If you found this useful, the full breakdown with math and simulation is on Substack → [mirador.finance] More deep-dives incoming.

Join our Telegram to get notified first → [https://t.me/Miradortelegram]

INTRODUCTION

Impermanent loss is an inherent risk for liquidity providers (LPs) in Automated Market Makers (AMM) pools. However, GammaSwap, the first on-chain perpetual (perp) options decentralized exchange (DEX), seeks to convert this risk into “Impermanent Gain” by allowing traders to bet on the tokens’ volatility via liquidity loan from AMMs.

How can GammaSwap turn the impermanent loss (IL) to impermanent gain (IG)?

Is hedging IL strategy on GammaSwap really financially viable?

This article is mainly to examine these two questions by deeply diving into its sophisticated mathematical formulas and simulating the hedging strategy’s profitability on GammaSwap.

KEY TAKEAWAYS

In this article, we will clearly examine the effectiveness of hedging IL strategy on GammaSwap via 3 main following sections:

• Section 1: Understanding about Impermanent Loss – the problem that LPs mostly concerns when providing liquidity into traditional AMMs like Uniswap V2

• Section 2: Explore how GammaSwap turns impermanent loss into impermanent gain and the essence of hedging strategy

• Section 3: Examine the viability of GammaSwap’s hedge strategy via realistic numbers

  ---

SECTION 1: THE PROBLEM - IMPERMANENT LOSS IN TRADITIONAL AMMS

1/ What Is Impermanent Loss?

For many people, providing liquidity in an AMM pool sounds like a safe strategy in the DeFi game because you don’t need to bet on price direction while still earning trading fees from swappers.

However, one strange-but-inevitable thing happens: even when the asset goes up and you make money, you still earn less than if you had done nothing and just held.

This hidden gap is called “Impermanent loss”.

So, theoretically:

Impermanent loss (IL) is the difference (loss) between providing liquidity (LP) to the AMM compared to simply holding (HODL) those tokens.

It’s called impermanent loss because your loss can be offset or recovered when the token price returns to its initial price, and not a complete loss of money. In practice, IL represents the opportunity cost of LP, trading off the risk of price volatility for the profit from fees.

IL happens only when the token’s price in the AMM pool changes from the time of the initial deposit. It means:

• In case of no price change, no IL will occur and LPs will earn a positive profit from the trading fees.

• In case of price change, if the trading fee cannot cover IL, the profit of LPs may turn negative, and vice versa.

IL is not caused by a bug, hack, volatility crashes or any bad executions but integrated directly into the algorithms of AMM because of its automatic rebalance mechanism, where the pool’s LPs are forced to sell their assets that are increasing in value and buy those decreasing in value, while token holders are not.

The most familiar spot where IL often occurs is on Constant Product Market Maker (CPMM) pools like Uniswap V2, where the most basic invariant is applied:

• X: the number of token A (i.e. $ETH)

• Y: the number of token B (i.e. $USDC)

• K: constant number

When ETH increases, the traders outside will buy ETH (swap USDC for ETH) from the pool as they see an arbitrage opportunity to sell it elsewhere for a profit. It means the AMM pool will automatically sell the increasing ETH, decreasing the quantity of ETH and increasing that of USDC, while their value is equal. And vice versa, when ETH decreases, the pool will buy more ETH. These arbitrage trades are exactly how the AMM continuously rebalances the pool, without the need for an order book or any intermediary.

Let’s compare the difference between simply holding tokens and LPing them into a Uniswap V2 pool.

Say you own 1 ETH + 3,000 USDC.

A) Case 1: HODL only

• Initial ETH price = $3,000 → Your total token value = $6,000

• When ETH price increases to $4,000 → Your total token value = $7,000

B) Case 2: LP in a Uniswap V2 pool (ETH/USDC) with the invariant: x*y=k

In this case, the initial parameters are: x = 1 (ETH); y = 3,000 (USDC) and k = 3,000 (constant)

• Initial ETH price = $3,000 → Your initial total token value = $6,000

• When ETH price increases to $4,000, some of your ETH will be automatically sold off as the AMM mechanism explained above. And now your token balance will be (i.e.):

• ~0.866 ETH (x_new)

• ~3,464 USDC (y_new)

• The new invariant (k) = 0.866*3,464 ≈ 3,000 (constant)

  → But your new total token value = $6,928 – slightly lower than HODL position ($7,000)

→The difference value between HODL and LP is $72 (~1.03%) = IL

That’s how IL was born and affects LP’s profit, causing LP to earn less than simply holding spot assets.

2/ How to calculate the IL by a simple formula?

As the example above, the IL value can be calculated by working out the difference between HODL and LP position.

So, theoretically:

And if we illustrate the IL via HODL and LP position on a graph, the IL is the blank space between the black line (LP) and the blue dashed line (HODL).

The impermanent loss stays between LP and HODL position

And converting to percentage, here’s what we got:

Where:

• V_LP: the new value of LP position at a new token price

• V_HODL: the new value of spot holding position at a new token price

But if we need to calculate the value of both HODL and LP position to work out that of IL, it would take a lot of time to reach the result. Meanwhile, the key point is that the IL only occurs when the token price changes.

So, the quickest way is to find a mathematical formula which can reflect the IL percentage through the price change only.

Let’s break them down into 2 different cases for clearer comparison.

Case 1: Simply HODL ETH and USDC in the wallet

Your wallet balance is:

• x0 ETH

• y0 USDC

The value of ETH balance is equal to that of USDC balance (x0*P0=y0).

And initial ETH price by USDC = P0 = y0/x0

Then, the initial total value is:

As ETH price changes (up or down), the new ETH price is P1, where:

(k = price change ratio; k≠0)

With the token balance unchanged, the new portfolio value will be:

Case 2: LP in a Uniswap V2 pool - 50/50 ratio with no fees (hypothetical)

Your LP balance is:

• x0 ETH

• y0 USDC

Initially: ETH price by USDC = P0 = y0/x0

As the pool is set with 50/50 ratio of token balance, so:

The initial total LP value is:

As Uniswap V2 is an invariant AMM where x*y=constant

It means:

( x,y: new ETH & USDC balance)

And the price in the pool is always:

This leads to:

Now we need to solve this to find the new token amount in the pool when ETH price changes.

First, apply [2] into the invariant:

Replace this value of x into [2], we have a new function:

Because the new LP value after the price changes is:

Substitute [3] into V_LP, this leads to:

As defined in the above formula:

Let’s replace [1] and [4] into the above IL formula:

Rewrite this formula, here’s the final formula:

Now, we will try using this official formula to work out the IL percentage by the price change ratio of ETH.

In the first example, we assumed the ETH to increase from $3,000 to $4,000. It means:

So:

(exactly the same as the above result)

That’s how we check the IL via the token price change ratio without the need of working out different positions’ value. And the IL curve will look like this in a graph, as reflected by Uniswap below.

Source: Uniswap docs

Obviously, higher volatility leads to greater impermanent loss, an unavoidable cost for AMM LPs. Although it cannot be removed, it can be hedged, an approach GammaSwap enables by turning LP exposure into tradable volatility positions.

SECTION 2: THE SOLUTION - HEDGING IMPERMANENT LOSS ON GAMMASWAP

1/ What is GammaSwap?

Source: DefiLlama

GammaSwap protocol is an on-chain perp options DEX launched in 2023, founded by Daniel Alcarraz, Roberto MarTinez and Devin Goodkin. In August 2025, this DEX’s total value locked (TVL) has reached its all-time high value at over $30M. Even though the TVL of GammaSwap, currently around $3M, is quite low, it doesn’t mean this protocol is a failure.

As a niche market allowing volatility trading (betting on the token price fluctuations) primarily via AMM liquidity and earning fees as an LP, in order to use GammaSwap, users need to learn about its products, especially main features on this DEX.

Simply put:

• If you think the price will fluctuate significantly (up or down) → use GammaSwap to open positions by borrowing liquidity from an available AMM pool (comes with borrow rates) and earn a profit

• If you think the price will be less volatile → you can act as a LP for others to “bet” on, earning swapping fees but may suffer from IL.

In terms of perp options trading, users on this DEX can trade the volatility of assets by borrowing liquidity from an AMM pool (primarily DeltaSwap - a Uniswap-like pool of GammaSwap). There are 2 main kinds of positions which traders can choose to open.

Straddle: Bet the token in the pool will be highly volatile

A straddle order on GammaSwap is a borrowed LP position where collateral is drawn at a 50:50 ratio between two tokens in the pool (the opposite of normal liquidity supply). In other words, the trader “borrows” LP tokens and exchanges them for two equal underlying assets without a need of rebalancing collateral. This position can be understood as Long Volatility position (but not the same as the directional Long below).

If you expect the token price to pump or dump sharply (high volatility) but you don’t know which direction (profitable either up or down), you can create a Straddle position, pay daily borrowing cost (also known as theta). If theta is lower than the price change, you earn a profit, and vice versa.

This position is suitable when the market is about to have major events (such as big news), or you want to hedge IL for your LP position because Straddle is the perfect “opposite” of LP, which can be implied as Short Volatility (but not the same as the directional Short below).

Example:

The price of ETH is currently $3000, and tomorrow, FED is going to announce future interest rates. As a trader, you believe that after or even before the news is announced, ETH will definitely fluctuate. Then, you can open a Straddle.

If the price pumps to $4000 or dumps to $2000, you make a large profit. But if it remains stable, you lose money slowly due to the borrow rate which you must pay to the LPs.

This position can be used to hedge IL in case you are an LP, too. In the next section, we will dive more into this.

B) Directional (Long/Short): Bet on the price direction, similar to Call/Put in traditional Options but different from Perp Futures

If you have a clear view on the direction of a token, you can open a Long to bet up or Short to bet down, but still benefit from volatility. This may sound like Perp Futures positions on other DEXs, but in fact, a long position on GammaSwap can still yield high profits even when the token drops sharply as reflected in the picture above. The remaining position (Short) is still similar.

Example:

You are bullish on ETH, opening a Long position at 3000. If Price goes up to 4000, you’ll get a quick profit. If it dumps to 2000, you will lose initially, but if volatility is high, GammaSwap can help reduce losses or even make a small profit.

