid rent from your bank.
That is no longer the case. The friction has come down enough that a small minority of crypto holders have started running their day-to-day finances through a stack that does not touch a bank at any point. Their salary lands in stablecoins. Their savings sit in a yield-bearing vault. Their card draws against that yield in real time. The bank, for them, is a vestigial appendage — useful for the occasional wire, irrelevant for everything else.
This guide is for people who want to assemble that setup. It walks through the four layers of the stack, the tradeoffs at each layer, and the specific tools that currently work in 2026.
What changed
To understand why this is suddenly viable, it is worth being precise about what changed. Three things happened roughly in parallel over the last twenty-four months.
The first was the maturation of liquid restaking and yield infrastructure. Five years ago, earning yield on ETH meant either running validators yourself or trusting a custodial staking service. Today, restaking protocols and curated vaults let you keep self-custody of your assets while earning across a portfolio of strategies. The yield is real, the audits are serious, and the abstractions are thin enough that experienced users can actually reason about what their capital is doing.
The second was the arrival of stablecoin rails that work. Stablecoins crossed somewhere north of 30% of crypto-on-chain transaction volume in 2025, and the supporting infrastructure — bridges, fiat onramps and offramps, payment processors — finally became reliable enough to use without checking the news beforehand. EUR-denominated stablecoins, in particular, eliminated the FX friction that had previously made on-chain spending a North American-only proposition.
The third was the quiet emergence of self-custodial cards. The earlier generation of "crypto cards" — Coinbase's, Crypto.com's, BlockFi's — were custodial, which meant they were really just bank accounts with an extra translation step. The newer generation, including ether.fi Cash, Gnosis Pay, and Monerium, lets you spend without giving up custody of your assets. They draw against your wallet at the moment of purchase. The card is a payment processor; it is not, in any meaningful sense, your money's home.
These three things — the savings layer, the yield layer, the rail and the spending edge — are what we mean by the on-chain spender's stack. Each layer can be assembled independently, and each one works on its own. But they only become a real financial life when they are connected.
Layer 1: The savings layer
Where does the money actually sit?
The honest answer for most readers is: in a self-custodial wallet, in some combination of ETH, BTC (or wrapped variants), and at least one stablecoin. The choice of wallet matters less than people think — Rabby, Frame, Metamask, Safe, and the various smart wallets all work — but the choice of what to hold matters quite a lot.
The naive version of this is to hold a single asset and try to spend out of it. This works badly. Your spending is denominated in fiat (rent is in EUR, groceries are in EUR), but your asset is volatile. A bad week of prices means you spend more of your stack than you intended. A good week means the opposite, but the asymmetry is real: people regret unnecessary selling more than unnecessary holding.
The working version is a barbell. A core position in a long-term asset (ETH or BTC), a working balance in stablecoins (USDC, USDT, or one of the EUR-denominated variants like EURC), and a buffer for emergencies. The proportions depend on conviction, but a useful starting point is something like 70% long-term, 25% stablecoin working balance, 5% buffer. The stablecoin balance is what your card actually draws against in normal months. The long-term position is what you do not touch unless you have to.
This is also the layer where it matters that you control your keys. Self-custody is not philosophical posturing here; it is what makes the rest of the stack possible. The moment your assets sit on an exchange or with a custodian, the bank-replacement architecture collapses back into a regular bank with extra steps.
Layer 2: The yield layer
Capital that sits idle is capital that is being slowly devalued by inflation. Nobody who lives on bank rails worries about this much, because checking accounts pay near-zero everywhere and have for decades. On-chain, the situation is different: stablecoin yields in the 4–8% range are routinely available from credible protocols, and ETH yields in the 3–5% range are the norm via restaking.
The yield layer is what turns the savings layer from a dormant pile into something that pays for itself. Done well, the yield from your stack covers a meaningful fraction of your card spending. Done very well — usually because you have a large enough stack — it covers all of it. People who have crossed that threshold describe it as the on-chain equivalent of living off the interest, except the interest is denominated in the asset they actually want to hold.
