Every currency in history has been managed by someone. U is managed by an optimizer that publishes its reasoning, cannot leave its bounds, waits a week in the open, and holds no key to the exit. This page is the whole of its authority, written down.
Most stable assets aim at a currency. U aims at a basket of consumption — the world index W(t), a five-bloc CPI composite that the protocol measures itself against but never holds. The optimizer's objective is simple to state: of all the baskets it is allowed to hold, pick the one whose backtested value in W(t) terms grew fastest.
That objective alone would buy nothing but bitcoin. What makes U a unit rather than a bet is the second half of the sentence: of all the baskets it is allowed to hold. The charter is where the stability lives.
Growth is the objective. The charter is the constraint. Neither is negotiable at runtime.
These are checked in code against any proposal before it can be queued. A proposal that violates one is not rejected by judgment — it simply cannot execute.
| Bound | Limit | Why it exists |
|---|---|---|
| Realized volatility | 8.0%/yr | A unit that swings like an asset is not a unit. This is the line between the two — and R3 moved it, deliberately and in public. |
| Worst drawdown | 20% | Bounds the pain a holder can be asked to sit through before the basket recovers. Widened from 12% at R3 because, without sovereign debt, nothing in the universe could reach the old bound. |
| Idle cash | none | Dollar stablecoins were removed at R2: a leg that neither earns nor hedges only melts. |
| Government debt | none | Removed at R3 to detach the unit from any single sovereign. This was the expensive one — bills were the only drawdown control the tokenized universe offers, and the charter above had to widen to absorb their loss. Redemptions are paid in kind, or from the euro leg. |
| Turnover per revision | ≤ 20% | Caps slippage and stops the protocol from reinventing itself in one step. |
| Revision frequency | ≤ 1 / quarter | An optimizer that can trade constantly is a trading desk, not a denominator. |
| Timelock | 7 days | Anyone who dislikes the next list can leave with the current one first. |
| Redemption rights | untouchable | No revision, no upgrade and no emergency may gate, pause or reprice the exit. |
Per-leg ceilings are part of the charter too, and they are set by how deep each tokenized market actually is — not by how much we like the asset. They appear in the table below.
Every row is a real token on Ethereum. The returns are twenty-year purchasing-power figures for holding that asset alone — which is the honest way to see why the optimizer weights them as it does.
| Asset | Tokens | Ceiling | Real / yr | Wobble | Worst fall | In R1 |
|---|---|---|---|---|---|---|
| Loading the live universe… | ||||||
Every stable thing on this list melts, and every growing thing shakes. The protocol's entire job is the trade between those two facts, made inside the charter.
The left column is what your U contains right now — fixed until a revision executes. The right column is the optimizer re-run against this morning's data. The distance between them is the case for the next revision, and it is public every day.
Loading…
Purchasing power, twenty years, the active list.
The same list over the recent decade.
What the unit's contents cost right now.
An AI advisor drafts the next quantity list from the evidence: rolling purchasing-power backtests against W(t), realized wobble and drawdown, live DEX depth per leg, and any newly qualified tokenized asset. The full argument publishes with the proposal — data, method, and the losing candidates.
The charter in §2 is enforced on-chain. The advisor argues; the contract decides what is arguable. It cannot widen its own bounds — that takes a separate, timelocked governance change that redemption rights are excluded from.
An accepted proposal waits seven days in public. Anyone who prefers the current list may redeem it in full before the change. On execution day the vault trades to the new quantities in one published rebalance, inside the turnover cap.
What the advisor can never touch: your right to redeem, the per-leg invariant, the fee schedule, or the pause switch that does not exist. Its entire authority is the next list of quantities, inside bounds, after a week of daylight. If it is ever wrong, it is wrong in public, slowly, and within limits — the only kind of wrong a monetary instrument survives.
The charter bounds the backtested drawdown at 12%. A backtest is not a guarantee, and the future is free to be worse than the past.
Nobody picks assets by conviction. The objective and the bounds are fixed; the optimizer is the only thing that moves, and only on schedule.
Most of these tokens are younger than the assets they hold. The backtests model the exposure — gold, bills, equities — not the token's own track record.
A six-leg mint on Ethereum L1 is expensive. Caps are sized to live pool depth, and the build spec is explicit about where the costs land.