Unlike Straddle, creating a directional position requires you to borrow liquidity and rebalance collateral to the bias ratio:

• Long (Up bet): Rebalance to ~60% volatile asset (i.e, ETH) / 40% stable (i.e, USDC) → similar to Call option.

• Short (Down bet): Rebalance to ~60% stable / 40% volatile → similar to Put option.

This kind of position is suitable when you are clearly bullish or bearish, but want to add leverage from volatility.

2/ How can GammaSwap solve the IL problem?

As discussed above, GammaSwap can’t completely eliminate IL but instead attempts to compensate by allowing LPs to hedge their own IL risk.

LPs can hedge IL on GammaSwap by opening a Straddle position (long gamma – volatility exposure).

In essence, traders opening a Straddle or any perp options positions on GammaSwap are borrowers, who take a liquidity loan from LPs in the AMM pool.

Source: GammaSwap Blog

• If you want to trade the volatility (i.e. Straddle), you need to deposit your collateral (ETH + USDC) to borrow LP tokens from GammaSwap

• These LP tokens will then be burnt, and you will receive reserve tokens (ETH + USDC) in the pool, which are still kept in the smart contract

• Now you hold the real asset, and if the price fluctuates sharply, LP tokens will lose value due to IL but the ETH and USDC you hold in the contract are not rebalanced. Besides, you will receives an NFT (ERC-721) as a receipt for position identification in the GammaSwap smart contract

• Then, when you close your position, GammaSwap will use the ETH and USDC in the contract to buy back LP tokens on AMM and your NFT will be burnt

• Because LP tokens are cheaper now, you can buy more LP tokens than when you burned, and the difference value becomes your profit, or called Impermanent Gain (IG).

So the interesting point is that the more volatile the reserve tokens are, the more LP tokens can be redeemed. This is how GammaSwap turns IL into IG.

Hence, the IL is the perfect opposite of IG, or:

(This negative calculation is like swapping LP position to HOLD position)

In our example above, we work out the result of IL as -1.03% when the ETH price pumps from $3,000 to $4,000. Applying the formula here, it means that LP will lose ~1.03% of their total token value while traders earn ~1.03% as a profit if opening a Straddle with the same position size as LP.

You can see the difference between these two parameters via the graph below, where the blue line (IG) is symmetrically opposite to the red line (IL).

Taken together, we can see that GammaSwap is like a lending market for LP tokens, which is essentially a zero-sum game (the loss of providers is the gain of borrowers/traders).

Now, let’s examine the effectiveness of this hedging strategy via a profit simulation based on realistic figures on GammaSwap.

SECTION 3: THE SIMULATION - FEASIBILITY OF HEDGE IL STRATEGY ON GAMMASWAP

Understanding about IL, IG and how the hedging strategy works by theory is not enough but its viability must also be considered.

In this section, we will try simulating the possibility of the hedge strategy compared to LP and HODL strategy only based on realistic numbers shown on GammaSwap UI, together with the explanation behind each formula of parameters for better understanding.

On GammaSwap UI, we will try making a hedge strategy via Straddle position to work out the estimated profit with the help from GammaSwap doc.

1/ Scenarios

Let’s outline three different positions to build a profit simulation accounting model.

Case 1: Straddle position in the DeltaSwap pool (WETH/USDC)

Here’s what we got from opening a Straddle position (borrowing liquidity from WETH/USDC pool) on the UI:

• ETH price = $3,205

• Initial deposit: $500

• Leverage (Debt/collateral ratio): 50%

• Position size: $16,797

• LIUs (Liquidity Invariant Units): 148.363

• Time to liquidation: 127 days

• Projected LTV (Loan-to-Value): 97.15%

• Theta (The daily cost to maintain a position): 0.6369%

• Borrow rate (12h): 6.9%

• LIU

For Straddle profit/loss calculation, LIU is one of the most important indicators to pay attention to since it represents the debt of the position. As traders need to borrow liquidity from the Uniswap V2-style pool with the basic invariant: x*y=k and 50:50 ratio, it means the debt value will be split equally into WETH and USDC.

Hence, here’s how we calculate the LIU:

Instead of trying to find out the exact borrowed quantity of WETH and USDC, which will take a lot of time, we will just need to utilize the ETH price and position size as a quicker way.

As explained above, the new HODL position value is calculated by the following formula:

From this:

(exactly the same as LIU shown on the UI)

B) Case 2: LP in the DeltaSwap pool (WETH/USDC)

And here’s what we got when LPing into WETH/USDC pool with the same position size as Straddle:

• Initial deposit: $16,797

• Supply APY: 13.5%

With this initial capital and ETH price at $3,205, based on the invariant and the 50:50 ratio of the AMM pool, here’s the token balance of our portfolio:

• Initial ETH: 2.62085

• Initial USDC: 8,398.63

In the first case, we deposit $500 to open a Straddle position of $16,797. In this case, however, if we only provide liquidity without opening any perp options (straddle) orders on GammaSwap, our LP strategy needs to include the initial deposit which is supposed to stay in your portfolio.

In other words, LP strategy is the total equity of your portfolio in case of providing liquidity only (no hedge).

Hence:

C) Case 3: HODL ETH and USDC spot position

In case you hold ETH and USDC spot position only, with the same token balance as follows:

• Initial ETH: 2.62085

• Initial USDC: 8,398.63

Your initial total portfolio value is:

2/ How to calculate Straddle PnL?

The success of hedging IL strategy depends on the profitability of Straddle position. If the IL of a LP is $5, in order to compensate this loss, a straddle position should make an IG of $5, too.

To calculate the profit/loss (PnL) as well as IG, we compare the current portfolio value with the portfolio value at the time the position was opened. In other words, the current PnL of the position will be the difference between the current and initial portfolio values.

Where:

• V: The portfolio value at the current time

• V0: The portfolio value at the initial time

A Straddle position on GammaSwap is essentially a portfolio consisting of short LP tokens and buying the underlying tokens that the LP token represents (i.e. WETH & USDC).

Therefore, the position value at the current time is the difference between the token value you are currently holding and LP token value you are borrowing.

Replace [5] into V:

Where:

• x: the quantity of the first underlying token (i.e. WETH)

• y: the quantity of the second base token (i.e. USDC)

• P: the current price of the first token relative to the second token (e.g., the price of WETH in USDC)

• LIU: the liquidity invariant of the amount of LP tokens borrowed at the time of opening the position

• t: the time (year) elapsed since the borrowing

• r: annual borrow rate of LP pool

• e^rt: represents the debt growth due to compounded interest

And this is the position value at the initial time (t=0):

Replace the V and V0 formula into the PnL’s above. Here’s the final result:

NOTE:

In this simulation model, we assume t=0 in order to isolate price-driven effects and evaluate the effectiveness of IL hedging. Time-dependent factors such as borrow rates (r) and time to liquidation (t) are therefore excluded from the model.

So, in this case:

Now, we shall apply this formula for hedge strategy simulation below.

3/ Hedging IL Strategy - Simulation Results

A successful hedge strategy is one where its return equals that of spot holdings and the IL of LPs can become the IG of traders (or the LP themselves if they open a trade). So, our goal is to make:

no matter how ETH fluctuates.

Now, it’s time to dive in the accounting model. Below is our result of the hedge strategy’s returns compared to LP and HODL only.

Where:

Example:

At a price increase of 10% for ETH ($3,524) (yellow-highlighted row), here’s the result of these parameters:

Applying the same formulas to other ETH price cases, it’s obvious to see that the more ETH price fluctuates, straddle position always earn higher profit than LP and HODL.

Specifically, looking at the IL, IG and two orange-highlighted columns, we can clearly see that the IL of LP has become the IG of straddle traders, making the hedge strategy equal to HODL strategy.

Or, it also means that the loss that LPs suffer from LP strategy can be compensated by opening a Straddle position with the same size to hedge and make it their profit gain. Adding it all up, LPs will suffer from no loss (IG + IL = 0) and then, their realized profit depends on their swapping fees.

So, from the above mathematical results, we can draw out a line graph to compare as below.

In this graph, you can’t see any particular lines for spot holdings because the return of hedge strategy (green line) is completely aligned with that of spot holdings (red line), whereas LP strategy is a downward curve (yellow curve) due to the impact of IL.

From the detailed simulation of the accounting model as explained above, together with all these assumptions based on realistic numbers on GammaSwap UI, the hedge strategy can be made successful to turn the impermanent loss of LP into the impermanent gain of traders. In simple terms, LPs can use this strategy to hedge their unavoidable IL and earn a positive profit from their borrow fees.

Yet, please note that actual results on GammaSwap may vary due to market conditions, funding dynamics, liquidity, and protocol-specific risk parameters. The above is our simulation based on real-world parameters, and it could become a reality if the price of ETH fluctuates significantly, as in the scenarios where these assumptions change.

CONCLUSION

While real-world outcomes may vary depending on market conditions and execution, the analysis and simulations above suggest that GammaSwap offers a compelling approach to managing impermanent loss. Impermanent loss is an inevitable event that every liquidity provider has experienced. Although it’s impossible to completely eliminate this loss, through GammaSwap, this exposure can be actively hedged, and under favorable market conditions, even transformed into a source of profit.

Disclaimer

This article focuses on explaining the mathematical design and core mechanics behind GammaSwap, particularly how impermanent loss can be hedged and transformed into impermanent gain through LP-based volatility positions. The analysis is intended to clarify the underlying logic of the protocol rather than to evaluate performance, compare protocols, or recommend specific trading strategies.

All models, formulas, and simulations presented are simplified in order to illustrate key concepts under idealized assumptions. Actual results on GammaSwap may vary due to market conditions, funding dynamics, liquidity, and protocol-specific risk parameters. This article is for educational purposes only and should not be considered financial or investment advice.

Ahead of the GammaSwap V2 launch, we will return with further deep-diving articles exploring other key components of the new architecture.

Leave a comment

INTRODUCTION

Liquidation is the safety mechanism that keeps a lending protocol solvent. For borrowers, it is the moment collateral is forcibly sold and the loss becomes permanent.

Aave V3 relied on a cap based liquidation design, using hard thresholds and fixed incentive parameters to keep execution simple and reliable on the EVM. The tradeoff is that the outcome can be coarse. A position that is only slightly unsafe can be liquidated in a size that is much larger than what is needed to restore safety, and fixed incentives can concentrate MEV competition at the liquidation boundary.

