Every PUD is the top of a column of real assets: pull the unit apart and its exact contents float out — layer depth is vault weight. Four legs held on-chain today, one admitted and waiting. Hover a layer to read what your coin actually owns.
Five steps from basket to bank-run-proof. Click through — or just scroll the story on the right.
One PUD is a fixed, published list of quantities — and every item on the list is an asset that exists on TON with real liquidity, because when you mint, the protocol buys those exact quantities on-chain, at that moment. No synthetic exposure, no posted rate to trust: the price of a PUD is what its contents cost, and full backing is the only guarantee that survives contact with a bad day.
Target after Revision R2. The RWA leg is admitted but dormant — until its jetton ships on TON, its 10% rides in USDT, so today's holdings are 50 / 25 / 15 / 10.
The composition is constrained by buyability, on purpose: a leg we cannot purchase and hold on-chain does not belong in the basket. The price of that discipline, on twenty years of data: +2.8%/yr of real growth (1.74x), 7.3% yearly wobble, a worst drawdown of -15.3%, and a deepest dip of -8.4% below entry. The multi-currency core returns as euro, yuan, franc and pound jettons ship. Full math in the mechanism.
Between revisions, the list above is a constant: nobody — including us — can change what your PUD contains. Once a year (and when a waitlist jetton ships), an AI advisor proposes the next quantity list, arguing from the strongest available evidence: twenty years of purchasing-power backtests against the world consumption index, live DEX depth on TON, drawdown and wobble under stress.
The advisor's power ends at the proposal. Hard rules are enforced by the contract, not by judgment: only TON-buyable assets; USDT 40–60%, gold 15–35%, no leg above its liquidity; at most 20% of the vault traded per revision; every proposal published with its full reasoning, then a 7-day timelock before it takes effect — and no revision, ever, can touch your right to redeem. Strong advice, weak power.
Revision R2 · published 2026-08-30 — the machinery’s first use. The advisor’s proposal: admit an RWA leg at 10%, funded from USDT (50 → 40). Acceptance criteria: tokenized real estate first, tokenized T-bills acceptable; activation is gated on a jetton with real DEX depth on TON — no qualifying jetton exists there today, so the leg is admitted but dormant, its slice riding in USDT. Backtested with the US home-price index standing in: 1.85× real over 20 years (+3.1%/yr), 7.4% wobble, −14.5% worst drawdown — a little more growth than the active basket, with a slightly shallower worst fall. Turnover on activation: 10% of the vault, inside the 20% bound.
Deposit USDT at the posted PUD rate. Every unit in existence was paid for in full — minting is depositing.
Open mint →Redeem to USDT at the posted PUD rate, any time. The exit has no queue, no gate, and — by design — no pause switch.
Why runs can't happen →Computed from public CPIs, FX and asset prices; posted on-chain with a 5 bp/day drift cap. External, auditable, boring.
Watch it live →