There are roughly three approaches.
The first is staking and restaking. You stake ETH directly through a liquid staking protocol (eETH, stETH, or one of the others) and let it earn the validator yield plus restaking points and rewards. This is the lowest-friction option and works well if your core position is in ETH.
The second is curated vaults. Several protocols now run vaults that route capital across multiple yield sources — lending, restaking, basis trades, RWA exposure — and rebalance automatically. The advantage is hands-off management and diversified risk; the disadvantage is one more layer of trust, since you are relying on the curator's judgment. ether.fi's Liquid vaults, Pendle, and a handful of others occupy this niche. The good ones are audited, transparent about strategies, and have track records you can verify.
The third is direct DeFi. You deploy capital yourself across Aave, Morpho, Pendle, or wherever you understand and trust. This is the most work and the most control. It also tends to be the highest yield for users with the time to manage it actively.
For someone setting up a spending stack from scratch, the first two are sufficient. The third becomes interesting once the stack is large enough that small yield improvements matter in absolute terms.
The thing to avoid is yield chasing. Anything advertising 20%+ on stablecoins is either a Ponzi or an unhedged exposure to something you do not understand. Boring 5–7% on stablecoins, paid by a protocol that has been audited and survived a market cycle, is the correct number to aim for.
Layer 3: The rail
The rail is what moves money between the on-chain world and the world where you actually need to spend it. Until recently, this layer was the worst part of the experience: bridges that broke, exchanges that delayed, fiat ramps that demanded a week of compliance theatre. Most of those problems have been solved well enough to forget about.
What you actually need from the rail layer is two things: a way to get money in (when you receive a salary in fiat, or want to buy more crypto), and a way to get money out (when you want to pay a bill that does not accept crypto). Both flows have become routine.
For inflows: Monerium, Bridge, and a small number of EU-licensed institutions provide IBANs that map directly to on-chain stablecoins. You give your employer or counterparty an IBAN; they wire fiat as normal; the stablecoin appears in your wallet. SEPA settles in minutes most of the time. There is no manual conversion step.
For outflows: the same rails work in reverse, but in practice most people use their card for outflows rather than wires. The card is faster, has no minimum, and works at the point of sale. Wires are reserved for things that cannot be done with a card — large rent payments to a landlord who refuses card, tax payments in some jurisdictions, occasional B2B transfers.
The bridge layer (between L1 and L2s, between chains) is mostly invisible at this point if you use modern wallets. Native bridges, intent-based routing, and the maturation of CCTP for stablecoins have removed most of the cognitive load. You move USDC from Ethereum mainnet to Base or Arbitrum without thinking about it; the wallet does the routing.
Layer 4: The spending edge
The card is where the stack meets the world. It is also the layer where the choice of provider matters most, because it is where every transaction touches a regulated financial network and where the economics of the underlying business are most visible.
The current field has roughly four serious options.
Coinbase Card and similar custodial offerings are the easiest and the worst. They are functionally a bank account at Coinbase that happens to debit your crypto when you spend. You give up custody, you accept Coinbase's business risk, and you usually get worse FX and fewer benefits than a regular bank card. They make sense for people who already keep substantial assets at Coinbase and do not want to manage a wallet, and that is roughly the limit of their use case.
Crypto.com Card is a marketing program with a card attached. The cashback rates are aggressive, the staking requirements are aggressive in the other direction, and the entire experience is built around persuading you to hold large amounts of CRO. This is fine if you want that exposure, ruinous if you do not.
Gnosis Pay is a serious self-custodial card that lives natively on Gnosis Chain. It works well in the EU, has a minimal interface, and is favored by users who want the cleanest possible architecture and do not mind a smaller universe of supported networks.
ether.fi Cash is the one we use and recommend for most readers. It is non-custodial, supports ETH, BTC, and stablecoins as collateral, works across the EU and most non-US jurisdictions, and ties into the broader ether.fi ecosystem of vaults and Liquid products if you want them. The card itself is a Visa, accepted everywhere Visa is accepted. The interface is clean enough that the card feels boring to use, which is exactly what you want from a card.