Aave V4 proposes a different approach. It replaces static caps with a Target Health Factor solver that sizes repayment to a chosen safety baseline, and it replaces fixed incentives with a state based bonus curve that adapts as risk deepens.

This raises two practical questions for borrowers and liquidators.

• Why does Aave V3 often liquidate more collateral than seems necessary for small breaches?

• Does Aave V4 actually make liquidation more proportional, and how does it change liquidator competition and MEV dynamics?

This article answers these questions by reconstructing the exact sizing and incentive mechanics, then validating the key behaviors using empirical Aave V3 on chain liquidation data.

KEY TAKEAWAYS

In this article, we examine Aave liquidation mechanics in three parts.

• Section A: Why cap based sizing in Aave V3 leads to overshoot, especially near the threshold

• Section B: How Target Health Factor sizing in V4 mathematically enforces a chosen post liquidation safety baseline

• Section C: How the state based bonus curve creates a Dead Zone and a Race Band, reshaping liquidator incentives and shifting MEV from a single trigger point to a competitive band

  ---

A/ WHY AAVE V3 LIQUIDATES MORE ASSETS THAN NECESSARY?

1. The Core Issue: Liquidation as Debt Recovery, Not Punishment

This section is a sizing question: when a position becomes liquidatable, how much debt should be repaid in the liquidation transaction ?

Aave V3 does not compute the minimal repayment needed to restore a position to a chosen safety buffer. Instead, it sets an upper bound on repay size, the Close Factor. In short, V3 provides a cap, not a target based solver.

Because the Health Factor (H) is continuous, a position at H just below 1 is only slightly unsafe, while a position at much lower H is deeply unsafe. A cap based rule does not scale smoothly with this risk depth. Within a given region, the cap is the same, and liquidators have a direct incentive to repay close to the maximum allowed size to maximize profit.

This is the core mechanism behind over liquidation in V3: post liquidation safety is not explicitly targeted. It is an emergent outcome of a coarse repay cap, profit maximizing liquidator behavior, and a liquidation bonus that scales collateral seizure with repay size.

2. The Mathematical Flaw of a Static Rule

2.1. State Variables and the Meaning of H < 1

To demonstrate why this “one size fits all” rule is inefficient, we define the core state variables (calculated in USD):

• D : Total Debt.

• V : Raw Collateral Value.

• c : Liquidation Threshold (the risk parameter of the asset).

• C : Eligible Collateral, where: C = c×V

• H : Health Factor, the ultimate safety metric.(this is not calculated in USD)

The Health Factor is calculated as:

A position becomes eligible for liquidation when H < 1 . Economically, this means the risk adjusted collateral value C is no longer sufficient to cover the debt under the protocol’s risk rules.

2.2. Close Factor in Aave V3 Is Piecewise Constant, Not Truly Fixed

Aave V3 does not apply a single fixed close factor in all states. Instead, it uses a piecewise rule that still behaves like a static cap within each region.

• Up to 50 percent of total debt can be liquidated when health factor is above 0.95 and both collateral value and debt value are at least 2000 USD each.

• Up to 100 percent of debt can be liquidated when health factor is 0.95 or below, or when either collateral value or debt value is below 2000 USD.

So the execution rule is still a cap based approach:

where f is 0.5 in the normal region and can jump to 1.0 in the severe risk or dust region.

This matters for the argument. The over liquidation problem is most visible in the near threshold region, roughly 0,95<H<1 , where the cap can still be very large relative to what is minimally required to restore safety.

2.3. Why a Cap Creates Overshoot

Aave V3 does not compute the minimal repay amount required to reach a target safety level. It only sets an upper bound. In most profitable scenarios, liquidators rationally choose to liquidate as much as the cap allows, because profit scales with size.

To see the mechanical effect, write the post liquidation health factor for a simplified single collateral single debt position. Let b be the liquidation bonus. If the liquidator repays R , then collateral seized is approximately (1+b)R in USD terms, and the post state becomes

This formula shows why sizing matters. As R increases, debt decreases, but collateral also decreases and it decreases faster than debt because of the bonus term.

This chart compares two repay sizes for the same position state.

• The yellow bar is the minimum repay needed to restore the position to a reasonable safety buffer (𝐻=1.05).

• The red bar is the maximum repay allowed under the V3 50% close factor cap.

• The diamond marker shows the resulting post liquidation health factor (𝐻𝑝𝑜𝑠𝑡) if the liquidator uses the cap.

Key takeaway: when the position is only slightly unsafe (Scenario A), the cap still permits a large repay, pushing (𝐻𝑝𝑜𝑠𝑡)​ well above the target buffer (overshoot). That higher buffer may reduce re liquidation risk, but it comes at a clear borrower cost: more debt is repaid and more collateral is seized than is strictly needed to restore safety.

3. Why V3 Used a Cap Based Model

3.1. Why a Simple Cap Worked Well on Ethereum Mainnet

Aave V3 was designed for a world where Ethereum mainnet transactions could be expensive and unreliable during market stress. In those moments, the protocol needs liquidations to execute quickly and predictably.

A close factor cap fits that requirement because it is extremely simple to compute. It only needs a few basic arithmetic operations.

• Take the user’s total debt D

• Multiply by a constant f

• Get the maximum repay size Rcap=f×D

This is what people mean by O(1) . It does not mean “better” or “more fair”. It only means the computation is constant and small. No matter how complex the position is, the contract does roughly the same tiny amount of work.

Important clarification: O(1) explains why a constant cap is attractive, but it does not explain why the cap is 50%. Any constant cap, 30%, 50%, 70%, is still O(1) . The exact number is a design trade-off.

• Higher cap: removes risk faster, fewer steps, better for solvency in fast crashes

• Lower cap: reduces borrower loss from overshoot, but may require more repeated liquidations

• The protocol also needs liquidators to show up, which pushes incentives toward simple, reliable execution

So the real reason V3 used a cap based design is reliability under stress, not mathematical necessity.

3.2. Why the Constraints Became Less Binding Over Time

Later, the environment changed in two practical ways.

First, more user activity moved to Layer 2 systems. Many L2 networks have lower fees and higher throughput than mainnet, which makes more complex on chain logic less risky to run in practice.

Second, the Dencun upgrade and EIP 4844 reduced the cost of sending rollup data to Ethereum using blobs. That lowered costs for many rollup transactions and made it easier for protocols to design richer execution logic in L2 environments.

The key point is not that computation becomes free. The key point is that the original constraints that strongly favored ultra simple liquidation logic became less severe. When costs and execution risk drop, a protocol can shift from “always do the simplest thing” toward “do the more accurate thing”, such as dynamically sizing the repay amount to hit a chosen safety target.

This is the design space Aave V4 is trying to use: replace coarse caps with a target driven mechanism, while keeping liquidations reliable enough that liquidators still participate during stress.

Example : How Alice Loses Capital

We can quantify this capital destruction through a step -by -step calculation of a minor risk scenario.

Initial state for Alice:

• Debt (D): 10,000 USD

• Collateral (V): 13,000 USD

• Liquidation Threshold (c): 0.80

• Close Factor (f): 0.50

• Liquidation penalty: 0.05 (5%)

Step 1: The Risk Trigger

The market price drops slightly, reducing V to 12,250 USD.

Observation: Alice is just barely below the 1.0 threshold. She only needs to repay a tiny fraction of her debt to become safe again.

Step 2: The V3 Execution

Instead of calculating what Alice actually needs, the liquidator applies the maximum 50 percent Close Factor rule to maximize their own profit:

Debt repaid (R): 10,000×0,50 = 5,000 USD

Collateral seized from Alice (including the 5 percent bonus): 5,250 USD.

Step 3: The Overshoot

We calculate the ledger state post liquidation:

• New Debt (D): 5,000 USD.

• New Collateral (V): 7,000 USD.

• New Eligible Collateral (c): 7,000 × 0,80=5,600 USD

• The new Health Factor is:

The key point is not that a higher post liquidation health factor is always bad. A buffer above 1.0 can be desirable, especially for volatile collateral, because it reduces the chance of immediate re liquidation. The issue in Aave V3 is that this buffer is not explicitly targeted by the protocol. It emerges mechanically from a coarse close factor cap and profit maximizing liquidator behavior. In Alice’s case, repaying the full 50% pushes the position from H=0.98 to H=1.12, which may be a larger buffer than necessary for a marginal breach, and therefore increases collateral sold and the liquidation penalty paid.

4. Empirical Evidence from On-chain Data

To test whether overshoot is a rare edge case or a recurring pattern, we analyze liquidation events on Aave V3 Ethereum using Dune Analytics. The chart below compares the Health Factor immediately before and immediately after each liquidation event.

Horizontal Axis (X): Health Factor before liquidation (Hpre).

Vertical Axis (Y): Health Factor after liquidation (Hpost).

Interpreting the plot is straightforward. Points near Hpre≈0.99 represent positions that were only slightly below the liquidation threshold. If liquidation sizing were tightly calibrated, we would expect many of these points to land just above 1.0 after liquidation. Instead, a visible share of them jump to much higher post liquidation health factors, with many exceeding 1.12. In other words, the data suggests that a meaningful number of near threshold liquidations result in a post state that is substantially safer than the minimum required to exit liquidation.

This pattern is consistent with the mechanism described earlier: within the close factor capped region, liquidators have an incentive to repay close to the maximum allowed amount, and the liquidation bonus causes collateral seizure to scale with repay size. The result is a larger buffer, but also higher borrower cost, especially when the initial breach was smal

A static cap is a reasonable engineering choice when reliability under congestion is the primary constraint, but it also makes liquidation outcomes coarse. The evidence above does not imply that every overshoot is bad. In volatile markets, a higher buffer can reduce the chance of immediate re liquidation. The core takeaway is more specific: in Aave V3, the post liquidation buffer is not explicitly chosen by design for each risk state. It emerges from a fixed cap plus profit maximizing execution.

This is the design motivation for a dynamic liquidation engine in Aave V4. Instead of relying on static constants to indirectly produce a safety buffer, V4 aims to compute repayment sizes that target a chosen post liquidation health factor, and then adjust incentives so liquidators still participate reliably across different risk states.

