For over a century, retail traders have faced the same structure under different names: an intermediary that is at once their counterparty (it profits when they lose), their information source, their custodian, and their gatekeeper. The harm is not anecdotal — it is the regulators' own finding, from ESMA's 74–89% loss rates on retail contracts-for-difference (CFDs) to the $6 billion of self-issued collateral discovered inside FTX. Rules have fined and capped this conduct for two decades without removing it, because it cannot be removed by rules — only by architecture.

On-chain markets are that architecture: self-custody replaces the custodian, atomic settlement removes the counterparty, a public ledger ends the information monopoly, permissionless access opens the gate, and two capabilities finance never had — composability and programmability — turn products into interoperable software whose rules are enforced by code. This is neither fringe nor theoretical: the Bank for International Settlements now proposes the same primitives as the blueprint for the next monetary system, and Hyperliquid, the largest on-chain venue, cleared roughly $3 trillion in 2025 at spreads matching the world's reference markets, inspectable by anyone.

But a venue can only take this halfway. A ledger can be public and still unreadable; access can be permissionless and still terrifying; settlement can be atomic and the path to it still a maze. That unfinished half is Arx — Latin for citadel. For two thousand years, money has lived inside someone else's fortress, on the fortress's terms.

Arx is the moment the fortress changes hands: not a better landlord, but the keys themselves, handed to the person who lives inside — self-custody and records they can check (trust), a way in that works like an app (access), a published intelligence scorecard (intelligence), and risk limits it cannot bypass (protection). Reading and reaching markets is the first thing done inside those walls; investing, payments, and borrowing — built by specialists — follow.

Picture where this leads. Someone with a self-custodial wallet — a student, a saver, an independent trader, whether in a financial capital or a country whose banks and currency have failed them — can reach the same markets once reserved for the well-banked, on the same terms, read them clearly, act with the downside known and capped, and never have to trust a stranger with their money or their truth. A fairer chance to participate than the old system allowed — rationed less by where you start or which institutions will have you. That is what decentralized venues and Arx are built to bring about — and every claim on the way there is written to be verified, not believed.

Strip away the product names and every era of retail trading shows the same architecture. The intermediary between a retail trader and the market holds four powers at once — each an advantage over the client that an open market would compete away:

Any one of these creates a conflict of interest. Held together, they create an equilibrium in which client harm is not an accident but a business model. The historical record is unusually consistent:

Two details in this record matter most. First, the harm is not explained by market risk alone. ASIC found that roughly 5% of losing CFD clients would have been profitable but for fees, and that loss rates rise steadily with trading frequency — from 59% for occasional traders to 76% for the most active (2026). The structure extracts through fees, spreads, financing, and execution as much as through direction.

Second, verification was impossible by design. A CFD trader cannot audit the feed she trades against. A brokerage client cannot see what the wholesaler paid for his order. An FTX depositor could not inspect the balance sheet. When researchers logged a retail broker's trades against its own data feed, client statements showed profits 50% lower and losses 50% higher than the broker's own API reported — and roughly one in thirteen trades vanished from the history entirely (Shahtahmassebi & Wright, 2021).

In every case the defining feature is not that the intermediary did cheat — it is that no client could ever check. Markets where one side cannot verify are markets where, eventually, the other side does not have to behave. Behind the four powers sit two quieter facts that make them durable: products live in silos (collateral trapped, an entire industry existing just to reconcile records), and rules are promises (fee schedules and "fair value" assessments the intermediary can waive, change, or self-grade). This is the problem on-chain finance addresses: not volatility, but a specific, recurring structure of conflicted intermediation that rules cannot remove — only architecture can.

A blockchain-settled market changes the structure of intermediation in six specific ways. Each is a paradigm shift with a name, a structural problem it dissolves, and a division of labor: the venue layer makes each shift possible; the application layer makes it real for a person. That division is the relationship between decentralized venues — Hyperliquid first among them — and Arx.

