# The stablecoin neobank stack is forkable, and most won't survive it

URL: https://www.thedeepfeed.ai/posts/2026-06-15-stablecoin-neobank-stack-is-forkable/
Category: Business
Published: 2026-06-15
Author: the-deep-feed
Tags: stablecoins, neobanks, fintech, crypto-cards, payments, distribution
Kind: deep

> Strip the logos off the stablecoin neobank boom and most of the apps are the same six vendors in a trench coat. A teardown of 29 operators, the unit economics that doom the skins, and the three escape routes that are left.

## TL;DR

- Strip the logos and most of the **29 stablecoin neobanks** in this teardown are the same app: a wallet, a KYC API, and a **Visa** card from one of **~6 shared vendors**. The card is infrastructure now, not the product.
- The money does not live in the apps. It lives in the vendor columns — **Bridge, Rain, Circle, Visa** — that every skin must rent. **~27%** of gross revenue leaks straight to rails the operator will never own.
- Interchange is a trap. **~60%** of a neobank's gross is **FX + float**, not the **~29%** from card swipes. An app that only earns interchange is structurally unprofitable before it pays an engineer.
- Three escape routes are left: own a **banking charter** (Xapo, Fiat24), own **uncopyable distribution** (MoneyGram, Phantom, Deel), or own a **defensible geographic wedge**. Everyone else is fighting over scraps.
- The honest number is not the **$33T** headline. Real-economy stablecoin payments are **$350–550B/yr**. Big enough for a handful of winners, far too small for 50 identical apps.

A few years ago, launching a bank meant a charter, a card program, a compliance team, a treasury desk, and millions of dollars in upfront capital. **Stablecoins** collapsed that to a weekend. Pick one vendor per layer off a menu, wire them together, slap your logo on a Visa card, and you have a neobank. The supply side noticed. There are now dozens of these apps, and a recurring observation has started to travel across fintech timelines: strip the branding off most of them and you are looking at the same six companies wearing different coats.

That observation is correct, and it is more damning than it sounds. This is a teardown of **29 stablecoin neobanks**, scored on a single axis: what each one actually *owns* versus what it *rents*. Then the unit economics that explain why the rented stack is a death sentence for most of them, and the three narrow paths that are left for the rest.

# The website moment for banking

The cleanest framing of what is happening came from a fintech writer in a widely-shared post this month:

> Banking is following the same path as websites. Twenty years ago, building a website required engineers, servers, and large budgets. Today anyone can launch one in an afternoon. The same thing is happening to financial products.
>
> — [@strato_money](https://x.com/strato_money/status/2065873428331291007), Jun 13, 2026

The piece runs a list that any operator will recognize. Need a card program? Rent it. Need KYC? Plug in an API. Need stablecoin accounts, local bank accounts, payment rails? Available, available, available. The conclusion is the part worth sitting with:

> Launching will become easy. Winning will become hard.
>
> — [@strato_money](https://x.com/strato_money/status/2065873428331291007), Jun 13, 2026

(Worth flagging the author's incentive up front: @strato_money runs a gold-backed stablecoin project, so the "infrastructure is commoditized" thesis is one that happens to favor differentiated-asset issuers. The argument stands on its own merits, but it is not disinterested.)

The website analogy is exact in a way that should worry every founder in the category. When publishing a website got easy, the number of websites exploded and the value did not accrue to the people who published them. It accrued to the platforms underneath — the hosts, the CDNs, the ad networks, the search engine. The cost of assembly went to zero, and the margin migrated one layer down to whoever owned the part you could not skip. Banking is now having that moment, and the layer you cannot skip is not the app. It is the issuer, the orchestrator, and the dollar.