B/ THE V4 SOLUTION: TARGET HEALTH FACTOR SIZING

Part A showed that Aave V3 is cap based: it defines a maximum repay size per liquidation, but it does not compute the repay size needed to reach a specific post liquidation safety buffer. This creates room for overshoot, especially when the position is only slightly below the threshold.

The chart below reframes the same liquidation events through a counterfactual question: for each real Aave V3 liquidation, what repay amount would Aave V4 have requested if the position were liquidated using a Target Health Factor solver. Each dot is one on chain V3 liquidation event.

Horizontal axis X: the maximum repay cap that Aave V3 would allow for that specific event state.
Vertical axis Y: the repay amount required by a V4 style target solver to restore the position to the chosen target safety level.

Points below the diagonal line indicate cases where the target solver would repay less debt than the V3 cap permits. In those cases, the required collateral seizure would typically be smaller, implying less borrower capital sold than a cap maximization outcome. In our sample, the points cluster below the diagonal, consistent with the claim that target based sizing can reduce overshoot near the threshold.

To understand why this happens, we now unpack the V4 solver itself. The core mechanism is the function _calculateDebtToTargetHealthFactor, which computes the debt to repay so the post liquidation Health Factor is steered toward a predefined target.

1. The State Variables

To establish the equation, we define a loan position with all parameters normalized to USD ( expect Health Factor) :

· D: Total current debt.

· V: Raw collateral value.

· τ: Liquidation Threshold. This is the risk parameter of the asset used to calculate safety (e.g., 0.80). It is distinct from the Loan to Value ratio used when initially opening a position.

· C: Eligible collateral value, where C=τ.V

· H: Initial Health Factor, where H = C/D. The position is eligible for liquidation when H<1,0.

· T: Target Health Factor configured by the system (e.g., 1.05). The algorithm must find a way to return H to this exact T baseline.

· b: Liquidation Bonus. While this bonus varies depending on risk depth, during the sizing step of a specific execution, this bonus is locked based on the initial state (H_pre) and enters the equation as a constant: b=b(H_pre).

· R: The exact Repay Amount (the target variable).

When a liquidator repays a debt amount R, they receive the equivalent raw collateral plus the bonus: R×(1+b).

Consequently, the eligible collateral (C) of the position will decrease by a corresponding amount: τ×R×(1+b).

The post liquidation ledger state of the position becomes:

· New Debt (D): D-R

· New Eligible Collateral (C’): C-τ.R.(1+b).

· New Health Factor (H’): (C^’)/(D^’)

2. Deriving the Repay Formula (R)

The system must solve the equation so that the new Health Factor perfectly matches the target (H’=T).

We apply algebraic transformation to isolate R:

The core formula executed within the Solidity codebase is:

The Mathematical Existence Constraint:

For this function to return a valid result (a positive repay amount), the denominator must be strictly positive:

This constraint is not just a technical requirement for division. It has a practical meaning. The term τ×(1+b) captures the effective haircut on collateral once the liquidation bonus is included. If governance sets the target T too low relative to this haircut, the solver can return an invalid or negative repay amount, or the transaction may revert. From a system design perspective, Aave governance must configure the target T and the maximum bonus limits so this inequality holds across supported collateral types and market conditions

The chart compares two sizing rules. The red dashed line is the V3 repay cap in the 50% close factor regime. The blue curve is the V4 target solver, the repay amount needed to reach T=H’ at each starting health factor H .

If the blue curve is below the red line, V4 would repay less than the V3 cap, so the borrower typically loses less collateral in near threshold liquidations.

Does target sizing create liquidation cascades?

A natural concern is repeated liquidations during a prolonged drawdown. If collateral prices keep falling, a position can be liquidated back to T, then quickly drop below 1 again and be liquidated again. This is a real trade-off, but it should be framed as frequency versus severity. A cap based system tends to produce fewer but harsher liquidations, while a target based system can produce more frequent but smaller liquidations by avoiding cap sized overshoot. Governance mitigates churn by choosing a target T that provides a meaningful buffer above 1, balancing the risk of immediate re liquidation against the borrower cost per event.

3. Safety Checks: Repay Clamp and the Dust Rule

The algebraic result R is referred to in the codebase as _debtToTarget. However, the smart contract does not execute this raw number immediately. It must route the value through two layers of boundary checks to ensure ledger safety.

Layer 1: The Execution Clamp

The actual executed liquidation volume (R_actual) is bounded by the total outstanding debt and the amount the liquidator is willing to cover:

Meaning:

• debtToTarget is the solver output, what the math would like to repay

• debtToCover is the liquidator’s chosen maximum, how much they are willing to repay in this call

• debtReserveBalance is the user’s actual outstanding debt in the repaid reserve, the real debt balance the contract can close, including interest accrual and rounding effects

Layer 2: The Dust Debt Exception

The Target HF rule is the default calculation method. However, it is constrained by the execution economics of the Ethereum network, specifically Gas fees.

If the remaining debt after liquidation (D -R_actual ) is too small (strictly less than a constant like DUST_LIQUIDATION_THRESHOLD, e.g., 1,000 USD), the future Gas cost to liquidate this tiny remainder will exceed the bonus value. No bot will execute it, leaving the protocol with toxic bad debt.

To prevent this, the codebase establishes an exception rule: If
D - R_actual<1,000 USD, the protocol abandons the Target HF formula. It allows the liquidator to repay 100 percent of the outstanding debt to completely wipe the position clean.

This design is smart for two reasons:

• It protects the protocol by ensuring the liquidation result is actually cleanable in practice, not just correct on paper. Dust is a real failure mode because bots follow economics, not intent.

• It also protects borrowers in the common case by enforcing a repay amount closer to what is needed for safety, instead of letting liquidators automatically push to a large cap.

The trade-off is that the dust path prioritizes system hygiene over borrower smoothness. When dust would be created, the protocol is willing to be more aggressive and close the position fully.

4. Execution Trace and V3 Comparison

To prove the precision of the function, we run a step -by -step execution trace.

Initial position parameters:

· Total Debt (D): 100,000 USD.

· Raw Collateral (V): 122,500 USD.

· Liquidation Threshold (τ): 0.80

· Eligible Collateral (C): 122,500×0,8=98,000 USD

· Initial Health Factor (H_pre): 98,000/100,000=0,98 (Below 1.0, triggering liquidation).

System parameters:

· Target (T): 1.05.

· The locked bonus at H_pre=0,98 is b=0,06.

Important note on T=1.05: this is not a universal constant. It is a governance configurable parameter, set at the Spoke level in V4. In this example, 1.05 is an illustrative choice for a reasonable safety buffer

Step 1: Calculate debtToTarget (R)

Step 2: Run through boundary checks

Remaining debt : 100,000 - 34,653 = 65,347 (USD)

Because this value is vastly larger than the dust threshold (1,000 USD), the final approved liquidation volume remains locked at 34,653 USD.

Step 3: Post liquidation ledger state

· New Debt (D’): 100,000 - 34,653 = 65,347 (USD)

· Collateral seized: 34,653×1,06=36,732 (USD)

· New Eligible Collateral (C’): 98,000-(0,8.36,732)=68,613)

Rechecking the final Health Factor:

The position is perfectly locked onto the T=1,05 target baseline.

Quantitative Comparison with Aave V3:

For this exact same position, the legacy V3 model allowed a maximum execution limit of 50 percent of the total debt (50,000 USD). Because liquidation bots are programmed to maximize absolute profit, they routinely submit executions right at this ceiling limit. The result is an artificially inflated Health Factor that destroys user capital.

By embedding control theory mathematics into the core Smart Contract, the _calculateDebtToTargetHealthFactor function in V4 caps the required intervention at 34,653 USD. This provides direct technical proof that V4 utilizes the absolute minimum force necessary to secure the system, completely eliminating the excess capital destruction zone inherent to the legacy model.

C/ VARIABLE BONUS AND LIQUIDATOR COMPETITION

When Target Health Factor sizing compresses the repay amount R(H) near the threshold, liquidators earn less absolute revenue per liquidation. Gas and execution costs do not shrink at the same rate. So the protocol needs a second lever to keep keepers participating, especially when positions are only slightly unsafe.

Aave V4 uses a state based Dutch auction mechanism to do this. There is no orderbook of bids. Instead, the protocol publishes a deterministic bonus schedule b(H). As the position becomes riskier and H falls, the bonus increases. The auction emerges implicitly because multiple keepers observe the same public schedule and choose when to execute based on their own cost structure.

If this is an auction, what is the clock? The market is the clock. H moves because collateral prices move and because debt grows over time from interest. That market driven drift acts like the auction tick. The reward increases until it reaches a keeper’s break even point.

1. The Piecewise Bonus Curve

The purpose of this subsection is to define the exact incentive schedule that keepers observe, because this schedule is the economic engine behind the state based Dutch auction. In the next steps, we will plug b_net(H) directly into the keeper profit function π_i(H) to derive the break even point (H^*)_i​, the Dead Zone, and the Race Band.

The codebase manages the raw liquidation bonus b(H) via three parameters: the maximum bonus cap (b_max), the maximum risk threshold (H_max), and the minimum bonus (b_min).

Intuition first. The schedule has three regions:

· If 𝐻≥1.0, liquidation is not allowed, so the bonus is zero.

· If 𝐻_𝑚𝑎𝑥<𝐻<1.0, liquidation is allowed and the bonus increases linearly as 𝐻H deteriorates, starting from 𝑏_𝑚𝑖𝑛.

· If 𝐻≤𝐻_𝑚𝑎𝑥​, risk is severe and the bonus is capped at 𝑏_𝑚𝑎𝑥​.

Mathematically, b(H) is strictly defined as a piecewise function:

However, the actual bonus captured by the liquidator is discounted by the protocol fee (f_protocol). The net bonus function is locked by the following formula:

Near the threshold (𝐻≈1), the repay amount 𝑅(𝐻) produced by target sizing is small, so keeper revenue is highly sensitive to the bonus level. As 𝐻 deteriorates, 𝑏_𝑛𝑒𝑡(𝐻) increases, raising expected revenue until it becomes profitable for some keeper to execute. This is exactly the mechanism that creates a break even root 𝐻^* and explains why keepers may wait inside a Dead Zone instead of racing immediately at 𝐻<1.

