Unit economics and cost structure for a self-custody, multi-chain mobile wallet with an integrated trading terminal — benchmarked against Phantom, MetaMask, Trust Wallet, Fomo and Trojan. Cost figures are derived from the operating codebase; revenue figures from published comparables.
Every dollar of cost and profit, per month. "Monthly active users" means people who open and use the app at least once in a given month. All figures are the base case, which assumes NYX earns the same revenue per user as Fomo does today — a directly comparable app charging the identical fee.
| Revenue — 0.5% fee on trades | $5,940,000 |
| Costs of running the service | Per month |
| Blockchain data and node access | $13,910 |
| Computing, hosting and data storage | $1,710 |
| Security and independent app review | $23,910 |
| Legal and compliance | $6,520 |
| Operations, support tooling and monitoring | $3,260 |
| Artificial intelligence and market analysis | $4,890 |
| Total cost of running the service | $54,200 |
| …as a share of revenue | 0.9% |
| Gross profit | |
| Revenue less cost of running the service | $5,885,800 · 99.1% |
| Discretionary investment | Per month |
| Marketing — funding 15% monthly user growth | $1,027,000 |
| Team — 6 people (from the founder alone at launch) | $66,700 |
| Operating profit | $4,792,100 |
| …as a share of revenue | 80.7% |
Running the service costs less than one dollar out of every hundred earned. The two large outflows — marketing and team — are discretionary and can be reduced within a single quarter. The business therefore has no structural cost problem at scale; it has a growth-investment decision.
The same picture at four sizes of business. Throughout this document, users means monthly active users — people who open and use the app at least once in a given month.
| Per month, US dollars | 10,000 users | 50,000 users | 100,000 users | 1,000,000 users |
|---|---|---|---|---|
| Revenue | 59,400 | 297,000 | 594,000 | 5,940,000 |
| Cost of running the service | 4,850 | 9,550 | 16,720 | 54,200 |
| Gross profit | 54,550 | 287,450 | 577,280 | 5,885,800 |
| Gross profit margin | 91.8% | 96.8% | 97.2% | 99.1% |
| Marketing | 9,800 | 48,900 | 97,800 | 1,027,000 |
| Team founder only at 10,000 users | 20,830 | 30,000 | 39,200 | 66,700 |
| Operating profit | 23,920 | 208,550 | 440,280 | 4,792,100 |
| Operating profit margin | 40.3% | 70.2% | 74.1% | 80.7% |
| Cost of running the service, per user | $0.485 | $0.191 | $0.167 | $0.054 |
One hundred times the users costs eleven times as much to serve. The cost of serving each individual user falls by a factor of nearly nine between ten thousand and one million users, because most of the cost base is fixed rather than growing with the user count.
The principal risk is neither cost nor scale — it is how much users actually trade, which is cyclical and largely outside the company's control. Section 11 models the downside. NYX stays gross-profit positive at every size even if users trade as lightly as they do in an ordinary crypto wallet, because marketing and team — the only large discretionary items — can be reduced within a quarter.
The revenue mechanism. NYX takes a flat 0.5% fee on every completed trade. The fee is deducted from the trade itself and settled to the company treasury across three routes: a direct transfer for Solana trades, and a collect-then-convert route for trades on Ethereum-compatible chains, TON and SUI. This is verified in the codebase and has been proven working on-chain.
Why the cost base is unusually small. NYX is non-custodial and runs on the user's device. Each user's phone connects directly to public blockchain infrastructure, so the services that limit usage per internet address — blockchain node access, price feeds, market data, push notifications — cost nothing at all, at any number of users, because the load is spread across a million individual phones instead of concentrating on company servers. There is no central server that every user flows through. This is a property of the architecture, not an optimisation that can erode over time.
| Product | Category | Fee per trade | Note |
|---|---|---|---|
| MetaMask | General Ethereum wallet | 0.875% | Also applies to bridging between chains |
| Phantom | Solana-first wallet | 0.85% | 1.5% flat on fee-free mobile swaps |
| Trust Wallet | General mobile wallet | 0% | By choice; currently subsidising transaction fees |
| Trojan · GMGN · Photon · BullX | Trading terminals | 1.00% | Axiom discounts to 0.75% at its top tier |
| Fomo | Mobile social trading app | 0.50% | Plus a $0.95 minimum charge per trade |
| NYX | Mobile wallet with trading terminal | 0.50% | No minimum — the cheapest in its category |
Fomo's $0.95 minimum means a $50 trade costs a Fomo user 1.9%. The same trade on NYX costs 0.5%. This is a real advantage when acquiring smaller and first-time traders — and a deliberate revenue trade-off.
