Economics

Where the yield goes, who holds uSOL, and why each side shows up.

The flow of SOL

1. A borrower deposits LST worth V SOL and borrows B uSOL, with B <= V x borrow limit.
2. Each epoch the pool rate rises by g (0.01724% for JitoSOL at epoch 1050). The position's shares gain V x g lamports of value.
3. On the next settle that gain is swept: V x g comes off the debt, and LST worth V x g moves from the position to the collateral's redemption buffer.
4. Any uSOL holder can burn uSOL and take the same number of lamports of LST from the buffers.

So every uSOL minted is matched first by collateral worth at least twice as much (at a 50% limit), and over time by LST in the buffer as the debt behind it is paid.

No fee, no interest

The program takes nothing. There is no borrow interest, no origination fee and no cut of the yield: 100% of the swept yield goes onto the debt. The only costs are Solana transaction fees and the rent for your position account, which close_position returns.

Time to zero

With the position synced every epoch, the debt falls by V x g per epoch, so:

epochs to zero = ceil( B / (V x g) )

The SDK's projectRepayment matches the program's own count to within one epoch in every row. Once the debt is zero, each position still held its starting SOL value in LST.

Measured inputs, from mainnet at epoch 1050: JitoSOL g = 0.0172378% per epoch, mSOL g = 0.0175876% per epoch, one epoch = 115,505 seconds (average of epochs 1040 to 1050). Annualised: JitoSOL about 4.8%, mSOL about 4.9%.

Who holds uSOL, and why

Most borrowers who want to spend will swap uSOL for SOL or a stablecoin. That sale is what puts uSOL in the market; the redemption buffer is what gives buyers a reason to take it.

Redemption keeps uSOL near one SOL

redeem(amount) pays exactly amount lamports of LST at each pool's own rate. If uSOL trades below one SOL, buying it and redeeming earns the gap, which pushes the price back up. If uSOL trades above one SOL, borrowing and selling earns the gap.

Redemption takes from every collateral's buffer in proportion to its SOL value, so every redeemer receives the same mix. A redeemer cannot pick the strongest LST and leave a weaker one for whoever comes last.

Limits

Each collateral has a debt ceiling: the most uSOL debt that may be open against it at once. Every uSOL in circulation is backed one of two ways:

• by an open debt, which is itself backed by collateral worth at least twice as much at a 50% limit, or
• by the buffer, when the yield has already paid the debt that created it. Those uSOL can be redeemed for the LST the sweep put there.

The launch setup uses a 50% borrow limit and a 60% unwind line for both JitoSOL and mSOL. The suggested debt ceiling at launch is 10,000 uSOL per LST, raised as the uSOL/SOL market deepens. There is no protocol fee on the yield.

Day one, said plainly

uSOL is one token for every LST, so it never splits into one token per pool. On day one the redemption buffer is empty: it fills at about 0.017% of all borrowing collateral per epoch. Until it has filled, the way out of uSOL is a uSOL/SOL pool, seeded at launch (the suggested seed is 100 uSOL against 100 SOL, with the uSOL borrowed at 25% against 400 SOL of JitoSOL). Jupiter routes through that pool once it has liquidity, so any wallet swaps uSOL to SOL in one step.

That makes the real price of the loan, for a borrower who sells their uSOL, the discount uSOL trades at under one SOL while the buffer is young. With 10,000 SOL of collateral at a 25% average loan to value, the buffer gains about 1.7 SOL per epoch, about 19% of the uSOL supply a year; at 50% average, about 9% a year. A lower average loan to value means a tighter peg.

When one LST is worth less

Each LST is valued at its own pool rate, never at par with another. If mSOL trades 2% under its pool rate, a loss there lands first on mSOL borrowers' own equity (a 2% rate drop moves a 50% position to 51%), not on JitoSOL borrowers and not on uSOL holders. Redemption pays from every buffer in proportion, so the last redeemers do not end up holding only the weaker LST.