The mathematical bounds of the V4 piecewise liquidation bonus curve, illustrating zero incentive above H=1.0 and a strict cap below H_max.

2. The Profit Equation and the Break-even Root

A bot’s decision to execute the liquidateCall() function does not depend on the nominal bonus percentage, but is strictly governed by its net profit function π(H).

Given a target repay volume R(H), gas cost g_i, and DEX slippage cost s_i specific to bot i, the net profit equation is established as:

A bot will only execute a transaction when π_i(H)>0. The break-even point for bot i, denoted as (H^*)_i, is the exact root of the boundary equation:

3. The Dead Zone and the Waiting Game Logic

To observe how the implicit Dutch Auction dictates bot behavior, we must model the competition between multiple liquidator bots. The auction does not happen via an explicit orderbook. It is a race governed by execution costs.

Assume a macro position where the Target is 𝑇=1.05, total debt is 𝐷=5,000, and 𝜏=0.80.

We introduce two competing liquidator bots with distinct structural costs (𝑐_𝑖=𝑔_𝑖+𝑠_𝑖):

· Bot A (Optimized Searcher): Utilizes private RPCs and highly efficient routing. Total cost 𝑐_𝐴=50 USD

· Bot B (Standard Liquidator): Operates on public mempools with average slippage. Total cost 𝑐_𝐵=120 USD.

Phase 1: The Dead Zone (𝐻=0.994)

The market dips slightly. 𝐶=4,970 USD →𝐻=0.994

At this nominal risk level, the interpolated bonus is extremely low, e.g., 𝑏=0.02

Applying the Target HF sizing solver:

Gross profit is: 1,196.58×0.02=23.93 USD

Let us calculate the net profit 𝜋_𝑖(𝐻) for both bots:

System Result: Both 𝜋_(𝐴 )and 𝜋_𝐵 are strictly negative. The transaction falls into the Dead Zone. No bot is economically willing to execute. The market clock ticks, forcing the auction to wait.

Phase 2: The Auction Execution (𝐻=0.990)

The market drops further. 𝐶=4,950 →𝐻=0.990.

The risk deepens, scaling the interpolated bonus up to 𝑏=0.06.

Recalculating the execution size 𝑅:

Gross profit expands to: 1,485.15×0.06=89.11 USD.

Now, we observe the Dutch Auction resolution:

System Result: Bot A’s Expected Value turns positive. Bot A immediately signs and broadcasts the liquidateCall() transaction. The debt is dynamically repaid, and the position is restored to 𝐻=1.05.

Bot B never gets the chance to liquidate. Because Bot A has a superior cost structure, it mathematically wins the Dutch Auction by intervening at a higher 𝐻 (a shallower risk depth) where Bot B is still trapped in the Dead Zone. This mechanism forces liquidator bots to optimize infrastructure and routing efficiency, so the protocol can be secured earlier with less collateral seized than a late, deeper liquidation.

4. The Race Band and the MEV Shift

The core shift in V4 is not that MEV disappears. The shift is that the trigger for competition changes shape.

In Aave V3, the system is close to a single ignition point. Once 𝐻 crosses below 1.0, a fixed bonus can make Expected Value turn positive for many liquidators at the same boundary. That concentrates competition at one line, often creating Priority Gas Auctions.

In Aave V4, profitability is state dependent. Because 𝑅(𝐻) is compressed near the threshold and the bonus increases as 𝐻 deteriorates, Expected Value does not turn positive for everyone at once. Instead, it turns positive first for the most cost efficient liquidator bots.

Each bot has different optimization capabilities for gas (𝑔_𝑖) and slippage (𝑠_𝑖). Their absolute execution cost is:

In this implicit auction, a bot effectively bids through its cost structure by signaling: I will execute as soon as the public reward covers my private cost. A bot with better routing or private orderflow has a lower 𝑐_𝑖​, so it can execute earlier at a higher health factor where the bonus is still lower, but profit is already positive for that bot.

This is what creates the Race Band.

Mechanism definition: The Race Band is the range of 𝐻H where the most optimized tier of bots transitions from 𝜋(𝐻)<0 to 𝜋(𝐻)>0. In other words, it is the golden window where Expected Value becomes positive for top tier bots, but is still negative for the rest.

System Conclusion: Aave V4 does not eliminate MEV or Priority Gas Auctions (PGA). It merely shifts the ignition point from a static legal boundary (H=1,0) to a dynamic economic band (H≈(H^*)_i). Once market conditions push a position into the Race Band, the Expected Value (EV) turns positive, and actors will immediately burn Gas to secure execution rights. Consequently, the protocol minimizes the seized collateral volume while forcing Keepers to compete aggressively on structural cost efficiency, rather than engaging in the blind speed-racing inherent to V3.

CONCLUSION

Through rigorous source code dissection and on-chain empirical simulation, Aave V4 demonstrates a fundamental architectural shift: the transition from static legal constants to dynamic mathematical control functions.

By embedding the Target Health Factor into the execution sizing core and interpolating the liquidation bonus via a state-based auction mechanism, the V4 architecture definitively resolves two core DeFi paradoxes:

1. Capital Efficiency: It completely eliminates the capital destruction zone inherent to the V3 50% Close Factor’s over-liquidation logic. This maximizes user asset protection, preventing excessive capital seizure during micro-violations.

• System Solvency: It restructures the game theory of the Keeper network. V4 shifts the MEV ignition point into a converging break-even Race Band, ensuring that toxic debt is reliably cleared by bots possessing the most optimized cost structures, requiring only the absolute minimum systemic intervention.

In Aave V4, liquidation is no longer a blind, punitive asset seizure. It has evolved into a quantitative self-rebalancing mechanism, mathematically secured and precise down to the wei.

DISCLAIMER

In this article, we focus on Aave V4 liquidation mechanics, specifically Target Health Factor repay sizing and the state based liquidation bonus curve. Rather than evaluating overall protocol performance or making broad claims that one design is always superior, our goal is to break down the core design logic and execution mechanics that shape borrower outcomes and liquidator incentives.

We do not avoid technical details. The article includes smart contract flow, mathematical formulas, and simplified game theory. However, all models, examples, and calculations are intentionally simplified and explained step by step to make the mechanisms clear and intuitive for readers with different levels of DeFi experience. Where we use empirical charts derived from Aave V3 on chain data, the results should be read as evidence of patterns under specific assumptions, not as universal guarantees.

Ahead of the Aave V4 launch, we will return with additional deep dives on other parts of the architecture, including risk parameter design choices, incentive tradeoffs, and edge cases under market stress.

Leave a comment

INTRODUCTION

GammaSwap is an on-chain perpetual options protocol launched in 2023 with a clear promise: to hedge impermanent loss.

But its concept also left new users confused: “What is gamma?”, “What does “Gamma Swap” mean?”, “Why does swapping gamma help reduce impermanent loss?” and “How does this work through perpetual options?”.

In this article, we focus on answering these questions in the simplest and most fundamental way, explaining the core ideas behind gamma, impermanent loss, and how GammaSwap connects them.

KEY TAKEAWAYS

• Section 1: Gamma explains why exposure changes with price, and why swapping gamma means trading volatility instead of direction.

• Section 2: Impermanent loss is simply negative gamma embedded in AMM liquidity provision.

• Section 3: GammaSwap turns this gamma exposure into perpetual options for hedging and volatility trading.

Writer’s note:

This article connects option models in both DeFi and traditional finance, and includes some mathematical ideas (fully simplified). We recommend reading it patiently, step by step, starting from the fundamentals. Each section builds on the previous one, and the full picture only becomes clear when the pieces are put together.

***

“Gamma” is not specific to GammaSwap, it comes from a broader framework used to describe risks of a financial position.

Gamma is part of the most the most powerful analytical tools - the Option Greeks. They are a set of financial measures, designed to quantify risk and show how an option’s price responds to changes in market conditions, where delta and gamma describe how an option’s price responds to price movements in the underlying asset.

In practice, Greeks measures are most commonly used in options trading (Option Greeks).

An option is a contract that gives you the right to buy or sell an asset at a predetermined price.

The cost of buying an option (option premium) represents the market’s expectation of how much the underlying asset will move. You make a profit if the asset moves more than what is priced in, and you lose money if the asset does not move, or moves too little to justify that cost.

Suppose you buy a call option with a strike price of $34,000 (the right to buy asset at price of $34,000 on expiration day), and you pay an option premium of $200.

• This $200 is the maximum loss you can incur. If the price stays below the strike, the option expires worthless and you lose the premium.

• The breakeven price is the strike plus the premium. In this case, breakeven is $34,200.

Here is how profit and loss work:

• If the asset price is below $34,000, you do not exercise the option. Your PnL stays at –$200.

• If the price moves above $34,200, you start making money.

Traditional options have an expiration date, which means your opportunity to profit is limited to that timeframe.

A perpetual option removes this time limit, it gives the holder the right to exercise at any moment, without an expiration date. This is the concept used in GammaSwap, offering full flexibility to buy or sell the underlying asset whenever it becomes strategically favorable.

In the first two section, we use delta and gamma as general tools for measuring risk across different financial positions (not only option). By doing so, we can better understand their underlying meaning, and ultimately see what “gamma swap” through “perpetual option” really refers to.

SECTION 1: GAMMA AND “GAMMA SWAP”

Gamma is often mentioned as if it stand alone, but in reality, gamma cannot be understood without first understanding delta.

Among these Greeks, one describe what your exposure is, while the other describe how that exposure changes.

1/ Delta - the root of Gamma

Fixed delta (future contract)

Delta measures how sensitive an option’s value is to price changes in the underlying. In simple terms, it tells you how much the option’s price is expected to move when the underlying price moves by $1.

This sensitivity represents your price exposure - how strongly your position reacts to market movements.

In a futures position, your exposure does not change as the price moves, every $1 movement in the underlying always results in the same $1 gain or loss. In other words, delta stays constant across all price levels (fixed delta).

Suppose you go long 1 BTC perpetual futures at a price of $50,000.

If the price rises to $55,000, you make $5,000.

If the price falls to $45,000, you lose $5,000.

For every $1 move in price, your profit or loss changes by exactly $1. This one-to-one relationship is what defines a linear payoff.

In this case, we have fixed delta is 1.

Put on a chart, the futures value is just a straight line.