The entire cost of running the service at one million users — $54,200 per month — amounts to less than half of one basis point of trading volume: 0.46 of NYX's 50. Raising the fee to 0.75% would still undercut every competitor except Fomo and would roughly double gross profit. This headroom is not assumed anywhere in the model.
The central modelling question is how much revenue each user generates per month. Published data across five comparable products produces a wide band that separates cleanly into two product classes roughly forty times apart.
| Product | Users | Fee | Reported revenue | Revenue per user, per month |
|---|---|---|---|---|
| MetaMask | 30 million flat for three years | 0.875% | $41.6M annualised DefiLlama | $0.12 |
| Phantom | 15–17 million | 0.85% | $105M–326M in 2025 sources conflict | $0.58–1.59 |
| Trust Wallet | 15–60 million estimates conflict fourfold | 0% | no swap-fee revenue | ~$0 |
| Fomo | 500–625 thousand traders | 0.50% | Q1-2026 $9.23M · Q2-2026 $7.24M | $4.40–6.20 |
| Trojan | 20k active / 145k registered | 1.00% | $28M daily volume | ~$420 per active user |
NYX is a trading product in a wallet's clothing — a self-custody mobile wallet whose actual product surface (price charts, a trading cockpit, copy-trading, new-token analysis, multi-chain discovery) is a trading terminal. That is precisely Fomo's shape, and Fomo charges the identical fee.
| Dimension | Fomo | NYX |
|---|---|---|
| Who holds the keys | The user | The user |
| Form factor | Mobile-first | Mobile-first |
| Fee | 0.50% plus $0.95 minimum | 0.50%, no minimum |
| Chains supported | Solana, Base, BNB, Monad | Solana, Ethereum-compatible, TON, SUI, XRP |
| Social and copy-trading | Core feature | Built |
| Sign-up | No seed phrase required | No seed phrase required |
Fomo: over $4 billion in cumulative trading volume in roughly twelve months · 625,000+ traders, growing by about 3,500 per day · revenue rising from $1.63M in the third quarter of 2025 to a single record week of $2.64M in August 2026 · trailing weekly revenue 86% above Phantom's · overtook Axiom as Solana's leading daily trading venue by volume · $75M Series B at a $550M valuation led by Index Ventures, with Union Square Ventures and Benchmark.
A mobile, self-custody trading app charging 0.5% can reach a $29M+ annualised revenue run rate on roughly 625,000 users. That is the category NYX competes in.
Three scenarios, each anchored to a named, published comparable rather than an internal assumption.
| Scenario | What it assumes | Trading volume per user, per month | Revenue per user, per month |
|---|---|---|---|
| Downside | Users trade as lightly as they do in an ordinary Phantom-style wallet — most hold rather than trade | ~$130 | $0.65 |
| Base | NYX reaches Fomo's ordinary, non-record monetisation (first- and second-quarter 2026 average) | ~$1,188 | $5.94 |
| Upside | Fomo's record weeks — full trading-terminal intensity | ~$3,600 | $18.00 |
The downside case applies Phantom's trading volume at NYX's lower fee — it assumes NYX both fails to attract more trading than a general-purpose wallet and charges 41% less than Phantom does for it. It is a genuinely pessimistic floor, not a mild haircut.
| Users | Downside — $0.65 | Base — $5.94 | Upside — $18.00 |
|---|---|---|---|
| 10,000 | 6,500 | 59,400 | 180,000 |
| 50,000 | 32,500 | 297,000 | 900,000 |
| 100,000 | 65,000 | 594,000 | 1,800,000 |
| 1,000,000 | 650,000 | 5,940,000 | 18,000,000 |
At one million users the base case implies $71.3 million in annual revenue — against Fomo's $29 million annualised run rate on 625,000 users. The model assumes NYX reaches Fomo's per-user monetisation, not that it exceeds it.