Shift 1 — Self-custody: from custodial trust to self-control. Ownership becomes a property of the asset itself — a venue cannot freeze, divert, or re-lend what it never holds. Celsius froze the withdrawal button on 1.7 million users with $4.7B inside; FTX diverted roughly $8B of customer deposits; the crypto-lender model quietly re-lent coins depositors believed were safe. On-chain, no one holds your coins — so no one can lock, lend, or lose them. Arx makes it livable: a wallet you open with an email, funded and cashed out with the payment methods people already use, keys created client-side and exportable, the exit a first-class flow.

Shift 2 — Atomic settlement: from conflicted middlemen to efficiency. Every trade has a moment of danger — when asset and money live on two ledgers, one side pays first and prays. Finance filled that gap with trusted intermediaries, and each one charges for the trust and can abuse the position. On a blockchain, asset and money share one ledger, so a trade is a single state change: both legs execute, or neither exists. The dangerous moment disappears, and with it the middleman who guarded it.

Shift 3 — The public ledger: from information monopoly to transparency. The order lifecycle, every position, and the venue's own solvency become shared state — "trust me" becomes "check for yourself." But raw transparency is a firehose, not a product: the venue publishes the truth; Arx makes it legible — plain-language reads with published accuracy, every displayed number traceable to public state.

Shift 4 — Permissionless access: from gated markets to fairness. Any wallet faces the same market on the same terms — and market creation itself became permissionless: builder-deployed perpetuals (HIP-3) brought gold, oil, indices, and stocks on-chain in months, and outcome markets (HIP-4) added prediction and event markets on the same rails. The venue opens the protocol; Arx opens the door: mobile-first, in the languages and funding methods users already hold.

Shift 5 — Composability: from product silos to compounding usefulness. Instruments are interoperable by construction: one margin account can back several at once, and a position can collateralize a loan without paperwork. The venue supplies composable primitives; Arx composes them into one coherent experience — order transmission, copying, intelligence, risk, and in time investing, payment, and lending, on a single self-custodial wallet.

Shift 6 — Programmability: from promised rules to enforceable fairness. The revocable promise — fees quietly changed, "fair value" self-assessed, margin waived when convenient — ends when rules are code: a published, deterministic liquidation formula that runs identically for everyone. Code changes the grammar of trust from "we won't" to "we can't," and Arx applies it to itself — risk rails no one at Arx can waive, an intelligence scorecard whose methodology is fixed in advance.

The strongest evidence that these six shifts have won the argument comes from their most conservative critic. The Bank for International Settlements — the central bank of central banks — now advances a "unified ledger" as the blueprint for the next-generation monetary and financial system (Annual Economic Report, 2025). That is the same physics with a different trust anchor, stated as an ultimatum:

The debate is no longer whether the shifts are superior — only who operates the ledger and how open it is. Arx's answer is that openness is not a detail: a ledger you cannot independently verify reproduces the information monopoly, whoever runs it.

Arx starts with the people who have the most to gain from honest markets and the least protection in today's ones: everyday traders and investors whose local options are an offshore broker that can bet against them or a bank that shuts them out. Two kinds of them, in particular, need each other.

The first, and the larger group, would rather follow a skilled trader than build a market thesis from scratch. They care more about not losing than about a spectacular win, and today they get burned in familiar ways: they follow a leaderboard that turns out to be self-reported, watch their own copied return come in quietly worse than the number that drew them in, or find they cannot withdraw when it matters most. What they lack is a way to tell skill from luck before they commit a dollar, and a record they can actually check.

The second already runs their own book — on a decentralized venue or a centralized exchange — watches funding daily, and stitches signals together from half a dozen tools. They do not need teaching. They need their edge in one place, their risk shown before they enter a position rather than discovered in a liquidation, and a track record that speaks for itself because it lives on-chain.

These two fit together. The follower's need for someone worth trusting is answered by the independent trader's need to be recognized for real skill, and the public ledger is what makes the trust between them genuine rather than promised. Arx's job is to make each side legible to the other, and to protect both from the exact failure modes the record documents.

