Ethena Pay launches a neobank where the deposit is USDe

Ethena has launched Ethena Pay, a self-custodial money app built on Avalanche that dresses its synthetic dollar, USDe, in the full costume of a neobank: savings yield, global transfers, a card that runs on Visa's network, and Apple Pay support. The beta opens with 400 users across 49 countries, and the pitch is the oldest one in banking, paid interest, made by a company that is pointedly not a bank.

The product, in detail

The structure is tiered, and the tiers are the business model. Standard users earn yield on up to 5,000 dollars of USDe and 4 percent cashback on card spend. Pro lifts the yield to 6 percent, the earning cap to 15,000 dollars and cashback to 5 percent; VIP keeps the same headline rates with a 50,000 dollar cap. Cashback pays out not in dollars but in AVAX, the native asset of the chain the app settles on.

Upgrading is where Ethena's own token enters the loop. Pro requires locking 2,000 dollars of ENA or referring ten users; VIP requires 10,000 dollars of ENA or fifty referrals. The design is circular by intention: the app drives demand for ENA, ENA locks deepen user commitment, and referrals compound distribution. It is the growth machinery of a crypto protocol bolted onto the interface of a banking app.

Geography tells the regulatory story. Forty-nine countries at launch, with the United States, European Union, United Kingdom and Canada explicitly listed as later markets. The jurisdictions with the most developed stablecoin rulebooks are the ones the launch avoids, and that is not an accident of sequencing.

What USDe actually is

None of this makes sense without understanding the asset underneath, because USDe is not a stablecoin in the sense the GENIUS Act means the word. A payment stablecoin holds dollars and Treasury bills; USDe holds a hedged trade. Ethena takes collateral, principally staked ether and bitcoin, and shorts perpetual futures against it in equal size, so the portfolio's dollar value holds steady regardless of price moves. The yield comes from two engines: staking rewards on the collateral, and the funding rate that perpetual futures markets have historically paid to the short side.

That design scaled from nothing to one of the largest dollar assets in crypto, around 4 to 5 billion dollars in circulation, precisely because the yield is real and native rather than passed through from a bank. It also carries risks a reserve audit cannot capture: funding rates go negative in bear markets, turning the income engine into a cost; the hedges live on exchanges, adding counterparty exposure; and redemption depends on unwinding trades rather than selling T-bills. Ethena maintains a reserve fund against funding inversions, and the design has survived stress it was widely predicted not to. But the honest description of USDe is a market-neutral hedge fund share that behaves like a dollar, and Ethena Pay is asking consumers to treat that share as a current account.

Why Avalanche

The chain choice is more than logistics. Ethena Pay runs its settlement on Avalanche rails, and pays its cashback in AVAX, which aligns the app's growth with the network whose blockspace it consumes. For Avalanche, a consumer money app is exactly the workload its architecture has courted: cheap, fast finality for payment-sized transfers, without asking users to know or care which chain they are on. For Ethena, a single settlement environment simplifies the hardest part of self-custodial UX, predictable fees and instant confirmation, and the cashback token doubles as an acquisition subsidy someone else's ecosystem fund is motivated to support.

The precedent being invoked, consciously or not, is the super-app pattern: pick one set of rails, hide them completely, and let the economics of the rails sponsor the growth of the app.

The regulatory fence, and the gap in it

The timing of this launch, the same week Treasury opened comments on GENIUS Act licensing and Singapore proposed banning interest on regulated stablecoins, is the whole story told twice.

Both regimes converge on the same bargain: licensed stablecoins get legal certainty and in exchange pay holders nothing. The prohibition is deliberate. It protects bank deposits from a perfect substitute that out-yields them, and it keeps payment instruments boring. Every major framework, American, European, Singaporean, draws the same line.

USDe lives on the far side of that line by construction. It is not a payment stablecoin under the Act, claims no licence, and is therefore bound by no yield prohibition. The fence built to keep regulated stablecoins from paying interest has, from another angle, granted the unregulated synthetic dollar a monopoly on the category's most wanted feature. Ethena Pay is the first consumer product built explicitly to exploit that monopoly, and its launch map, everywhere except the fenced jurisdictions, is the compliance strategy drawn as geography.

The open question is duration. Regulators noticed the same gap; Singapore's consultation this week discusses yield-adjacent structures, and the American service-provider rules landing through 2028 will decide whether US-facing platforms can list synthetic dollars at all. The product works today. The perimeter is still moving.

Self-custody meets the card network

Technically, the most interesting claim is the pairing of self-custody with a Visa card, two systems with opposite assumptions. A card authorisation must clear in under a second and can be reversed for ninety days; a self-custodial balance moves only when the user signs, and never moves back. Making them meet requires choreography: spending balances pre-authorised into a card float, conversion executed at tap time, disputes absorbed by the programme rather than unwound on chain. The user keeps custody of savings; the spending leg necessarily passes through regulated intermediaries, the BIN sponsor and processor that every card programme drags behind it, plus Apple's wallet rules on the devices where Apple Pay works.

That compromise, sovereignty for the vault, delegation for the till, is quietly becoming the standard architecture of crypto consumer finance. Ethena Pay is its most aggressive expression yet, because the vault also pays 6 percent.