Excluded at management's instruction. NYX has a built and tested referral system that pays a share of the fee to whoever introduced the user. Users who arrived without a referrer still yield the full fee. At high referral take-up the company's share of gross fees would fall to roughly 71%. These revenue figures must not have referral revenue added on top without applying that reduction.
Cost to acquire a customer. Industry benchmarks for 2026: a crypto wallet user costs roughly $25 to acquire through general advertising; the best-performing campaigns reach $6–8, the worst $44. Targeting people who already hold crypto cuts this by 50–70%. Paying individual crypto influencers a flat fee has collapsed as a channel — Forbes documented a $30,000 payment that produced a single sign-up — with budgets moving to short-form video clipping at a fraction of the cost. Planning figure: $13 per user.
How long a user stays. Deliberately conservative at six months. Retention in this market is poor: the average holding period for a speculative token is around 100 seconds, and 96.6% of Solana wallet addresses that trade once disappear within a day. A twelve-month assumption would not be defensible.
| Downside | Base | Upside | |
|---|---|---|---|
| Revenue per user, per month | $0.65 | $5.94 | $18.00 |
| Total revenue per user over six months | $3.90 | $35.64 | $108.00 |
| Cost to acquire that user | $13.00 | $13.00 | $13.00 |
| Ratio of user value to acquisition cost | 0.30× | 2.74× | 8.31× |
| Time to earn back the acquisition cost | never | 2.2 months | 0.7 months |
| Conventional healthy threshold | 3.0× | 3.0× | 3.0× |
Reading this honestly. The base case sits just below the conventional 3.0× threshold when measured on an average user — but earns the acquisition cost back in 2.2 months, which is strong. Two factors materially improve on the average:
| Company | Valuation | Users | Per user | Multiple of revenue |
|---|---|---|---|---|
| Fomo | $550M after its June 2026 round | 625,000 traders | ~$880 | ~18.8× |
| Phantom | ~$3 billion | 15 million | ~$200 | — |
| Consensys / MetaMask | $7–12 billion reported listing range | 30 million | ~$233–400 | — |
Costs are derived from the operating codebase, not from vendor price lists.
This is the market-data layer: live prices, price charts and token information. The server caches results so that many users viewing the same token produce a single request to the data provider rather than one per user. Cost therefore grows with the number of distinct tokens being viewed, not with the number of users — which is why it grows only elevenfold as users grow a hundredfold.
| Users | Peak simultaneous | Distinct tokens viewed | Peak requests per second |
|---|---|---|---|
| 10,000 | 175 | ~61 | 112 |
| 50,000 | 875 | ~302 | 553 |
| 100,000 | 1,750 | ~598 | 1,094 |
| 1,000,000 | 17,500 | ~4,800 | 8,779 |
$0 at every size — spread across user devices: Ethereum-compatible node access · Jupiter · DexScreener · GeckoTerminal · SUI · push notifications · self-hosted analytics.
One server at ten thousand users, rising to six to ten servers across two data centres at one million. The same capacity rented from Amazon or Google would cost $16,000–43,000 per month; dedicated hardware is a ten- to twenty-fold saving.
Data storage grows from about 70 gigabytes at ten thousand users to 5–10 terabytes at one million, and costs 0.1–0.4% of the total at every size — it sits on disks already paid for in the server fleet.
At one million users, roughly 2 gigabytes per user per month means about 2 petabytes of data leaving the servers every month. Delivered from Amazon or Google at $0.05–0.09 per gigabyte, that alone would cost $100,000–180,000 per month. Delivered from Cloudflare, which charges nothing for outbound data, plus unmetered dedicated servers, it costs nothing. This single architectural decision is worth more than every other line in this document combined, and it is already in place.