A third group sits right beside them: traders who already deal contracts on gold, indices, or currencies through a CFD or futures broker, and have done the math on its overnight fees and its house-versus-client incentives. The same instruments now exist on rails that cannot take the other side of their trade. For them Arx is a bridge, not a pitch.

What unites all three is a dangerous gap: they want upside, but the mechanics that decide whether they keep it — liquidation, funding, slippage, correlation, and their own behavior at three in the morning — are exactly the ones the old venues never taught and never guarded. Arx's risk discipline is built for that gap: not a nag, but the difference between surviving your first year and funding someone else's.

These are where Arx begins, not where it ends. The same foundation — checkable records, enforced risk limits, honest reporting — serves everyone who comes next: the systematic trader who wants programmatic access, the allocator who assembles their own selection of Participants, the first-time saver who arrives to save rather than trade, and eventually the software agents that act on a person's behalf within limits their owner sets. The ambition is the whole spectrum the vision names. The discipline is earning it one group at a time, so the product widens without ever being rebuilt.

Two thousand years ago, on the highest hill in Rome, stood a fortress the Romans called the arx. Inside its walls, beside the temple of Juno Moneta, the Republic struck its coins. That fortress is why your money is called money, and why the place it is made is called a mint. Money was born inside a citadel — and it never left.

Every era since has rebuilt the same arrangement with better marble: the vault, the clearing house, the brokerage, the exchange. The walls grew taller and the names grew friendlier, but the deal never changed — your wealth lives inside their fortress. They set the price you see, hold the keys you don't, and decide when the gate opens. For two thousand years, you have waited outside the walls of your own money.

Then, for the first time in the history of the hill, walls learned to belong to anyone: a public ledger has no privileged inside. And so the citadel changes hands — that is the name, and the thesis. Arx is your citadel, not ours. A treasury only you can open, a hall where trades settle the moment they are struck, walls of glass no one can lie behind, a gate open to every market on the same terms, stones that lock into one another, and laws that bind even the castle's builders. You are not the visitor this time; you are the sovereign. Arx stands where a citadel's best servants stand — on the watchtower, at the council table, on the walls. Never on the throne.

Arx is a mobile-native, non-custodial SaaS analytics and order-transmission platform that pairs access to on-chain markets — spot and perpetuals on crypto, and perpetuals on equities, commodities, and FX as the Connected Protocol lists them — with accountable intelligence, self-directed Copy Activity, and enforced risk discipline, for eligible users who hold their own wallet. A decentralized exchange — the Connected Protocol — settles trades, matches orders, and holds liquidity, honestly and in public, at a capital-intensive scale deliberately not Arx's job. But a venue gives everyone the same raw rails and stops there. Arx is the layer above it — and above every provider that follows. Trust is the foundation the other three rest on:

The relationship is complementary, not competitive: the providers beneath supply the primitives; Arx supplies the users, the trust, and a reason for a non-expert to arrive at all. These upper layers are won on experience and trust, not raw capital — which is why an application, not the infrastructure beneath it, owns them.

Each refusal below is structural — it is how Arx avoids re-acquiring the four powers:

The venue layer reached parity only in the last two years; builder-deployed markets made multi-asset breadth real only in the last year; and the populations with the most to gain — users in access-constrained markets — remain unserved by anything that combines self-custody with genuine usability. Arx connects to Hyperliquid — its current Connected Protocol — because it is the first on-chain venue whose architecture (a fully public order book, no internalized flow, permissionless multi-asset market creation) takes a real step toward the vision this document describes; Arx supplies the layer above it — trust, access, intelligence, protection.

Copy Activity is really about observing behavior. A Participant's on-chain activity is verifiable fact — the Trust foundation — so it becomes one more input you can use two ways: follow it, watching a Participant's moves and folding them into your own read; or copy it, replicating each move at your own election inside limits you set. Either way the behavior is weighed alongside the rest of your intelligence — never followed blind. It is also the most demanded feature in retail markets and the most abused. Arx rebuilds it on the public ledger, with four commitments the incumbent structure cannot make.