With K is strike price of a future contract

• A long futures position has a straight-line PnL: you gain as the price rises above K and lose as it falls below K.

• A short futures position is the exact opposite: your PnL decreases linearly as the price moves above K and increases as it moves below K.

This linear relationship remains exactly the same even when leverage is added.

For example, if the price moves by +1%, your PnL also moves by +1% (or −1% if you’re short).

That corresponds to a delta of +1 (or −1 if the position is short).

With 5× leverage, the PnL simply scales to +5% or −5%, but the line is still perfectly straight, nothing about the shape changes.

This means the position’s delta is +5 (or −5 if short), but it is still fixed.

In conclusion, fixed delta means that the value of a contract always changes at a constant rate relative to the price of the underlying asset.

Non-fixed delta, therefore, simply means that this relationship is not constant: the sensitivity itself changes as price moves.

Non-fixed delta (option contracts) and the root of gamma

Option contract and volatility trading

The most representative example of non-fixed delta is the option contract, a product that is best understood as a trade on volatility rather than on price direction.

The idea of trading volatility becomes clear only when we look at option structures such as a straddle, where directional exposure is neutralized and the payoff is driven primarily by price movement itself.

But if we only consider the payoff of an option at expiration, things like delta and gamma are not really meaningful.

The diagrams below show the payoff of basic option (call/put) positions at expiration, which is still linear from breakeven price (breakeven price is strike price plus premium):

• For a long call, the payoff is zero when the price stays below the strike price (explained below). Once the price moves above the strike, the payoff increases linearly as the price rises.

• For a long put, the payoff increases linearly as the price falls below the strike price, and becomes zero once the price moves above it.

• The short positions are simply the opposite.

However, these payoff diagrams only describe options at expiration.

In reality, what makes options interesting is that their value changes continuously before expiration. At any moment, an option’s price reflects the market’s expectation of how likely it is that the payoff will end up above zero.

Because of that expectation, with options, delta is not fixed. It changes depending on where the underlying price is relative to the strike, whether the option is out of the money (OTM), at the money (ATM), or in the money (ITM).

• Out of the money (OTM): The option has no intrinsic value. A call option is OTM when the price is below the strike, and a put option is OTM when the price is above the strike. In this region, delta is close to zero because price moves have little impact on the option’s value.

• At the money (ATM): The price is close to the strike. Small price changes significantly affect the probability of the option finishing in the money, so delta changes rapidly. This is where gamma is highest.

• In the money (ITM): The option already has intrinsic value. A call option is ITM when the price is above the strike, and a put option is ITM when the price is below the strike. Here, delta moves closer to one (or −1 for puts), and the option starts to behave more like the underlying asset.

As the underlying price moves, that probability changes, and so does the option’s value.

This is where we should mention “non-fixed delta”.

Non-fixed delta in option

For example, suppose there is a call option with a strike price of $300.

(The strike price is the predetermined price at which holder of an option has right to buy (call option) or sell (put option) the underlying asset at expiration)

Strike price is also the reference price that determines whether the option is OTM, ATM, or ITM.

The chart below shows how the delta of this call option changes as the price of the underlying asset moves relative to the strike price.

This chart shows how the delta of a call option changes as the underlying price moves.

• When the price is far below the strike ($300), delta is close to 0. For example, If delta is around 0.0, this means if the underlying price increases by $1, the option price only increases by about $0.05.

• As the price approaches the strike, delta increases rapidly. For example: delta may rise to 0.50. Now a $1 move in the underlying leads to roughly a $0.50 move in the option price.

• Once the option is deep ITM, delta flattens out and moves close to 1.00. For example: delta can reach 0.95 or higher. At this point, a $1 move in the underlying results in almost a $1 change in the option price, making the option behave very much like the underlying asset itself.

Explanation:

When the price is far below the strike, delta is close to zero because the option has little chance of finishing in the money. As the price approaches the strike, small price moves have a much bigger impact on that probability, so delta increases rapidly.

Once the option is deep in the money, the outcome becomes more certain. The option is very likely to finish in the money, so further price moves change that probability only slightly, and delta flattens out near one.

And this is the key to identifying convexity.

What makes this behavior possible is gamma.

And now, what is “Gamma”?

Gamma measures how quickly delta itself changes as the underlying price moves.

Put another way, gamma (Γ) is the first derivative of delta (Δ) with respect to the underlying asset price (P).

With a linear payoff (fixed delta) like futures.

• When the underlying price P increases by 1 unit, the position value always increases by a constant amount. As a result, delta is constant.

• Since delta is a constant, its derivative with respect to price is zero => Gamma = 0

For example: If you hold a futures contract, for every P rise in price, your PnL increases by a fixed amount. Delta is always 1. The rate of change of 1 is gamma, and it’s 0.

With a non-linear payoff (non-fixed delta) like options

• As the underlying price P changes, the option’s sensitivity to price (delta) also changes.

• Delta is a function of price, so its derivative with respect to price is not zero => Gamma ≠ 0

For example: With a Call Option, as the price S rises, the option becomes more likely to finish “in-the-money”, causing delta to increase from 0.50 to 0.60. This rate of change is the gamma.

To understand the way Gamma measures how fast delta changes as the underlying price moves, let’s make a simple example:

• When the option is far out of the money, delta is close to zero and barely changes. For example, if the price moves from $80 to $85, delta might only increase from 0 to 0.01 - gamma is very low (0.01 - nearly 0) in this region.

• Near the strike, delta change quickly. As the price moves from $95 to $100, delta can jump from 0.30 to 0.50. Another small move from $100 to $105 may push delta further to 0.80. This rapid change in delta is exactly what high gamma (nearly 0.3) looks like.

• Once the option is deep in the money, delta is already close to one and starts to stabilize. For example, a move from $120 to $125 might only change delta from 0.96 to 0.97, meaning gamma is low again (0,01 - nearly 0).

In short, gamma is highest around the at-the-money region, where small price moves cause large changes in delta.

——————————————————————————————

Once you fully understand all of this, a natural question comes up: why bother with delta, gamma, and all these moving parts, if profit and loss are ultimately determined at expiration?

In practice, traders often sell options before expiration, locking in gains when option prices rise, or cutting losses before the option’s value decays to zero, rather than waiting until the final day.

In that sense, realized PnL is driven by how the option’s price changes over time, not just by its final payoff.

This perspective becomes even more important when studying perpetual options, where there is no fixed expiration at all. Without a final settlement date, understanding how option value responds to price movements (through delta and gamma) is not optional, but essential.

——————————————————————————————

Once you understand what gamma is, the next question naturally arises: what exactly are you swapping when saying “swap gamma”, and why would you want to do so?

2/ Why swap the “gamma”?

As discussed above, gamma can be viewed as having two states: zero (comes from fixed delta position) and non-zero state (comes from non-fixed delta position).

Gamma swapping is, in essence, the exchange between these two states: swapping zero (or nearly zero) gamma for non-zero gamma, or vice versa.

From this perspective, taking one side of the swap results in a long gamma position, while taking the opposite side results in a short gamma position. Whether you are long or short gamma determines how your exposure evolves as the underlying price moves.

This distinction matters because gamma determines how your exposure changes as price moves.

• When you are long gamma, you are not betting on the price going up or down; instead, you benefit from delta becoming more favorable as the price moves in either direction.

• When you are short gamma, your delta becomes less favorable as the price moves, meaning the market moves faster than your exposure can adjust.

Demand for gamma swapping

(1) The first demand comes from participants who want to be genuinely long or short gamma.

• Some are willing to pay to gain convexity, benefiting from large and uncertain price movements.

• Others prefer to sell convexity, earning a premium in exchange for absorbing the risk that comes with adverse delta adjustments.

For these participants, swapping gamma is a direct way to express a view on volatility without taking a directional bet on price.

(2) The second demand comes from hedging needs.

Many positions are naturally exposed to gamma as a consequence of their structure rather than as a deliberate choice. In such cases, swapping gamma allows them to neutralize or reshape their gamma exposure without unwinding the underlying position.

In the context of this article, the focus will be primarily on hedging demand for gamma.

What makes this especially interesting is that: Although gamma exposure may sound abstract or unfamiliar, many market participants are not realizing it while already exposed to it for a very long time. It is often the core reason behind losses that many of you may already be experiencing.

One of the most common examples is impermanent loss, faced by liquidity providers.

The next section will make this connection explicit, showing how liquidity providers are effectively short gamma, and why impermanent loss is best understood through that lens.

SECTION 2: IMPERMANENT LOSS - A RESULT OF NEGATIVE GAMMA AND THE NEED FOR GAMMA HEDGING

While liquidity providers (LPs) may not think of gamma, the mechanics are familiar: as prices move, their exposure evolves in an unfavorable way.

When prices move, the AMM of liquidity pool automatically rebalances LPs’ position. This rebalancing leaves LPs’ position worth less than if they had just held the assets outside the pool. That gap is what we call impermanent loss.

A very detailed explanation of this concept has been published in “GAMMASWAP: FROM IMPERMANENT LOSS TO IMPERMANENT GAIN – HOW IS IT POSSIBLE?” article.

Referring back to the impermanent loss curve established in our previous analysis, we have:

• Blue dashed line (Holding a 50/50 portfolio): This represents the case where you do NOT provide liquidity. You simply hold 50% of one asset and 50% of the other in your wallet (for example, 50% ETH and 50% USDC).

• Black solid line (LP value in a X*Y=K AMM): This represents the case where you provide liquidity to a standard AMM pool (such as Uniswap V2), which follows the X*Y=K formula.

(1) Observing how Delta changes

Let’s look at the solid black curve that represents the value of an LP position.

As price goes up (moving to the right on the x-axis): The black curve becomes flatter. Its slope decreases compared to the steep slope near the starting point.

This can be explained that: when price increases, AMM automatically sells part of the asset that is going up in order to keep the pool balanced. Because you now hold less of the increasing price asset, the growth of your portfolio slows down.

In other words, your delta decreases as price moves away from the starting point.

These come directly from the formula used to compute the value of a LP position in a constant-product AMM:

• The value of the pool can be written as:

• Taking the first derivative of V with respect to price P, we have delta:

From this expression, the intuition is immediate: since P appears in the denominator, delta decreases as price increases.