Independent security review is treated as a fixed, non-negotiable cost from about ten thousand users onward. This is deliberate and material: because NYX never holds customer funds, it carries no custody risk — but a compromised app release or a poisoned software dependency would affect every user at once. The potential damage is wide even though the company holds nothing.
| Component | What it is | Cost | Per month |
|---|---|---|---|
| Pre-launch security audit | Independent firm reviews the wallet, key handling and cryptography line by line. Firms of the class of Hacken, BlockApex, Trail of Bits, Cure53. | $16–54k once | — |
| Annual re-audit | Repeat review as the code changes | $43–76k / year | 3,600–6,300 |
| Mobile penetration testing, twice a year | Specialists actively attempt to break the iPhone app, Android app and the interfaces behind them | $27–43k / year | 2,300–3,600 |
| App hardening | Anti-tampering, detection of modified phones, code obfuscation, screenshot blocking | $12–41k / year | 980–3,700 |
| Bug bounty reserve | Paying independent researchers who report vulnerabilities, via Immunefi | reserve | 3,300–8,700 |
| Attack filtering and monitoring | Web firewall, denial-of-service protection, intrusion monitoring, scanning of third-party software dependencies | — | 470–6,400 |
| Breach-response retainer and cyber insurance | Specialists on standby, plus cover | — | 1,500–5,200 |
The scoping question is the single largest legal variable, and it is unresolved until counsel opines. It is addressed in full in Section 8.
| Scenario | Basis | Year-one cash | Ongoing / month |
|---|---|---|---|
| Base — no licence required | Pure non-custodial software: no custody, no order matching, no handling of customer money. Cost = scope opinion, company formation, terms of service and privacy documentation, trademarks, ongoing counsel. | $30–79k | $870–6,520 |
| Contingency — licence required | Only if the copy-trading service is left holding customer keys. Legal file $43–217k · application fees $5.4–27k · regulatory capital $54–163k · local director ~$37k/year · compliance officer and local hires $109–272k · 9–15 month timeline. | $272–652k | $11–27k |
Shown as growth investment rather than as a cost of running the service, so that gross profit is not obscured by discretionary spending.
| Users | Lean — organic and referral-led | Growth — 15% more users each month | Aggressive — 30% more each month |
|---|---|---|---|
| 10,000 | 2,200 | 9,800 | 27,200 |
| 50,000 | 6,500 | 48,900 | 137,000 |
| 100,000 | 13,000 | 97,800 | 274,000 |
| 1,000,000 | 65,200 | 1,027,000 | 3,196,000 |
| Block | 10,000 | 50,000 | 100,000 | 1,000,000 |
|---|---|---|---|---|
| A · Blockchain data and node access | 1,360 | 2,620 | 4,660 | 13,910 |
| B · Computing, hosting and storage | 180 | 300 | 640 | 1,710 |
| C · Security and independent app review | 1,630 | 3,800 | 6,520 | 23,910 |
| D · Legal and compliance | 870 | 1,630 | 2,720 | 6,520 |
| E · Operations, support tooling and monitoring | 160 | 440 | 980 | 3,260 |
| F · Artificial intelligence and market analysis | 650 | 760 | 1,200 | 4,890 |
| Total per month | 4,850 | 9,550 | 16,720 | 54,200 |
| Per user, per month | $0.485 | $0.191 | $0.167 | $0.054 |
Four layers, drawn from the source tree rather than from marketing material. The arrangement below is what keeps the company outside the scope of a financial licence — the argument set out in Section 8.
Keys are generated and stored on the user's phone and are never transmitted. There is no mechanism — technical or administrative — by which NYX could move a user's assets.
The only server in the trading path caches prices and charts. It contains no order-handling code at all, so there is no order for NYX to receive, hold, route or forward.
Wallet tracking now notifies rather than trades: a blockchain webhook arrives, the server forwards a push notification, and the user decides. No key reaches the server and no order is placed for anyone.