![Seven near-identical bank-card silhouettes in different grey tones cast a single shared shadow, revealing the same object beneath; one card edge marked in red](/post-images/2026-06-15-stablecoin-neobank-stack-is-forkable/same-six-vendors.jpg)

# What the operators actually own

To make "they all own nothing" precise rather than rhetorical, score each operator on six things that confer durable advantage, drawing on the [verified production stacks at neobank.build](https://www.neobank.build/stacks): a card issuer relationship they control, distribution they own, their own chain, their own stablecoin, a regulatory license, and their own on/off ramps. Weight them by how hard each is to copy — a card-issuer principal membership and real distribution are worth the most, a license worth a banking charter tops out the scale. The result sorts 29 operators into four tiers.

| Tier | Score | What it means | Operators |
| --- | --- | --- | --- |
| Chartered | 11–13 | A literal banking license | Xapo Bank, Fiat24 |
| Owners | 7–10 | Durable moat, captures value | Wirex, MoneyGram, Gnosis Pay |
| Hybrids | 3–6 | One real edge, rents the rest | Phantom Cash, Deel, Payy, Hyperbeat, ether.fi Cash, Plasma One |
| Skins | 0–2 | Pure assembly, owns nothing | Altitude, Tria, Kosh, RedotPay, Based, Avici, SecondFi |

Twelve of the twenty-nine land in the bottom tier. These are the pure skins: a Sumsub or Persona for KYC, a Bridge or Iron for stablecoin accounts, a Privy or Turnkey for wallets, a Rain for card issuing, a Visa BIN on top. Tria rents Sumsub, Turnkey, and Third National and competes on chain-abstraction UX. Kosh stitches together Sumsub, Bridge, Iron, Privy, and Rain and competes on an Asia freelancer wedge. RedotPay runs Paybis, Fireblocks, and Visa and competes on an existing user base. None of them owns a single layer of the stack. Their edge, where they have one, is a niche or a geography — not the machine.

At the top sit two operators that did the unfashionable, expensive thing and acquired a real charter. [**Xapo Bank**](https://www.xapobank.com/) holds a Gibraltar banking license and a VASP registration, pays 4% APY denominated in Bitcoin, and converts stablecoin rails into protected USD savings — a product the rented-stack crowd is legally barred from offering. [**Fiat24**](https://www.fiat24.com/) runs a Swiss banking charter on-chain, with accounts represented as NFTs and its own USD24 and EUR24 tokens mirroring real fiat against a native Swiss IBAN. These two are not skins on rented rails. They are banks that chose to live on a blockchain, and the moat is the part that took two years and a regulator's signature to build.

# The card stopped being the product

The reason the skins are in trouble is that the thing they compete on is the thing that got commoditized. Walk the feature list of any consumer crypto card today and it is the same list: cashback, better exchange rates, a metal tier, referral bonuses, crypto rewards, lower fees. A hands-on reviewer who tested four of these apps side by side put the structural problem plainly:

> Anyone can slap a Visa or Mastercard on top of a wallet and call it a neobank. All of them want your balance, not just your one-time card swipe.
>
> — [@Kaffchad](https://x.com/Kaffchad/status/2064286661811610002), Jun 9, 2026

Every item on that feature list is copyable. Cashback can be matched, an exchange rate can be undercut, a metal card can be ordered from the same manufacturer your competitor uses. @strato_money's version of the point lands in the same place from the other direction: "None of these are moats. Every feature can be copied. Every reward can be matched. Every fee can be undercut. Every card eventually looks the same. The card becomes infrastructure. Not the product."

This is not a hypothetical race to the bottom. It is already visible in the numbers the operators publish to prove traction. KAST advertises 7% APY and 3% cashback. Plasma One pairs 3% cashback with a 6% yield, fasset offers up to 8% in points plus 5% cashback, Gnosis Pay gives 5% tied to token holdings. These are not differentiators. They are a subsidy war funded by venture capital, and a subsidy any competitor can match is not a moat — it is a cost of staying in the game. The card is now what the Visa logo on it always was: rails.

# Follow the money, and it leaves the building

If the apps own nothing and the features are a wash, the natural question is where the value actually accrues. The teardown answers it by counting vendors. Across the verified production stacks, the same names appear again and again in the columns the operators cannot build themselves. Visa shows up in the clear majority. [Rain](https://www.rain.xyz/) issues the card for a large share of the rest. Solana clears settlement underneath. [Bridge](https://www.bridge.xyz/), the stablecoin orchestrator Stripe paid more than a billion dollars to acquire, moves the dollar flows. Circle's USDC is the unit of account almost everywhere.

