How Utilisation Sets the Price of Borrowing in DeFi

In an on-chain lending market, the interest rate is a function of one variable: what share of the deposited capital is out on loan. That share is utilisation, the function is published in the contract, and the result is that anyone can compute the rate for any level of demand before committing capital. This is a genuinely different arrangement from any conventional credit market, and its consequences are worth working through.

The curve and its kink

The shape is consistent across the major designs: a gentle slope up to a target utilisation, then a much steeper one beyond it. The point where the slope changes is usually called the kink.

Below the kink, rates rise slowly with demand. This is the intended operating region — enough capital is lent to be productive, enough remains idle for lenders to withdraw.

Above the kink, rates rise sharply. The steepness is not a revenue decision; it is a defence. High utilisation means little idle capital, which means a lender wanting to withdraw may be unable to. The steep segment exists to make borrowing painful enough that borrowers repay and lenders are attracted, both of which restore the buffer.

So the kink is the protocol stating where it thinks the boundary lies between productive lending and a liquidity problem, and the steepness after it states how aggressively it will act to get back. Two protocols with different kinks have made different judgements about how much withdrawal risk to impose on lenders, and that judgement is legible in a chart.

What a lender is actually holding

A deposit into such a market is not a fixed-term instrument. It is a claim on a pool whose availability depends on borrower behaviour.

At moderate utilisation, withdrawal is instant. At very high utilisation, it is not available until someone repays or someone else deposits. The deposit did not change; the conditions did.

This is the single most misunderstood property of on-chain lending, and it is not a bug. The alternative — guaranteed instant withdrawal — requires holding capital idle, which means paying lenders less. The curve is the mechanism that trades a small probability of delay for a materially better rate, and it does so transparently.

The practical consequence for anyone sizing a position: check utilisation before depositing capital you might need at short notice, and treat a market persistently sitting above its kink as one where withdrawal is a negotiation with the borrower base.

The spread, and where it goes

Borrowers pay one rate; lenders receive a lower one. The gap has two components worth distinguishing.

The first is arithmetic rather than a fee: interest is paid on borrowed capital and distributed across all deposited capital, including the idle portion. At half utilisation, a borrow rate is spread across twice as much deposited capital, so the lender rate is roughly half. Nobody took anything.

The second is the reserve factor: an explicit share of interest diverted to the protocol's own reserves, which exist to absorb bad debt. This is a real charge, it is published as a parameter, and a higher reserve factor means lenders receive less in exchange for a larger buffer against loss.

Confusing the two produces the common complaint that a protocol is taking most of the interest. Usually it is taking a modest published share, and the rest of the gap is the idle capital doing what idle capital does.

Reading the parameters that matter

Beyond the curve, a handful of published values determine how the market behaves under stress.

The collateral factor sets how much can be borrowed against a given asset — lower for volatile collateral, which is the protocol's judgement about how fast the price can move before liquidation completes.

The liquidation threshold and penalty set when a position becomes liquidatable and what the liquidator earns. The penalty has to be large enough to make liquidation worth executing during congestion and small enough not to be punitive, and that balance is where several protocols have learned expensive lessons.

Supply and borrow caps limit total exposure to a single asset. A cap is the clearest signal a protocol gives about which of its listings it does not fully trust.

Before depositing

• Look at utilisation relative to the kink, and treat anything persistently above it as a withdrawal constraint.
• Read the reserve factor to know what share of interest is diverted, rather than inferring it from the spread.
• Check the collateral factor and cap on the assets backing the loans, since that is what stands between you and bad debt.
• Compare the rate against what the same asset earns elsewhere, including the withdrawal terms attached to each.
• Remember that every parameter here is changeable by governance while your capital is deposited.

The last point is the one that connects this to everything else about on-chain lending. The curve is transparent and computable, and the authority to change it sits with whoever controls the protocol.