Verified by reference count in the source tree, so this reflects what is built rather than what is planned.
| Category | Providers | Why it matters legally |
|---|---|---|
| Trade routing and liquidity | Jupiter (Solana) · KyberSwap (Ethereum-compatible chains) · pump.fun · NOXA-DEX (Robinhood Chain) · STON.fi and DeDust (TON) · Aftermath (SUI) · the XRP Ledger's own order books and market maker · Relay and ChangeNOW (cross-chain) | All non-custodial. The liquidity pool is the counterparty to each trade — never NYX. |
| Robinhood Chain | Tokens launched through the NOXA launchpad trade in pools that no external aggregator indexes, so NYX calls those pool contracts directly — quoting, buying and selling on-chain. | Direct contract calls from the device. No intermediary holds anything. |
| XRP Ledger | The ledger performs its own routing, so no aggregator exists. NYX builds a payment routed through the ledger's order books and automated market maker, plus the trust line a first purchase requires. | Signed locally with the user's own key. NYX never holds the key or the asset. |
| Sell strategy — take-profit | Jupiter Trigger V1. Each tier becomes an on-chain limit order: "sell N tokens for at least M". The user signs once on the device; Jupiter's permissionless keeper network fills it even with the app closed. The app calls Jupiter directly, not through any NYX server. | Non-custodial. Coins stay in the user's wallet. No key on any server. |
| Sell strategy — stop-loss | Jupiter Trigger V2. A stop-loss cannot sit on-chain unattended, so it needs a watcher. Jupiter's V2 deposits the tokens into a Jupiter-managed vault and its keeper signs the fill. | Custodial — by Jupiter, not NYX. Disclosed in-app before the feature can be enabled. |
| Market data | Birdeye · DexScreener · GeckoTerminal · CoinGecko · RugCheck · Blockscout | Read-only information. No customer funds or orders involved. |
| Buying crypto with cash | MoonPay · Transak · Coinbase Pay · Ramp · Onramper | Each is a licensed payment provider running its own identity checks, delivering funds straight to the user's own wallet address. NYX never handles the money. |
| Blockchain access | Helius and public community nodes | Standard infrastructure, treated as a shared public resource rather than an outsourcing relationship. |
Assessment: most likely no. The one exception has been designed out and ships in build 204.
Europe's crypto regulation, known as MiCA, requires a licence — a "crypto-asset service provider" authorisation — from firms that provide certain defined services. The transitional period for existing firms ended on 1 July 2026, so this question is live rather than hypothetical.
The decisive principle is that MiCA regulates services, not software. The licence is triggered by holding customers' assets, running a trading venue, acting as the counterparty to a trade, or handling customers' orders on their behalf. Taking each in turn:
| Regulated service | Does NYX do this? | Assessment |
|---|---|---|
| Holding customer assets | No. Keys are created and stored on the user's own phone. NYX has no access to them and cannot move a user's funds under any circumstances. | Outside scope |
| Running a trading venue | No. NYX matches no orders and operates no marketplace. | Outside scope |
| Acting as counterparty to trades | No. The other side of every trade is a public liquidity pool, never NYX. | Outside scope |
| Executing orders for customers | No. The user's own device signs and submits every transaction. | Outside scope |
| Receiving and passing on customer orders | The one genuinely arguable category — examined below. | The live question |
Under guidance from the European securities regulator (ESMA, question-and-answer 2653), this category is triggered where a firm forwards a customer's order to a third party which then completes the trade. Importantly, that guidance states the assessment must follow the operational reality of what the firm does, not how the firm describes itself.
NYX's operational reality, verified in the codebase:
This is a materially stronger position than a typical decentralised-exchange website, which normally routes orders through its own backend infrastructure. NYX does not have an order path to regulate.
Section 7 shows this visually: the green path — the signed transaction — runs outside the NYX server box entirely. That diagram is drawn from the source tree rather than from marketing material, and is intended as the technical exhibit accompanying a legal scope opinion.
Copy-trading previously stored users' signing keys on the NYX server and placed buy and sell orders on their behalf. That was the one component capable of pulling the company into licensing scope, and it has been redesigned to alerts only: a blockchain webhook reaches the server, the server forwards a push notification, and the user decides and signs on their own device. The server holds no key and places no order.
The decision layer is built and tested — suppressing dust, collapsing repeat buys, capping any one wallet's share of the feed, and marking stale trades rather than hiding them. Alert wording is constrained by test to report what happened and never to instruct, because instruction is advice and advice is regulated.