![An exploded stack of pale cracked layers floating above one solid red base tray; coin-dots fall through dotted lines from the thin top layers and collect in the red base](/post-images/2026-06-15-stablecoin-neobank-stack-is-forkable/where-margin-lives.jpg)

Read that vendor list the right way and the "no moat" thesis stops being an insult and becomes an accounting identity. Value capture lives in those columns, not in the apps stacked on top. Every operator in the skins tier is a customer of Bridge, Rain, Circle, and Visa — and customers do not capture the margin of their suppliers. When a new neobank launches this week with an identical stack, it is not competition for Bridge and Rain. It is *demand*. The supply shock at the app layer is a revenue event one layer down.

The incumbents are not waiting to be disrupted, either. They are walking downstream into the same product with distribution the startups cannot match:

> neobanks with their existing distribution are coming to crypto and payments. New players that are focusing solely on crypto neobank should pay attention and think about what gives you the edge and moat against incumbent neobanks. We have seen the slow death and consolidation of the first wave of neobank from a decade ago. Most aren't profitable, and crypto can be a way to make neobank profitable.
>
> — [@0xcoconutt](https://x.com/0xcoconutt/status/1983918433226867138), Oct 30, 2025

That last line is the quiet threat. The previous generation of neobanks, the Revoluts and the Chimes, spent a decade learning that the card business is a thin, unprofitable grind unless you own deposits and float. The crypto entrants are speedrunning the same lesson, except the incumbents who already survived it are now arriving with tens of millions of users and a stablecoin toggle.

# Why the skins bleed

The unit economics are where the thesis stops being a vibe and becomes a P&L. Model a representative emerging-market neobank — a freelancer-focused app in the Kosh shape. A user receives roughly $3,000 a month in dollars, holds about $2,000 in USDC, and spends $1,500 a month on the card. The headline that goes in the pitch deck looks like a money printer.

| Revenue line | Per engaged user / yr |
| --- | --- |
| Interchange (0.8% on spend) | $144 |
| FX markup (0.75% inbound) | $270 |
| Float spread (1.5% on balance) | $30 |
| Off-ramp fees | $54 |
| Gross revenue | $498 |
| – Vendor rail pass-through | –$137 |
| – Support / AML / KYC / issuance | –$24 |
| Net contribution / yr | $337 |

A $337 annual contribution against a $25 acquisition cost pays back in under a month. On a slide, that is a category-defining business. Two facts dismantle it.

![A small grey card-swipe slice on the left diverges into two much larger flows, one of them rendered in red, signalling that the swipe is the smallest source of revenue](/post-images/2026-06-15-stablecoin-neobank-stack-is-forkable/interchange-vs-float.jpg)

The first is hidden in the revenue mix. Interchange, the thing the whole product is marketed around, is only about 29% of the gross. Roughly 60% comes from FX markup and float spread. The card swipe is the smallest line on the page. An operator that only earns interchange, with no inbound FX capture and no accumulated balances to spread, is structurally unprofitable. It is paying for users it cannot monetize.

The second fact is the cohort. That $337 figure is the engaged user, and engaged users are a minority. Blend in the people who signed up and went quiet, and in consumer fintech the silent majority always dominates, and the math collapses.

| Cohort | Share | Net / yr |
| --- | --- | --- |
| Power users | 10% | +$641 |
| Active | 25% | +$337 |
| Casual | 30% | +$70 |
| Dormant | 35% | –$5 |
| Blended net / signup / yr | | $168 |

Half the headline. Sixty-five percent of signups are casual or dormant, carrying compliance and support cost while barely transacting, and they drag the blended contribution to $168 per signup before a single dollar goes to engineering, marketing, or general overhead. Then subtract the vendor pass-through that eats 27% of gross on every transaction that *does* happen — the cut Rain takes on the card, the cut Bridge takes on the FX, the cut that, for the skins, is gone forever. At 100,000 signups, the blended net is about $16.8M a year pre-opex. Thin, fragile, and entirely dependent on retention holding in a category where switching costs are a logo swap.

