Kamino vs Jupiter vs Nolus: Leveraged SOL Costs

Every leveraged position carries four costs: getting in, holding, getting out, and being wrong. Interest gets the attention because it is the number every app prints on its front page. It is also the cost least likely to decide how a trade ends.

This article prices one trade on four Solana lending apps, Nolus, Kamino, Jupiter Lend, and Save, using each app's published terms and its live rates on September 17, 2026. Where the wider industry is the better yardstick, it says so and widens the view to Aave, Morpho, and Compound, the largest lending apps on Ethereum. Over a month, what you pay to hold differs by about the price of a lunch. The gap in how far SOL can fall before anything happens to the position, and in what that costs when it does, is where the apps really part ways. On both counts Nolus comes out ahead of the other three.

The Same Trade, Four Apps

Say you put in $1,000 and open a 2.5x long on SOL. You now hold $2,500 of SOL, $1,500 of it borrowed, with SOL at $101. All four apps route the swap that gets you in and out through the same Solana liquidity, so that cost is about the same everywhere and is left out of the table, as is the network fee of a few cents per transaction.

Rates move daily on three of the four apps. On Nolus, the rate a position gets is fixed the moment it opens.

Getting In and Out

Nolus charges nothing to open a position and nothing to close one. The protocol's margin sits inside the interest rate shown on the position form, so there is no second charge to find. Jupiter Lend charges nothing either way too. Kamino's leverage product costs about two cents each way in flash loan fees, and Save takes 0.10% to borrow the USDC, which is $1.50 on this trade. For scale, Aave charges 0.25% on swaps made through its own app.

None of these numbers decides the trade. But it is worth noticing that on Nolus the fee column is simply zero, and that the rate you see is the whole price.

A Quiet Month

Suppose SOL goes nowhere for thirty days. You close where you opened, and the only thing you paid for was time.

Nolus charged the most interest that month: $9.60, against $5.68 to $7.00 on the other three. There is also a second line in the other three apps' favor: the SOL you bought with the borrowed money keeps earning the app's supply rate while it sits as collateral. At the snapshot's rates that yield more than covered the interest on Kamino and Jupiter Lend, so those positions ended the quiet month about two dollars ahead, and Save roughly broke even. Nolus paid its $9.60 in full, because the asset behind a Nolus position earns nothing.

So the honest number for a quiet month is nine to twelve dollars on a $2,500 position, depending on the app, in the other apps' favor. Part of that is the price of a fixed rate. The rest is yield that collateral on Nolus does not earn.

What the fixed rate buys is a number that cannot move on you. The other three charge a variable rate that follows how much of the pool is borrowed, and every one of them publishes a curve with the same shape: nearly flat until the pool is roughly 80 to 94% lent out, then close to vertical. Each curve also says how high the rate goes when the pool is fully lent out: 15% on Jupiter Lend, 37% on Kamino, and 150% on Save. On this trade that is $18, $45, and $185 for the same thirty days, the light part of the bars above. Utilization moves with the market, not with you, and it tends to jump in exactly the weeks a leveraged position is under pressure. A Nolus position takes the rate it opened at and keeps it, whatever the pool does afterwards.

A short trade in a quiet market is cheaper on a variable rate. A position held through a busy one is where the fixed rate earns back its premium, and either way you know the worst case on the day you open.

A Bad Month

Now suppose SOL slides instead.

The first app to act is Save. At $84, a 17% drop, it starts selling your collateral. Kamino follows at $81, down 20%, and Jupiter Lend at $71, down 29%. A Nolus position is untouched at all three of those prices. Its first liquidation line sits at $67, a 33% drop from where you opened.

What happens at that line is a partial liquidation. The protocol sells only as much SOL as it takes to bring the position back to a healthy level, at the market price, and the position stays open. If SOL keeps falling, the next slice goes at about $62 and a third at about $57. All three prices are on the position form before you open.

Kamino, Jupiter Lend, and Save also sell in slices rather than closing the whole position at once. The difference is what each slice costs you. Nolus charges nothing on a partial liquidation: the collateral goes at the market price, and that is the whole cost. Jupiter Lend takes $10 of every $1,000 it sells. Kamino takes $10 at the line and up to $100 the further past it your position has slipped. Save takes $49 to $69.

The picture does not change outside Solana. On the three largest Ethereum lending apps, the penalty per $1,000 sold runs $45 to $75 on Aave, $44 on Morpho's cbBTC market, and $50 to $100 on Compound. And on Kamino, Save, Aave, and Compound, part of that penalty is kept by the protocol itself, so the app that sells your collateral also earns from selling it.

Put the two together and the bad month reads like this. Everywhere else the selling starts earlier, and every slice carries a fee, part of which the protocol keeps. On Nolus it starts later, and a slice costs what the market pays for it.

Nolus stays first if you push each app to its limit. Nolus and Jupiter Lend both go to 5x, the highest of the four, while Kamino stops at 3.8x and Save at 2.9x. At those limits a Nolus position survives an 11% drop before its first liquidation, Save 9.7%, Jupiter Lend 5.9%, and Kamino 1.3%.

For Lenders

The pricing difference has a mirror image on the lending side, and here the view widens again. On Nolus the pool rate goes to lenders in full; the protocol's margin is charged to the borrower on top of it rather than taken out of it. Kamino, Jupiter Lend, Save, and Aave keep between 10 and 50% of the interest borrowers pay before anything reaches lenders. Compound keeps the gap between its borrow curve and its supply curve. Morpho takes nothing at the market level, though the vault curators between a lender and the market take 0 to 25%.

What You Are Choosing

The four apps are not pricing the same thing differently by accident. Three of them lend you money at whatever the pool charges today, pay you a yield on your collateral, take a cut when they have to sell it, and take a cut of the interest before it reaches lenders. Nolus charges nothing to get in or out, quotes one rate and holds it, draws the liquidation line further out, sells at market when it is crossed, and passes the pool rate to lenders whole. What it does not do is pay yield on the collateral. That, plus the fixed-rate premium, is the lunch.

In a quiet month that trade-off costs you about the price of a lunch. In a bad one it is the difference between selling that starts at $84 with a fee on every slice and selling that starts at $67 with nothing on top. Rates on all four apps will have moved by the time you read this. The shape of the choice will not.

Sources

• Nolus: Liquidations & risk and Borrowing
• Kamino: Liquidations and Fees & Interest Rates
• Jupiter Lend: Borrow and Liquidation Mechanism
• Save: Liquidations and Fees
• Aave V3: Pool contract, liquidationCall and Swap features
• Morpho: Liquidation
• Compound III: Liquidation
• Interest rates, pool parameters, and liquidation terms were read from each app's public API and on-chain state on September 17, 2026, with SOL at $101. Where an app's documentation and its on-chain parameters differ, the on-chain value is used.