Status: the decision layer and its tests are complete; the final removal of the key-passing line from the client, and the webhook-to-push server piece, land in build 204. Until that build ships, the legacy path remains present in the codebase.
The base case — no licence required, $870 to $6,520 per month in ordinary legal costs — is the reasonable planning assumption for the core wallet. Policymakers have signalled that self-custody stays outside the regulatory perimeter, and major self-custody wallets are not licensed as service providers.
With copy-trading redesigned to alerts, none of the six triggers is met and the $272,000–652,000 authorisation figure in Section 9 is a remote contingency rather than a live exposure. A written legal opinion costing $16,000–43,000 converts this from an open question into a documented position — and is worth commissioning before fundraising, since it removes a several-hundred-thousand-dollar uncertainty from the diligence conversation.
This section is a commercial assessment of publicly available regulatory guidance. It is not legal advice and does not substitute for a formal opinion from qualified counsel. European legal and regulatory costs are incurred in euro and converted here at the rate stated in the assumptions.
Outside the monthly running costs.
| Item | What it covers | US dollars |
|---|---|---|
| Pre-launch security audit | Independent review of the wallet, key handling and cryptography | 16,000–54,000 |
| Legal opinion on licensing scope | Formally documents the position set out in Section 8 | 16,000–43,000 |
| Company formation and corporate work | Incorporation, share structure, contracts | 5,400–16,000 |
| Terms of service, privacy policy, app-store documentation | Across the jurisdictions where the app is distributed | 5,400–10,900 |
| Trademark registration | NYX name and logo, EU trademark office plus key markets | 3,300–8,700 |
| Settlement contract deployment | One-off on-chain deployment cost | ~650 |
| Apple and Google developer accounts | $99 per year plus $25 once | ~135 |
| Total year-one one-time cost | 47,500–134,500 | |
| Contingency — licence application | Only if copy-trading is left custodial and the legal opinion places NYX in scope (Section 8). Excludes $54,000–163,000 of regulatory capital, which is held on the balance sheet rather than spent. | 272,000–652,000 |
| Per month, US dollars | 10,000 users | 50,000 | 100,000 | 1,000,000 |
|---|---|---|---|---|
| Revenue — 0.5% fee on trades | 59,400 | 297,000 | 594,000 | 5,940,000 |
| − Blockchain data and node access | −1,360 | −2,620 | −4,660 | −13,910 |
| − Computing, hosting and storage | −180 | −300 | −640 | −1,710 |
| − Security and independent app review | −1,630 | −3,800 | −6,520 | −23,910 |
| − Legal and compliance | −870 | −1,630 | −2,720 | −6,520 |
| − Operations and support tooling | −160 | −440 | −980 | −3,260 |
| − Artificial intelligence and market analysis | −650 | −760 | −1,200 | −4,890 |
| = Gross profit | 54,550 | 287,450 | 577,280 | 5,885,800 |
| Gross profit margin | 91.8% | 96.8% | 97.2% | 99.1% |
| − Marketing, funding 15% monthly growth | −9,800 | −48,900 | −97,800 | −1,027,000 |
| = Profit after marketing | 44,750 | 238,550 | 479,480 | 4,858,800 |
| − Team 1 person → 6 | −20,830 | −30,000 | −39,200 | −66,700 |
| = Operating profit | 23,920 | 208,550 | 440,280 | 4,792,100 |
| Operating profit margin | 40.3% | 70.2% | 74.1% | 80.7% |
The team line begins with the founder alone, at $250,000 per year fully loaded (salary, employer contributions, equipment and workspace) — roughly $20,830 per month — and grows from there: about two people at fifty thousand users, three at one hundred thousand, and six at one million, with later hires costed at a lower average than the founder. The count is deliberately lean because the non-custodial architecture removes the heaviest support burdens — there are no account recoveries, frozen balances or withdrawal tickets, because the company never holds customer funds.
This is a placeholder for planning, not a hiring plan. Hiring pace, seniority mix, location and the split between employees and contractors would each move it materially. It is the least developed line in this model and should be rebuilt against an actual org plan before it is relied upon.