(2) From decreasing Delta to negative Gamma

As we have mentioned before, gamma measures how delta changes as price moves. At its core, gamma is simply the derivative of delta, simply written as:

Taking the derivative of delta with respect to P, we have:

This result shows that anyone providing liquidity to a constant-product AMM (for example, Uniswap V2) is structurally exposed to negative gamma. The curvature of the LP payoff is an inherent consequence of the AMM design, not of market conditions or trading behavior.

Geometrically, negative gamma produces a downward-curving (concave) shape. This explains why the LP curve (black) always lies below the HODL curve (blue dashed line) as price moves further away from the initial level.

(3) Impermanent Loss

For comparison, consider the HODL position. Its value changes linearly with price, which means its delta is constant. Since delta does not change as price moves, its derivative with respect to price is zero. In other words, the HODL position has zero gamma.

Mathematically:

This contrast highlights the core difference between holding assets and providing liquidity. In other words, the red “loss” area on the chart is a direct consequence of negative gamma pulling delta down.

Therefore, no matter whether the price goes up or down, keeping your funds in a liquidity pool results in a lower total value than if you had just held the same assets in your own wallet. Moreover, this loss becomes increasingly pronounced as the asset price continues to rise.

————————————————————————————

This explains why swapping gamma for hedging purposes becomes necessary for liquidity providers. Once impermanent loss is understood as a direct consequence of negative gamma, it is clear that LPs are exposed to a form of risk they did not explicitly choose.

Viewed through this lens, GammaSwap went live in 2023 to address impermanent loss by allowing LPs to hedge their gamma exposure, using a financial instrument that is directly tied to gamma itself - options (perpetual options).

The next section will walk through perpetual options trading in GammaSwap and explain the concept of PnL when trading on it.

SECTION 3: GAMMASWAP – ONCHAIN PERPETUAL OPTION

Since options are the clearest type of derivative that having both zero and non-zero gamma, they are a natural way to swap gamma.

And to keep hedging continuous without having to roll positions, perpetual options (options with no expiration), are the most practical choice.

In this section, we do not go into a detailed breakdown of how GammaSwap perpetual trades work (this is covered in a previous article). Instead, the focus is on explaining how profits and losses arise when taking positions that are effectively bets on volatility, or gamma.

In essence, providing liquidity is economically similar to holding a short straddle option position, where you are short volatility and lose as price moves further away from its initial level.

Straddle Position – Hedging Impermanent Loss

A straddle position is a borrowed LP position where the withdrawn collateral is split in a 50:50 ratio between the two tokens. This makes it the direct opposite of a standard LP position.

GammaSwap gives you the opposite exposure of a liquidity provider. When you open a perpetual option position (such as a straddle, long, or short), the protocol burns the LP token and locks the underlying assets in a smart contract. Your position is then tracked using the liquidity invariant, the same x⋅y=k relationship used by the contant-product AMM like Uniswap V2.

For example: Suppose Alice borrows $1,000 worth of ETH and USDC when ETH is priced at $2,000. This burns $1,000 of LP tokens and leaves them holding 0.25 ETH and 500 USDC in a perpetual option straddle.

The position follows the constant-product rule x⋅y=k, so here

Now assume ETH rises to $3,000, the new pool price implies:

Now solve the system:

We have result is x≈0.204

Then, after rebalancing, we have:

Meanwhile, the borrowed assets are no longer subject to AMM rebalancing, because the LP position is burned and the underlying assets are locked in the perpetual option contract rather than remaining inside the pool. Their value simply follows the market price:

0.25 ETH × 3,000 + 500 USDC = $1,250

As a result, the straddle owner holds collateral worth $1,250 while owing a debt worth $1,225, earning a profit of $25 (ignoring fees).

Therefore, “impermanent loss” here becomes “impermanent gain”.

The mechanism is essentially a swap: loss arises when an LP position underperforms HOLD position, while gain emerges when that same LP position is borrowed out and converted back into a HOLD exposure.

The swap-design we have mentioned above is enabled by the way that: Payoff on GammaSwap is created by flipping the LP’s position.

As a result, the trader moves from a payoff that curves downward (loses from volatility) to one that curves upward (profits from volatility). Paying a fixed cost in exchange for unlimited upside from price movement is exactly the defining feature of swapping gamma.

Although GammaSwap positions are fundamentally built on a straddle-like payoff (benefiting from volatility rather than pure direction), long and short positions introduce an important modification: directional rebalancing.

Instead of keeping the borrowed LP position neutral, the protocol intentionally rebalances the collateral toward one side.

Long Position

In a long position, more weight is shifted toward the volatile asset (e.g. ETH), tilting the straddle upward and making it behave more like a call option.

For example, if a trader opens a $1,000 notional long position, 60% of the position will be in ETH ($600) and 40% of the position will be in USDC ($400).

At its core, this position remains a straddle, the PnL curve is still a convex, volatility-driven payoff, not a directional linear exposure like future.

What changes is not the structure, but the weighting.

By adjusting the token ratios, the trader intentionally introduces a directional bias. This reweighting tilts the payoff profile, making one side of the straddle more dominant and expressing a preferred direction, without abandoning the underlying long-volatility thesis.

In practice, even with directional bias, the position retains its convex PnL shape.

Therefore, a sufficiently large move in the opposite direction can still generate positive PnL, because a straddle profits from price displacement rather than direction.

Short Position

In a short position, more weight is shifted toward the stable asset (e.g. USDC), tilting the straddle downward and making it resemble a put option.

For example, if a trader opens a $1,000 notional short position, 40% of the position will be in ETH ($400) and 60% of the position will be in USDC ($600).

Crucially, this does not turn the position into a linear future. The payoff remains convex, and delta is still dynamic. The rebalancing simply biases the straddle so that as price moves, exposure naturally leans toward the winning side: ETH when price rises, or USDC when price falls, while still retaining the gamma-driven nature of the position.

Position PnL in practice due to AMM volatility

Across all three strategies, the PnL generated during downside price moves is consistently higher than that from equivalent upside moves.

Straddle PnL

Long PnL

Short PnL

In theory, a straddle produces a symmetric, convex payoff: gains increase as price moves away from the entry in either direction.

The theoretical PnL of a straddle forms a smile-shaped curve. As the price moves away from the strike in either direction, the payoff rises symmetrically.

In options theory, this shape is commonly referred to a smile shape.

This smile reflects the idea that position’s PnL increases as the market price moves deeper into OTM or ITM regions relative to the strike.

However, real markets rarely behave in such a balanced way. In practice, the smile is often distorted into a skew, where PnL rises more on decrease side of the price distribution than the other.

In GammaSwap, the notion of delta differs fundamentally from that of traditional options. Delta does not arise from a predefined option payoff, but is endogenously generated by the AMM liquidity formula itself.

As the underlying price moves, the pool’s inventory composition changes mechanically, resulting in a price-dependent delta exposure of the position:

With:

• S is the current price,

• K is the strike price

• L is the liquidity constant (a variable set by protocol)

Therefore, we have:

• When price falls (S→0 ): The term ​1/√S, grows very fast. As a result, delta changes sharply, making the left side of the PnL curve rise much more steeply.

• When price rises (S→∞): The term 1/√S slowly moves toward zero. Delta approaches a constant value 1/√K, so the right side of the PnL curve still increases, but in a smoother and more controlled way.

For the same percentage price move, downside moves cause delta to change much more than upside moves. This mathematical asymmetry is why GammaSwap PnL curves are naturally skewed and much steeper on the left side.

Because delta accelerates rapidly as price falls, a long volatility position effectively turns into a strong short position the moment the market starts to panic.

This means you profit significantly more from a 20% price drop (for example) than from a 20% price increase. This makes GammaSwap an especially powerful hedging tool for portfolios that are heavily exposed to spot tokens.

CONCLUSION

Impermanent loss is not simply an unavoidable cost of DeFi liquidity provision, but as a direct consequence of negative gamma exposure embedded in AMM design. By analyzing liquidity pools through the lens of options theory, we see that providing liquidity is economically equivalent to selling volatility, a short straddle position with fixed gamma and adverse convexity.

GammaSwap introduces a new on-chain primitive: perpetual options built from liquidity borrowing, allowing this gamma exposure to be explicitly transferred, hedged, or traded.

DISCLAIMER:

This article serves as a conceptual foundation for understanding GammaSwap.

Our focus here is not on the operational details of the protocol, but on the idea of “gamma swapping” from first principles, connecting options theory with impermanent loss in AMM mechanics. We aim to help readers clearly understand the structural logic behind impermanent loss, negative gamma exposure, and how perpetual options can be used to reshape that risk.

We do not avoid formulas or mathematical reasoning, but everything is simplified and presented step by step so that the core intuition becomes clear.

In the next article, we will move from theory to mechanism, breaking down step by step how GammaSwap structurally flips negative gamma exposure and turns “impermanent loss” into what can become “impermanent gain.”

Leave a comment

INTRODUCTION

Irrefutably, on-chain privacy demand is no longer limited to retailers or normal users. Institutions and enterprises, especially in finance, are increasingly seeking ways to reconcile transparency with confidentiality.

As stated by A16z Crypto:

“Privacy will be the most important moat in crypto in 2026. Many industries and users (like finance and healthcare) require companies to keep sensitive data private. It’s also a massive blocker for the institutions looking to tokenize real-world assets (RWAs) right now.”

The first era of privacy, exemplified by Tornado Cash, has disclosed the limits of full anonymity in cryptocurrency transfer under regulatory pressure. Today, newer privacy protocols like Railgun, Midnight, Fairblock built on Ethereum, alongside the evolution of early privacy chains like Monero with a more advanced model, have shifted toward selective disclosure as a balance between privacy and transparency in the traditional and decentralized finance, between Crypto Asset Service Provider (CASP) and regulators.

However, while selective disclosure may be necessary, it alone is unlikely to fully drive mass institutional adoption in 2026.

This article examines the crypto privacy landscape from an institutional perspective to explain why.

KEY TAKEAWAYS

Our research is structured around 3 key sections:

• Section 1: Major requirements for institutional participation in DeFi

• Section 2: Why selective disclosure is necessary, and how it improves upon earlier anonymity models such as Tornado Cash

• Section 3: The limitations of selective disclosure in attracting large-scale TradFi capital flows in 2026

  ---

SECTION 1: KEY FACTORS OF INSTITUTIONAL ADOPTION OF DEFI

What institutions need before entering DeFi?