The owners keep the cut the skins pay away. Wirex, a Visa and Mastercard principal member, issues its own cards and keeps the issuing margin. Gnosis Pay owns its chain, its stablecoin, and its EMI relationship, and keeps the spread the others rent. That is the whole game in one sentence: the engine is not a feature, it is the unit economics.

# Is the pond even big enough

The bull case for building one of these is a single staggering number: stablecoins settled somewhere between $33 trillion and $46 trillion in 2025, more than Visa and Mastercard combined. It is true and it is nearly useless. Roughly 70% of that volume was automated bot activity, and another large slab was trading and exchange flows moving between venues. The number that matters is real-economy payments, the actual contested pie these apps are fighting over, and it is closer to **$350–550 billion a year**, per the [BCG and Allium stablecoin work](https://digital-assets.fct.bcg.com/).

There is a second reason to keep the bull case in proportion. A reviewer who tracks the category closely put the scale in context:

> Crypto card spending has grown from roughly $100M per month in early 2023 to more than $1.5B per month by late 2025. But compared with mainstream fintech and card networks, crypto neobanks are still tiny. Wise reported $194.06B in cross-border volume and 15.6M active customers in 2025. American Express reported $1.67T in worldwide billed business in 2025.
>
> — [@simonthekid_](https://x.com/simonthekid_/status/2065014227828371954), Jun 11, 2026

A category spending $1.5 billion a month is real and rising fast. It is also a rounding error next to a single legacy card network, which means the contest is not yet for a mature market. It is for the right to be the app that converts the next wave of stablecoin users into primary accounts, and that has not been proven for any operator in the field.

That contested pie is a genuinely large market, growing fast, and it is not infinite. It comfortably supports a handful of winners with real margins. It does not support 50 identical apps each spending venture money to subsidize cashback for dormant users. The reason so many are launching anyway is not that demand suddenly appeared. It is that the cost of supply collapsed. This is a supply shock dressed up as a market opportunity — the barrier fell, the assembly got cheap, and everyone with a wallet and a weekend shipped the same product. The total addressable market did not grow to meet them.

# The three doors that are still open

So if infrastructure is rentable, cards are commodities, features are copyable, and most of the apps are unprofitable, what is left? The teardown and the timeline converge on the same answer, and it is a short list. @strato_money names the survivors' moat as three things, distribution, trust, and network effects, and the supply-side data agrees, expressed as three concrete doors.

**Door one: own the issuer or the charter.** This is the Xapo and Fiat24 and Wirex path. A banking license or a card-issuer principal membership is expensive, slow, and regulator-gated, which is exactly why it is a moat. It lets you offer products the rented-stack crowd legally cannot, such as protected savings, real yield, and deposit insurance, and it lets you keep the margin everyone else pays away. Two of twenty-nine operators went through this door. That scarcity is the point.

**Door two: own uncopyable distribution.** MoneyGram has a physical cash network across the planet that no startup can replicate. Phantom Cash sits on 40 million-plus existing wallet users. Deel leads with global payroll, so the money lands in its product before the user is ever asked to top up an account. None of these owns the full stack, and Phantom and Deel both rent Bridge, but they each own the one thing money cannot buy quickly: a place users already are. The point was made bluntly:

> The new crypto moat is not just better tech. It is distribution. Base has Coinbase. Hyperliquid has traders. Pumpfun has degens. Circle has USDC. Robinhood has retail. Users don't wake up looking for better infrastructure. They go where the activity already is. The best product does not always win. The product with a clear path to users usually gets more chances to win.
>
> — [@wyckoffweb](https://x.com/wyckoffweb/status/2065072727090766040), Jun 11, 2026

This is the door the incumbents and the brands will walk through, and it reframes the whole category. If banking infrastructure is free, then the bank gets embedded inside whatever already owns the customer. A creator with ten million followers can issue a card overnight. A retail chain can launch a wallet to capture its own float. The most valuable financial products of the next decade may never call themselves banks at all:

> Retail giants are now becoming banks. If the Starbucks app can hold billions in user cash, imagine when retail giants start issuing their own stablecoins to capture the float and bypass credit cards.
>
> — [@ghostweb3](https://x.com/ghostweb3/status/2065768963816980964), Jun 13, 2026

**Door three: own a defensible geographic wedge with real dollar demand.** This is the Kosh-shaped bet done right — pick a market where access to dollars is genuinely scarce, where the FX spread is real revenue rather than a subsidy, and where local trust and compliance are a barrier to the next entrant. The wedge is narrower than a charter and shallower than global distribution, but in the right corridor it produces the one thing the generic skins lack: users who transact because they have to, not because they were bribed with cashback.