After the cost of running the service, marketing and team:
| Users | Downside | Base | Upside |
|---|---|---|---|
| 10,000 | −28,980 | +23,920 | +144,520 |
| 50,000 | −55,950 | +208,550 | +811,550 |
| 100,000 | −88,720 | +440,280 | +1,646,280 |
| 1,000,000 | −497,900 | +4,792,100 | +16,852,100 |
The downside case loses money only because it is modelled with full marketing spend and a full team. Removing the discretionary marketing budget returns it to profit at every size:
| Users | Downside revenue | − cost of running the service | Gross profit | Margin |
|---|---|---|---|---|
| 10,000 | 6,500 | −4,850 | +1,650 | 25.4% |
| 50,000 | 32,500 | −9,550 | +22,950 | 70.6% |
| 100,000 | 65,000 | −16,720 | +48,280 | 74.3% |
| 1,000,000 | 650,000 | −54,200 | +595,800 | 91.7% |
The cost base is small enough that even wallet-like trading levels cover it at a 25–92% gross margin. Marketing and team are discretionary and can be reduced inside one quarter. The company does not have a structural cost problem in any scenario — it has a growth-investment decision.
| Risk | Assessment | Mitigation |
|---|---|---|
| How much users trade is cyclical | The primary risk. Revenue varies twenty-eightfold between the downside and upside scenarios while cost barely moves. Speculative-token trading recently fell below 5% of Solana's total trading volume, the lowest share since 2023. | The cost base is small and largely fixed; marketing can be reduced within a quarter. Supporting several blockchains reduces dependence on any single ecosystem. |
| Copy-trading design | Formerly the one feature that could require a European licence. Redesigned to alerts only — the server holds no key and places no order. The final client change ships in build 204. | Once shipped, none of the six licensing triggers is met, and the $272,000–652,000 contingency becomes remote rather than live. |
| Wide security exposure | Holding no customer funds removes custody risk, but a compromised app release or poisoned software dependency would affect every user at once. | The largest cost block at every size — independent audit, twice-yearly penetration testing, app hardening, bug bounty, dependency scanning. Budgeted from ten thousand users. |
| Three known engineering ceilings | Copy-trade tracking saturates at around sixty tracked wallets; the market-data cache runs on a single server; there is no database yet. All are well below one million users; two are immediate. | Two fixes are a few hours of work. The remainder is a single six- to ten-week engineering project, to begin before fifty thousand users. |
| Users do not stay | Speculative-token traders churn quickly — around a 100-second average holding period. Even the six-month assumption may be generous. | Modelled conservatively. Wallet features — multi-chain custody, offline payments — give reasons to stay that pure trading terminals lack. |
| Competition | Fomo, Axiom, Trojan, Photon and GMGN, plus Phantom moving into trading. | Cheapest fee in the category; differentiated on self-custody, chain coverage and offline capability. |
| Revenue concentration | A small group of heavy traders generates most of the revenue. | Known and modelled; drives the segment-based acquisition strategy in Section 5. |
| Assumption | Value | Basis |
|---|---|---|
| Share of monthly users active on a given day | 25% | Trading-app engagement norm |
| Share of daily users online simultaneously at peak | 7% | Standard mobile usage curve |
| Share of simultaneous users viewing a price chart | 35% | Product judgement |
| Number of distinct tokens actively traded | 12,000 | Estimate — measurable from the company's own systems |
| Trading fee | 0.5% flat | Verified in the codebase |
| How long a user stays | 6 months | Conservative against observed market retention |
| Cost to acquire a user | $13 | 2026 benchmark for crypto-native targeting, mid-range |
| Growth rate in the marketing case | 15% per month | Management scenario, half of it paid |
| Team size and cost | 1 → 6 people | Founder only at launch, $250,000/year fully loaded, growing from there. Indicative — the least developed line in this model. |
| Exchange rate for euro-denominated costs | 1 EUR = 1.087 USD | European legal, hosting and some security costs are incurred in euro |
Excluded from the model: referral payouts (Section 4), token or points programmes, revenue share from buying crypto with a card, derivatives revenue, and paid subscription tiers. Several represent upside not reflected here — Fomo, for example, earns additional revenue from derivatives routing.