Stabolut
How it works

The mechanism, in full.

Where the yield comes from, the limits the reserve runs under, and exactly how you get your money back.

USB is a dollar you can hold, send and redeem, backed by a reserve of digital assets held on-chain. The reserve is put to work. What it earns, after our commission, flows to USB holders. This page is the long version — the mechanism, the limits it runs under, and exactly how you get your money back.

What happens when you deposit

a hundred dollars, followed end to end

1

You send $100 of eligible digital assets. The contract prices them against independent oracle feeds and mints you 100 USB — no more. The amount that can be minted is capped on-chain against the value that demonstrably arrived, so no key, and no person, can create USB that isn't backed.

2

Your $100 joins the reserve and is deployed across the strategies below. It is never lent to us, never lent to a single counterparty, and never sits in a bank account.

3

The reserve earns. We take a commission on what it earns — not on your principal — and the rest accrues to USB holders.

4

You ask to redeem. The price is fixed at the moment you ask, not at the moment you're paid. Positions are unwound in an orderly way and you receive full backing value.

That last point is worth pausing on. In most designs the redemption price is struck when the payout happens, which means the first person out gets a better price than the last. That turns a nervous week into a race. Fixing the price at the moment of request removes the prize for winning the race — there is no advantage to panicking first.

Where the yield comes from

Every point of return has someone on the other side paying it. We can name the payer in every case. The reserve runs three kinds of strategy, and inside each one it spreads across a number of independent venues rather than concentrating in the highest-paying option.

Strategy Mechanism · who pays · what would break it

Over-collateralized lending

We lend the reserve into established on-chain money markets — Aave, Morpho, Spark, Fluid, and the Sky savings rate. Every borrower has posted more collateral than they've borrowed, and if that collateral falls toward the value of the loan, the protocol sells it automatically — the position closes before it can become a loss to the lender. This is the oldest and most tested mechanism in decentralised finance, and the closest thing the market has to a base rate.

Who pays

borrowers who want liquidity without selling the assets they hold.

What would break it

a flaw in the lending protocol's code, or collateral that falls faster than it can be sold. Both are covered on the risk page.

Fixed-rate contracts

Some holders of variable-rate positions would rather know their return in advance, and will accept a lower expected number in exchange for certainty. We take the other side, mainly through Pendle, where a yield-bearing position is split into a principal part and a yield part that trade separately. Buying the principal part at a discount fixes the return on the day of purchase. It is known when we enter, not guessed.

Who pays

holders of variable rates buying certainty.

What would break it

needing to exit before maturity in a market that has moved against the position — which is why maturities are kept inside the redemption window.

Market-neutral strategies

Positions structured so that the direction of the market doesn't decide the outcome — the return comes from spreads and market structure rather than from prices rising. A position that gains when prices rise is paired with one that gains when they fall, and what's left is the fee the market pays whoever is willing to hold both sides. In practice: curated market-neutral vaults on Morpho and Euler, and basis positions on major derivatives venues, with the collateral held in off-exchange settlement rather than on the exchange itself.

Who pays

leveraged traders paying to hold their positions, and takers crossing the spread.

What would break it

a venue failing, or a market dislocation wide enough that the two sides stop offsetting cleanly.

On naming venues

These are the kinds of venue the reserve uses, not a fixed list. The live allocation — every protocol, every position, every address — is published on the reserve dashboard when USB goes live, and it changes as conditions do. What doesn't change are the limits in the next section.

Why this matters more than the headline number

Each of these answers to a different kind of demand. Borrowers want liquidity. Rate-payers want certainty. Traders want leverage. Those three appetites don't disappear at the same time, which is why a reserve spread across all three earns something in most conditions — rather than everything in one condition and nothing in the next.

The limits the reserve runs under

These are enforced by the system, not by anyone's judgement on the day. A rule a person can waive under pressure is not a rule.

Concentration

No single protocol holds more than a set share of the reserve

No single curator manages more than a set share, counted across every venue they touch — not per venue, which is how concentration hides

No single yield-bearing token is held above a set share, and there is a cap on the total held in that category

No position larger than a set fraction of what that venue could realistically absorb in a day — because size you cannot exit is not liquidity

What we will not hold

Nothing whose backing includes managers we cannot see or name

No asset where a single key can create new supply

No position in any lending market that prices a fluctuating asset at a hardcoded one dollar

Nothing we can't exit inside the redemption window

No exposure to USB's own backing — no circular structures, ever

Each of those exists because of a specific failure elsewhere in this market. The last section of this page names them.