Everything outside those three doors is a logo on someone else's rails.

# What the bulls are underpricing

The strongest version of the optimist's case is not that the skins will all survive. It is that the easy-launch world creates so many new financial products, with every brand, community, game, and marketplace issuing its own, that the total demand for the underlying rails explodes, and a marketplace connecting all that new distribution to all those offers becomes inevitable. That is the bet @strato_money lands on, hedged honestly: "Not sure if this is the right idea yet, but this is the gap I keep seeing."

It is a reasonable bet, and it has a tell. Every version of the bull case routes the value back through the layers the skins do not own — the rails that issue the cards, the orchestrators that move the dollars, the networks that own the distribution. Even the optimistic future is one where the app is a thin client and the margin lives underneath. The supply shock is real. It just is not a supply shock in the apps' favor.

A second thing the bulls underprice is that the category has run this experiment before. The first wave of neobanks raised billions, acquired tens of millions of users on the promise that a better card and a cleaner app were a business, and mostly never turned a profit. The ones that survived owned deposits, owned a charter, or owned a captive distribution channel. The crypto wave has better rails and a genuinely new revenue line in stablecoin float — but it is being built on the same flawed premise by most of its entrants, that distribution is something you can buy with a referral bonus. It is not. It is the only thing that was ever scarce.

The honest read is the one both the supply-side teardown and the demand-side essays arrive at independently: the stack is forkable, the fork is nearly free, and a free fork is worth exactly what it costs to make. Thousands of these apps will launch. Most will have identical cards, identical rewards, identical banking partners, and identical infrastructure, and most will disappear — not because they built something broken, but because they never built a reason to exist that a competitor could not assemble over a weekend. Launching a bank is easy now. That was never the hard part.

## Sources

- [@strato_money — Everyone Can Launch a Bank Now (X Article)](https://x.com/strato_money/status/2065873428331291007)
- [@simonthekid_ — Your Bank Got Forked (hands-on review)](https://x.com/simonthekid_/status/2065014227828371954)
- [@Kaffchad — The 3 types of crypto card and their moat](https://x.com/Kaffchad/status/2064286661811610002)
- [@wyckoffweb — The new crypto moat is distribution](https://x.com/wyckoffweb/status/2065072727090766040)
- [@0xcoconutt — Incumbent neobanks come downstream](https://x.com/0xcoconutt/status/1983918433226867138)
- [@ghostweb3 — Retail giants becoming banks](https://x.com/ghostweb3/status/2065768963816980964)
- [neobank.build — verified production stacks](https://www.neobank.build/stacks)
- [BCG / Allium — Stablecoin Payments Dashboard (real-economy payments $350–550B)](https://digital-assets.fct.bcg.com/)
- [BCG — Stablecoin Payments: The Truth Behind the Numbers (white paper, Jan 2026)](https://www.bcg.com/assets/2026/white-paper-stablecoin-payments-truth-behind-numbers.pdf)
- [Bridge — stablecoin orchestration (Stripe)](https://www.bridge.xyz/)
- [Rain — card issuing for stablecoins](https://www.rain.xyz/)
- [Xapo Bank — regulated Bitcoin bank](https://www.xapobank.com/)
- [Fiat24 — Swiss on-chain banking](https://www.fiat24.com/)

---

Canonical: https://www.thedeepfeed.ai/posts/2026-06-15-stablecoin-neobank-stack-is-forkable/
Site: https://www.thedeepfeed.ai
Full corpus: https://www.thedeepfeed.ai/llms-full.txt