The Participants shown are real wallets whose complete history — every entry, exit, drawdown, and dollar of realized PnL — lives on Hyperliquid's public ledger. Arx displays that public record neutrally; it does not select, rank, verify, endorse, or produce it, and any user can check any figure against the chain. Profiles are readable labels over real addresses, one tap from the underlying wallet; no performance figure is ever synthesized, and placement is never sold.

Every wallet shown leads with maximum drawdown and realized profit and loss — money actually closed and banked, not paper gains or a cherry-picked win — wins and losses both on screen, and past performance does not indicate future results stated plainly. Two facts decide whether replicating is wise.

First, alignment: the Participant risks their own capital on the very same position — a real wallet, not a signal sold from the sidelines. Copy Activity carries no performance fee today: where Arx rewards a trader others follow, it pays that reward from its own funds — for contributing the activity, never a cut of any user's gains, losses, or volume.

Second, honest divergence: because replication mirrors each move at your own size and funding clock, your realized return is never the Participant's published one — and across trades that gap is why only a fraction of a headline return reaches copiers elsewhere. Arx computes your true copier return from the public ledger and shows it beside the headline before you commit a dollar.

Every wallet shows how much live history stands behind it — active days, number of trades — as a plain fact, so a lucky fortnight is never mistaken for a long record. Arx makes no judgment of skill and applies no ranking or endorsement; you filter and sort the public data yourself and study what you would be replicating before you replicate it. Surfaces are study-first by design — never "copy now."

No Copy Activity can start without the user setting a loss limit, a position-size limit, and a cap on leverage; a circuit breaker pauses replication when the loss limit is hit. Copy Activity is self-directed, scoped, and revocable at any time — replication runs inside the user's own custody, never through a handover of funds. The Participant has no access to, control over, or discretion over the user's wallet, cannot see who replicates them, and the user can stop at any time.

The unfinished half established the intelligence gap: public data, private edge. Arx's intelligence layer exists to close it — with a design constraint most trading products refuse: everything it says must be checkable, and its accuracy must be published.

Markets move on several kinds of information at once: how the traders you can follow are positioned, what wallets more broadly are doing on-chain, how the order book itself is behaving, what an asset fundamentally is, and what the chart says — each one input, none read alone. A trader normally stitches these together from half a dozen tools and loses the signal in the noise. Arx fuses them into one cohesive decision flow, held to three standards: coverage (nothing that matters is missing), depth (each signal read properly, not headline-skimmed), and timeliness (arriving while the decision is still open). Fewer, better, faster reads — and the user decides.

Rather than an endless headline list, Arx's news layer looks back over each market-moving event's last 24 hours to show why it matters to your book: the on-chain flow around it, how similar events moved price before (with a confidence label drawn from the sample size), any related prediction- or event-market odds the protocol lists, and the assets moving with it. Corroboration in place of doom-scrolling — the newest dimension of the decision flow, and the one still shipping.

A trader rarely wants "a perpetual" — they hold a view, an exposure, a horizon, and a loss they can live with. Arx's harder job is to translate that intent into expression: given what you believe and what you can risk, it lays out how the same view could be taken across the instruments available — spot, perpetual, and in time options and event markets — with the cost, leverage, and downside of each shown side by side, the option of not trading among them. It maps; it never recommends. You choose the instrument, the size, and whether to act at all.

The deepest input is people. The public ledger gives everyone the same raw facts about every wallet, but raw facts are not context — and a single statistic can mislead, since a flattering Sharpe ratio can be six lucky weeks. So Arx computes informational indicators over that public record: how much history stands behind a wallet, how positions are distributed, how crowded or stressed a market is — decision-grade context a person interprets for themselves, every indicator sitting on public data the user can verify beneath it. These are informational analytics, not a rating, ranking, or endorsement. How the indicators are computed is proprietary; the inputs are public.