The yield arithmetic, honestly

Six percent on a dollar balance demands an answer to the oldest question in finance: paid from what? For USDe the components are legible. Staked collateral earns consensus rewards, historically low single digits. The short perpetual position earns funding, which in strong crypto markets has run high single to double digits annualised, and in weak ones has gone negative for stretches. Blend them, subtract Ethena's cut and the reserve fund's accrual, and 6 percent on capped balances is plausible in favourable regimes and subsidised in hostile ones, with the caps, 5,000 to 50,000 dollars, sized so that subsidy remains affordable.

Compare the alternatives a user actually has. Licensed stablecoins pay zero by law. Treasury-backed tokenised funds pay the bill rate to qualifying investors, with transfer restrictions. Bank deposits in the launch countries pay local rates in local currency, with local currency risk attached. Against that menu, USDe's offer is genuinely differentiated, which is exactly why the caps and the geography, and not the marketing, are the parts to watch.

The competitive field it lands in

The launch window is crowded from both directions. The same week: Visa and Bridge expanding stablecoin-funded cards toward a hundred countries, on fully regulated rails, paying no yield. Twenty-one banks committing to a consortium issuer whose token will, by statute, pay nothing. Singapore proposing to ban interest on regulated tokens outright. The entire licensed world is converging on zero, at scale, with distribution Ethena cannot match.

Ethena's bet is that zero is beatable, that a measurable slice of global dollar demand will accept legibility-instead-of-licensing in exchange for six points of yield. The banks' bet is that yield without a licence has a short shelf life. Both can be right for years; the fence between them is where every interesting product in this category is now being built.

The history it is arguing with

Yield-plus-card has been tried, and the archive is cautionary. BlockFi built a beloved rewards card on top of yield accounts and followed the 2022 credit cascade into bankruptcy; Celsius promised bank-beating rates and dissolved into litigation. Those failures shared a structure: pooled customer assets, rehypothecated into opaque credit risk. Ethena's counterargument is architectural, the user holds the keys, the yield source is a visible market position rather than a loan book, and the risk, whatever it is, is at least legible on chain.

It is a genuinely better structure. It is not a risk-free one, and the difference will matter the first time funding rates stay negative for a quarter. A 6 percent current account is never free; the question Ethena Pay puts to 49 countries of users is whether they can price what they are being paid to hold.

How the beta will actually be judged

Four hundred users is a number chosen for control, not for data volume, and the metrics that matter in the beta are operational rather than growth-shaped. Conversion latency at the card edge: whether tap-to-approval stays invisible when the funding leg is a signed, self-custodial draw-down. Dispute mechanics: the first chargeback that has to be honoured against an irreversible on-chain spend will define the programme's real cost structure. Withdrawal behaviour: whether users treat the yield balance as savings, parked and stable, or as a hot wallet, in constant motion, decides how the hedging book behaves underneath. And support load: self-custody moves the recovery burden onto the user, and the beta's ticket queue will reveal whether 49 countries of normal people can actually carry it.

None of those show up in a dashboard of signups. All of them decide whether the caps ever rise.

The distribution problem yield solves

Strip the mechanism away and Ethena Pay is solving the same problem as every launch this week: how a dollar token acquires users it does not already have. The banks answer with trust, the networks with acceptance, Samsung with default placement. Ethena's answer is price. Six percent is not a feature, it is customer acquisition cost, paid continuously, from the protocol's own economics rather than a venture budget, and referral-gated tiers turn the subsidy into a growth loop. It is the most honest acquisition model in the category, the cost sits in the open, and the most fragile, because it survives only while the basis trade pays for it.

What to watch

Three dials. The caps: 5,000 to 50,000 dollar yield limits keep the product small and the risk retail-sized; raising them is the signal that Ethena is chasing deposits at scale. The funding regime: a sustained negative-rate stretch is the stress test that matters, and the reserve fund's behaviour through it will settle the safety argument one way or the other. And the fenced markets: whether the US, EU and UK launches arrive as licensed products, restructured to fit the perimeter, or never arrive at all, will say which side of the regulatory fence the next generation of crypto banking chooses to build on.

The category it is trying to name

Ethena calls the product a money app, and the vocabulary struggle is genuinely informative. It is not a bank: no deposits, no insurance, no licence. It is not an exchange: nothing is traded. It is not quite a wallet: the point is the yield and the card, not the keys. The nearest honest description is a self-custodial cash-management account denominated in a synthetic dollar, a phrase no marketer will ever print, which is why the industry will end up calling all of these things neobanks and regulators will end up insisting they are not.

That naming fight has a history of mattering. The last generation of yield platforms marketed themselves in the language of banking, interest, accounts, deposits, and the enforcement actions that followed leaned heavily on exactly that vocabulary. Ethena's materials are noticeably more careful, and the carefulness is itself a market signal: the products on the far side of the regulatory fence have learned to stop borrowing the regulated world's words.

Banks spent this week announcing their entry into stablecoins. Ethena spent it demonstrating the counter-thesis: that the most interesting dollar products will be built by whoever is not required to behave like a bank.