Arkis: Why are interest rate models such a big deal in DeFi?
Oleksandr: They set the cost of borrowing and the return for lending — essentially the engine of on-chain credit markets. Whether it’s a retail yield farmer or an institutional fund running arbitrage, every trade depends on how predictable these rates are.

A: What’s the standard approach today?
OP: Protocols like Aave, Euler, and Morpho use utilization-based models, where rates rise as pools fill up. This keeps lending pools solvent and follows basic supply-demand logic: scarce liquidity should cost more.

A: Sounds good. So what’s the problem?
OP: Utilization-based models are great for solvency, but they also introduce noise and unpredictability. Small jumps in utilization can trigger sharp spikes in borrow costs. For example, a looper might see plenty of liquidity, borrow it, and suddenly cross what’s called the kinkpoint, where the borrow rate suddenly gets much steeper. This scenario would instantly make the trade unprofitable.

A: So what’s the Arkis approach?
OP: At Arkis, we anchor borrowing rates to the expected yield of the collateral, rather than just utilization. There are three main levers in the model. The first is the target spread, which makes sure there’s always enough margin for a trade to make sense. The second is the minimum lending rate, which guarantees that lenders are earning at least a fair baseline return. And finally, there’s utilization feedback. We still include utilization, but it plays a much smaller role. It’s there as a smoothing signal, not the main driver of rates. 

A: Why does this difference  matter?
OP: Because the collateral yield is the actual economics of a trade. For example, Pendle principal tokens have an implied APY known in advance. If the collateral yields 10%, asset managers only need confidence that their borrow costs won’t exceed that. By tying borrowing costs to yield, Arkis directly aligns rates with profitability.

A: What about liquidity providers? Do they benefit?
OP: Yes, they do. Lenders earn a minimum baseline return, plus yields that rise as collateral yields rise. On top of that, we use a fee-splitting mechanism. When utilization is low, lenders capture a larger share of the fees. When utilization is high, more of the fee burden shifts to borrowers. That way, both sides are fairly compensated as market conditions change.

A: How does this compare to existing models?
OP: Other platforms can sometimes give borrowers really attractive spreads, but those spreads can be volatile. In certain conditions, they even turn negative and force unwinds.

With Arkis, even at very high utilization, the spread stays positive. In a Pendle PT USDe looping trade performed in September 2025, Arkis had the lowest standard deviation of spread and the highest stability score across platforms. For institutions, that predictability matters.

For institutions, stability and predictability matter more than chasing occasional outsized spreads.

Q: What are the implications for the future of DeFi?

OP: Markets scale when borrowing is predictable, because that makes capital more cost-efficient for everyone involved. If strategies can run without the risk of sudden cost spikes, capital providers are more willing to commit deeper liquidity, and asset managers can operate at institutional scale. The Arkis model is designed to provide that kind of predictability, keeping spreads stable so asset managers and liquidity providers remain aligned as markets grow.