Every signal is evaluated against what subsequently happened and carries a measured hit rate; signals that do not demonstrate value are retired; the engine's own confidence is calibrated against realized outcomes. The published record includes the failures — because a signal product that reports only its wins is a marketing product. One honest caveat, stated plainly: unlike copy records, which anyone can recompute from the ledger, these hit rates are Arx's own measurement — the commitment is procedural (methodology fixed in advance, every signal scored, misses published), and the underlying inputs are public for anyone who wants to check the reads against the chain.

This engine cannot be built where the book is private — the data does not exist to be read. It is possible on Hyperliquid because the order lifecycle is public down to the wallet level, and the evidence is blunt: on a transparent venue, those who read the book get measurably better outcomes than those who do not. Arx's purpose is to move that advantage from the 0.1% who can build the tooling to everyone else. Signals inform; they never advise, predict prices, or size positions — the user decides.

Discipline compounds only if a trader can see their own pattern. Arx keeps the record most venues discard: the thesis behind a trade, what actually happened, and how a signal or a replicated Participant performed for you over time. Wins and losses become a personal, checkable history — the raw material for getting better, and the quiet reason disciplined users stay. Most platforms optimize for your next trade; this optimizes for your next year.

Signal Radar is the public face — a continuously updated, informational view of how tracked wallets are positioned, where market structure is diverging, and a plain-language daily read, with a gated tier for individual indicators as they form. The watchtower of the citadel: informational analytics made visible, never a recommendation.

Lucid is the same intelligence, conversational and strictly factual: screen-aware, grounded in the same public inputs, able to walk a user through a market, a position, or a Participant's record in plain language. It explains and structures; it never predicts prices or sizes positions. It is one rung of a deliberate trust ladder — self-guided, copilot-assisted, delegated — where the user chooses how much to hand over, every grant is scoped and revocable, and unattended autonomy is not on the ladder today.

The opportunity opened on an industry that profits when its customers lose. Arx's answer is protection built into the product, not sold as an upsell: the full downside on screen before you act, and — where you replicate another Participant — a loss floor the platform cannot switch off.

Before you commit, the ticket puts the consequences on screen: the margin you pay, where liquidation would sit and how far away, the risk and reward on the stops you set, and what is left after fees and slippage. Beside the fields, Arx surfaces the on-chain liquidation walls — where the nearest risk actually sits — as context, never a recommended value: it shows you the truth and never suggests your stop or your size. Every number is venue-native, computed from price, side, and leverage, not a platform attestation, so the survival math is visible before stress, not discovered during it.

Copy Activity carries a mandatory drawdown stop — a loss limit that closes the session the moment your equity falls past it, with no "off" toggle in the interface, set with the Participant's own worst 90-day drawdown shown right beside the slider. A position-size limit caps any single mirrored order, and a cap on leverage clips any Participant position above your ceiling. These rails run inside your own custody and pause replication before a Participant's blowup can run past your limit. It is the literal inverse of a house that keeps winning while you cannot stop: protection the platform cannot switch off, because the user sets it and no one at Arx can override it.

Enter by studying, not funding: browse markets, study Participants, and watch Lucid work before any key or deposit exists; an email login creates a self-custodial embedded wallet, exportable at will — no seed phrases, no blockchain know-how. Fund and exit in one operation each, fees shown, not hidden — because getting money out is the number-one trust moment in this industry. And reach real breadth — crypto, equities, commodities, and FX on the Connected Protocol — chosen for liquidity and on-chain verifiability, with tradable depth surfaced as a first-class fact so a headline price never hides a market you cannot actually reach.

The right question before funding is not "do I like this app" but "what can this app do to my money — and what can't it?" Arx's answer is a narrow architecture. Keys and collateral live at the edges — the wallet and the venue — never in the middle. Arx transmits orders and reads public state; it does not custody, match, or settle.

Notice what kind of trust this is. Arx does not ask you to trust a balance sheet or a brand — the two the last era ran on. It offers a different stack: cryptographic trust (your keys, your funds), data trust (records and hit rates you can recompute), social trust (real traders with their own capital at risk), and plain-language trust (the downside shown before you take it). Each is checkable alone; stacked, they are why "don't trust, verify" is the architecture here, not a slogan.