The economics work like this. You provision a Cash account. You fund it with ETH, BTC, USDC, or any of several supported assets. When you spend, the card draws against that balance — selling exactly as much as needed for the purchase, at the moment of purchase, at the spot rate. Your stack stays in the asset you chose to hold; only the spent portion converts.
Get ether.fi Cash →
Putting it together
A working setup, end to end, looks roughly like this.
Salary lands in fiat. It hits a Monerium IBAN, becomes EURe (or USDC, depending on currency), and arrives in the wallet within minutes. The wallet routes 70% of it into a long-term position — say, eETH staking — and parks 25% in a curated stablecoin vault yielding 5–7%. The remaining 5% sits in a working balance for the month's spending.
When the card spends, it draws against the working balance. As the working balance depletes, the wallet tops it up from the stablecoin vault, automatically or manually depending on preference. Over the course of a typical month, the yield generated by the stablecoin vault and the eETH position covers somewhere between 30% and 70% of the card spending, depending on the size of the stack. The rest is funded by drawdowns of the principal, which the next month's salary replenishes.
The point is not that the math is dramatically better than a high-yield bank account, although it is. The point is that the entire flow happens in one stack, with one set of credentials, on one set of rails, without ever crossing a bank's permission boundary. The portability, the privacy, the lack of counterparty trust — these are the real wins. The yield is a bonus.
What does not work yet
It would be dishonest to suggest the stack is finished. It is not.
Tax remains the worst part of the experience. Every card transaction is technically a disposal event in most jurisdictions, which means most users end up needing software (Koinly, CoinTracking, Crypto Tax Calculator, ether.fi's own reports) to keep up. This is improving, and some EU jurisdictions have begun treating qualifying transactions more favorably, but it is still meaningfully more work than a bank card.
Regulatory access is uneven. The cards we cover work cleanly in most of the EU, the UK, and much of LATAM and Asia. The US is a patchwork; some products do not serve US persons at all, and others have meaningful restrictions. If you live in the US, your options are narrower, and you should verify support for your specific state before committing to a setup.
Customer support, when something goes wrong, is asymmetric. A traditional bank has a call center. An on-chain card has a Discord and a documentation site. For 95% of issues this is fine; for the remaining 5%, it is not. Plan accordingly: keep a backup bank account around for emergencies, and do not run your entire financial life through the stack until you have lived with it for several months.
These are not deal-breakers. They are the kinds of friction that exist in any new infrastructure, and they are getting smaller every quarter. But they are real, and people who switch over without acknowledging them tend to switch back within a year.
The verdict
The interesting thing about all of this is not that it is now possible to live without a bank. It has been possible, in a half-broken way, for years. The interesting thing is that the experience has finally become boring enough to recommend.
The stack works. The card spends. The yield compounds. The transitions between layers are mostly invisible. For the first time, you can run an ordinary financial life on chain and the inconvenience is roughly comparable to running it through a regular bank — which is to say, considerable, but no longer disqualifying.
Whether you should make the switch depends on what you value. If you hold a substantial amount of crypto and dislike the friction of selling it to spend it, the stack pays for itself within a few months. If you live in a jurisdiction where banking is unreliable or expensive, the calculation is even easier. If you are mostly in fiat and only dabble in crypto, the stack is overkill and a regular bank account remains the correct choice.
For everyone in between — and that is a much larger group than it was three years ago — the on-chain spender's stack is now a real option. The infrastructure is here. The tools are good enough. The remaining work is figuring out which configuration fits your particular financial life.
The next several pieces in this publication will walk through specific configurations, specific tools, and specific edge cases. We will keep updating this guide as the stack evolves.