How fast the reserve becomes dollars

A redemption promise is only as good as the reserve's ability to turn into money on the day it is asked. So the reserve is built as a ladder, and every position is placed on it according to how long it takes to exit — not according to what it pays.

placeholder — final is a raytraced glass render (brief 03)
Minutes

Idle stablecoins in custody, and savings positions with no waiting period. This is the layer that absorbs an ordinary day.

Within a day

Blue-chip lending positions and unwind capacity on hedge venues.

One to three days

Paced unwinds, curated vaults with contractual withdrawal terms inside the window, and short-dated fixed-rate positions.

Three to seven days

Positions with a native cooldown, longer maturities, and anything else that takes real time to exit.

The line we don't cross

Nothing enters the reserve with exit terms longer than a week. Ever. A position that takes a month to exit is not part of a reserve backing a redeemable dollar — it's a bet that nobody asks.

The ladder is stress-tested against a large share of supply being redeemed inside two days, with haircuts applied for the things that actually go wrong in a crisis: lending markets seizing up as everyone withdraws at once, and forced sales going through at worse prices than the screen shows. The point of publishing the shape of the ladder is that you can check the redemption promise against it yourself, rather than taking our word for the promise.

Redeeming, step by step

Three ways out, all at full backing value.

Immediate

Served from the instant layer, priced with a fee that reflects what it costs to serve you now. The fee rises as that layer drains — so the people leaving first pay for the liquidity they're consuming, rather than the people leaving last.

Standard

Free, through a published window. The price is fixed at the moment of request.

Under stress

The window can extend, but only within a hard maximum written into the contract, and only when two specific measured conditions are both met. Nobody at Stabolut decides this.

Three things that can never happen

We can't stop processing. Even in a declared emergency the contract keeps paying out at a floor rate. Freezing withdrawals is how a loss event becomes a collapse — it is the single most common last act of a failed stablecoin, and it is not available to us.

We can't change the terms on you. Redemption terms carry a long mandatory delay before any change takes effect, inside bounds the contract fixes. A dollar whose exit terms can be rewritten overnight is not a dollar.

We can't quietly gate you. The dashboard publishes coverage — how the pending queue compares to available liquidity — never a queue length. A visible queue is a starting pistol; every run in financial history has been made worse by one.

What we don't do

The clearest way to describe a design is often to name what's absent from it.

No banks.

No accounts, no banking relationships, nothing in an institution that can freeze it.

No RWAs.

No treasury bills, no private credit, no tokenised funds. The reserve is crypto-native and verifiable on-chain.

No token.

The yield comes entirely from what the reserve earns. There are no emissions, no unlock schedule, and no reason for the return to collapse when a token price does.

No discretion.

No strategy that depends on someone's judgement about where the market is going. If we can't see it, we don't run it.

No hidden managers.

Every counterparty is one we can name.

What this looks like in a bad year

We target 5–7% a year. In a flat, low-activity market, the realistic number is closer to 3–4% than 7%. Lending rates compress, spreads narrow, and there is no honest way to manufacture the difference.

The distinction that matters

In that environment holders earn less — but USB's backing is unaffected. Yield variability is a return risk, not a solvency risk. The two get confused constantly in this market, usually right before something breaks, and keeping them separate is the point of the whole design.

We would rather set that expectation now than explain it later. The risk page covers what happens in the cases that are genuinely dangerous.

Why we're built the way we are

Five synthetic dollars died in the last seventeen months. Not one died from the mechanism.

Stream Finance

Undisclosed off-chain managers. Unverifiable backing. $93M lost, $285M of contagion.

Elixir

Sixty-five percent of its backing lent to a single counterparty. Down 98%.

Resolv

One compromised key minted 80 million unbacked tokens. Eighteen audits — all of the contracts, none of the infrastructure that actually failed.

Usual USD0++

Changed redemption terms overnight. Solvent the entire time. Never recovered.

They died from unverifiable backing, from concentration in one counterparty, from a single key, and from rewriting the terms after people had already committed. Not from the arithmetic.

Meanwhile the largest crypto-backed dollar shrank by almost sixty percent — and its peg never moved. It redeemed two billion dollars in twenty-four hours, at full value, during the worst crash in years. The design held while the business shrank, which is exactly what a well-built reserve is supposed to do.

The architecture works.
The operators failed.

That is why the limits above are enforced by the system rather than by policy, why mint capacity is capped on-chain rather than trusted to a signer, why there is no counterparty we can't name, and why redemption terms can't be changed without long notice. Each of those rules is a specific failure we've read the post-mortem on.

Be there when USB launches.

Launch dates, reserve reports, and each new currency as it goes live.

Get updates