The first question a trader should ask any platform is "are you betting against me?" Arx's answer is structural: it does not run the order book, does not internalize flow, and does not custody funds — it cannot take the other side of a trade. Revenue comes from activity and service, on-chain and inspectable.

The first line is proven practice, not projection: front-ends transmitting orders through this same builder-code mechanism already earn real revenue on the venue — a single wallet application has collected roughly $19.7M in builder fees (Grayscale, 2026).

One nuance, stated plainly: copying carries no performance fee today. The first-order business model layers on top and is Planned — trading clubs monetized by membership, and Arx's own strategy vaults, each with its own disclosed, opt-in terms defined with counsel. None is offered today; each would be reflected in the Terms & Conditions before launch. Payment for a service performed, never a tax on copying.

Early users make a platform real, and Arx plans to recognize them through a loyalty points program, audited like everything else here: points would accrue for genuine, sustained engagement and tenure, with anti-farming curves and Sybil defenses on the same public ledger, attribution wallet-anchored and inspectable, each season's parameters published before it opens, and accrual never conditioned on trading volume, deposits, or order activity. Stated plainly — Arx has issued no token, and this document promises none. Points would confer no ownership, no revenue right, and no expectation of profit; they are a loyalty mechanic and a tamper-evident record of who showed up first, not an investment.

The alignment is not perfect, and we say so: fee-on-volume rewards activity, not user success alone. But the sharpest conflict in retail markets — the intermediary that wins when the client loses — is structurally absent here, and the fee that remains is public. That is the difference between an incentive to serve and a license to harvest.

Everything here is Planned — not offered today, and introduced only if and when reflected in Arx's Terms & Conditions and cleared with counsel.

The synthesis grows richer: the verifiable on-chain base overlaid with technicals, fundamentals, news, and event-market signals, plus sharper reads of how tracked wallets are positioned, how crowded a strategy is getting, and when a trader's behavior drifts. Trust deepens with it — more records made checkable, honest divergence on every copy, the scorecard kept public. These are the layers a venue can never build and the ones that improve with every user, so they come first.

As traders with checkable records, and the people who follow them, accumulate, the next step is the trusted community around them: trading clubs where such a trader shares rationale on a paid membership. This is the first-order business model — earned, not scheduled: it opens only once the trust and intelligence beneath it are proven, reflected in the Terms & Conditions, and cleared with counsel.

The mandatory protections and a walk-in door ship from day one; from that proven core Arx widens both — more markets the protocol lists (prediction and event markets, structured products, options), more languages and funding rails for the people the old system shut out, and protection extended across each new product. The same self-custodial account grows into where capital lives and compounds — strategy vaults, and in time investing, payment, and lending, each built by a specialist protocol with Arx curating it, never becoming the bank. Eventually the same rails let a person's own software agent act on their behalf, inside limits they set.

Three rules set the pace: nail the category before adding the next; ship nothing that cannot be checked on-chain; and promise nothing here — no market, no milestone, no token, no fee — that is not already true.

Four crafts, one build — strategy, product, growth, infrastructure. Every commitment this document makes has an owner.

Arx and its parent group are backed at the group level by SoftBank, GSR Ventures, Tiga Investments, Provident Capital, Unicorn Venture, and eGarden Venture — each name shown subject to that investor's display consent.

Partners: Hyperliquid (settlement venue); Privy (embedded self-custodial wallets); further infrastructure partners — data, on-ramp, and security — are announced as each integration goes live rather than pre-announced here. Arx is in closed beta as of this draft; traction metrics — funded accounts, transmitted order volume, and retention evidence — will be published at public launch rather than projected here.

Arx binds itself to publish, at the named moments: the security auditor, scope, and report, together with contract addresses, before public launch; traction metrics with closed-beta retention evidence at public launch; and, if a points programme launches, each season's parameters before that season opens. Market figures in this paper are re-verified at each revision; regulator and market-structure figures were last checked against source documents on 2026-07-02.