First, we need to clarify the “institution” here.

Institutional adoption is often discussed as a single trend, but in practice, there’re two main kinds of institutions, including: crypto-native funds and traditional financial institutions.

Crypto-native funds, like hedge funds, market makers, or proprietary trading firms, can experiment security tools relatively quickly as their compliance requirements are lighter and their competitive advantage depends on confidentiality in their operations.

However, TradFi institutions operate under stricter legal constraints. Banks, pension funds, and asset managers must provide disclosure to regulators and auditors regarding all stakeholders, cash flows, and risk sources. For them, privacy is not just a technical feature but a legal responsibility.

To attract trillions of dollars from TradFi, DeFi in general and privacy protocols in particular need to meet some main requirements, as follows.

1/ Confidentiality And Auditability

Institutional transactions often reveal sensitive information: what bonds they are buying, how much stablecoin they are withdrawing, how much collateral they are depositing, etc. Some enterprises don’t want to publicize their investment portfolios, capital sources or their customers’ data.

If these information are all publicized and fully tracked on-chain, their commercial secrets can be exposed, and their competitors may take advantage of this.

Also, regulations related to anti-money laundering (AML) and combating the finance of terrorism (CFT) require regulatory bodies to be able to examine and retrieve transaction history when needed. For example, FATF Travel Rule requires retaining information on senders/recipients of transactions of over $1,000 or more and sharing it with relevant VASPs (exchanges, wallets).

Organizations need to report financial statements periodically, demonstrating that on-chain assets match off-chain records. Regulators, like SEC, FCA, MAS, require access to transactions (with orders) to investigate fraud, tax evasion, or violations of sanctions. It means: if there’s a lack of auditability, the organization might be perceived as concealing information.

2/ Legal Certainty

An institution in TradFi requires an absolute legal certainty to protect themselves from the risk of being sued and sanctioned. If there are no clear regulations on “what constitutes a legal private transaction,” regulators may automatically assume it as violating AML/CFT.

If there’re still many “gray areas” where the legislation of crypto or DeFi is ambiguous, institutions may need to hire their own lawyers for transaction analysis, which is both costly and hard to scale.

3/ Operational Reliability And Security Assurance

Every individual and organization tapping in the DeFi world also need to be careful, especially when it comes to investment. TradFi institutions, like JP Morgan, Goldman Sachs or other banks often execute large volume of transactions, up to billions, and they don’t want to risk their own portfolios if a privacy-based protocol faces chain downtime or a single network error. Traditional systems like SWIFT, Fedwire, and DTCC offer nearly 99.999% uptime, and organizations expect similar performance if they switch to on-chain.

Therefore, a privacy protocol that can maintain a robust infrastructure and secure no vulnerabilities is a requirement from an institution to pour their money in.

4/ Institutional Precedent

To broaden the institutional adoption, DeFi needs a successful pioneer from TradFi who leads the privacy trend. The larger the business, the more cautious they must be with every step they take. If a privacy protocol can truly prove their spectacular success with the participation of major organizations, a network effect will rise up and this would lead to involvement from many other parties.

These are four main factors that TradFi institutions need before tapping in DeFi, where privacy sits at the center. Without confidentiality and compliance, institutional participation becomes impossible.

This is where selective disclosure emerges as a practical fix to the limitations of earlier privacy models such as Tornado Cash. But whether it’s an ideological feature or not remains a question that needs to be answered.

SECTION 2: SELECTIVE DISCLOSURE – THE FIX FROM THE OLD RATIONALE (TORNADO CASH)

1/ The Failure Of Tornado Cash

In the first era of privacy, from 2009 to 2021, besides Monero and ZCash protocol, Tornado Cash, launched in 2019, is one of the most notable decentralized privacy protocols that enables transferring crypto without being traceable on public chains. Tornado Cash uses smart contracts and zero-knowledge proofs (ZKP) to hide the on-chain connection between the sender and receiver address of a transaction, which can be tracked easily on other public chains like Ethereum or Arbitrum.

One way to think about ZKPs is like proving you know the password to a vault without ever saying the password out loud. The verifier gains confidence, but no sensitive information is exposed.

However, despite good technology behind, this protocol was sanctioned by the OFAC (The Treasury Department’s Office of Foreign Assets Control in the US) for allegedly facilitating money laundering ($7B in virtual currency since launch) on Aug 8th, 2022. This sanction led to the founder’s imprisonment and the halt of project development.

What got Tornado Cash into trouble is its non-compliance by design. All funds deposited in Tornado lie in the same pool (i.e. Tornado Cash 0.1 ETH/1ETH/10 ETH/100 ETH) without any filtering or classification as clean or hacked money. Its mechanism doesn’t consider any proof to acknowledge whether money from sanctioned address or clean source.

Source: Etherscan

This feature drew significant attention from regulators, who became concerned about the potential for criminal activities such as money laundering, terrorism financing or other forms of illicit finance. As such, this caused Tornado Cash involved in regulatory non-compliance, including nearly $150M laundered by North Korea’s Lazarus Group hackers.

This model may serve legitimate privacy needs, but it also makes it easier for illicit funds to infiltrate, creating unacceptable risks for organizations.

In order to survive, privacy CASP must find a way to harmonize between privacy need and compliance. And selective disclosure, or programmable/selective privacy, is currently the most viable solution.

2/ The Need For Selective Disclosure And A Typical Case From Railgun

Institutions don’t seek to hide from the law but to avoid broadcasting sensitive business information to the world. And selective disclosure is a way to prove that some information is true or clean without revealing all underlying data. In practice, this is often achieved through cryptographic tools such as ZKPs or FHE (Fully Homomorphic Encryption) to create “selective proof”.

In crypto, it solves the privacy and compliance paradox we discussed earlier: If you want to be highly private (anonymous/secure), you’re easily suspected of illicit activity (money laundering) by regulators. Yet, if you want to be compliant (follow the law), you have to reveal everything, which means losing privacy.

Source: Coingecko

Since early 2024, despite some ups and downs, the total value of crypto treasuries held by companies and governments has grown at an unprecedented pace, with the all-time high of over $200B on October 2025. This trend highlights the need for programmable privacy core infrastructure, which allows many companies to both hide their investment strategies/commercial secrets from public scrutiny or competitor front-running, while still remaining regulatory-compliant.

In order to attract meaningful inflows from TradFi, selective privacy is a requirement. Recognizing this need, from the late of 2025 to early 2026, a new wave of protocols with compliant, programmable privacy designs emerged, including some notable names such as Aztec, Railgun, Fairblock, Midnight ...

Source: Ethereum Daily

Now, we will take Railgun as an example for clearer comparison in the design.

Railgun is a privacy protocol designed to strike a balance between privacy and compliance. The link between the sender and receiver addresses is still severed while the protocol still follows the law of OFAC.

Source: Etherscan

Unlike Tornado Cash, shielded transactions on Railgun still reveal the exact token amounts being transferred, as the protocol allows users to send arbitrary amounts rather than fixed denominations. It means, users who want their transactions to be harder to trace often need to standardize their transfer sizes to common base amounts, such as 0.1 ETH, 100 USDC, or 1,000 USDT.

Source: PPOI

In terms of compliance, by using Private Proof of Innocence (PPOI), before assets are “shielded” in the privacy pool, Railgun checks the AML/sanctions blacklist to verify that the token transferred doesn’t come from risky-tagged addresses before adding it to the privacy pool. This feature is to block suspicious funds from the start by using tag data from sources such as Chainalysis and sanctioned lists (i.e. OFAC).

Besides, Railgun users can also export their transaction history for use in the Koinly tax software, a part of compliance tools, helping users file their own taxes even when using Railgun‘s privacy wallet, without sacrificing basic privacy. This feature is specifically helpful for companies to declare and calculate their taxes.

Currently, the total volume of Railgun is about to reach $250M per month, reflecting the increasing adoption of this protocol and a rising demand of private payments in the crypto world.

SECTION 3: WHY IS SELECTIVE DISCLOSURE NOT ENOUGH TO ATTRACT TRILLIONS INFLOW FROM TRADFI IN 2026?

From all the figures and arguments above, we can’t deny that privacy is an important piece in the growth of DeFi. Despite the growth of privacy chains, selective disclosure, which is considered by some as privacy 2.0, securing the confidentiality and auditability, is not enough for DeFi to attract abundant inflows from TradFi in 2026 because of some following limitations.

1/ Delays In Regulatory Developments

One of the biggest stumbling blocks that hinders the growth of privacy techs in 2026 is the legislation progress. Public financial institutions are always careful with the legal framework, which should be clear and transparent enough for them (who manage huge assets) to avoid the risk of fines or asset freezing.

The year 2025 witnessed significant legal progress. Some crypto regulations has been in effect, including MiCA (Markets in Crypto-Assets), which provides a unified framework for cryptocurrencies within the European Union (EU). The US is currently promoting numerous laws to create a regulatory framework that is scalable and predictable, including GENIUS Act (for stablecoins), and CLARITY Act (for digital asset clarity). However, this clarity has not been fully achieved.

• GENIUS Act (Guiding and Establishing National Innovation for U.S. Stablecoins Act): The effective date of this act is the earlier of January 18th, 2027, or the date that is 120 days after the date on which the federal banking regulators issue implementing regulations.

• CLARITY Act (Digital Asset Market Clarity Act): This act is suffering from delays since the Senate Banking Committee postponed its markup/discussion of amendments to mid-January 2026 after Coinbase withdrew its support for the current version, meaning progress on market structure legislation has been delayed.

In reality, there are still many “gray areas”. The US and many other countries are completing the draft. With privacy, the issue is more sensitive because there is still no specific law on “privacy tokens”. A negative signal is that EU intends to ban privacy coins like Monero (XMR), ZCash (ZEC) on platforms permitted to operate within the block (i.e, exchanges, licensed CASPs). These regulations are expected to come into effect on July 1st, 2027.

Source: Cointelegraph

Moreover, reported by Chainlink, large organizations can’t disclose customer identities on public blockchains due to regulations like GDPR and GLBA (General Data Protection Regulation and Gramm-Leach-Bliley Act), which require personal data protection. Therefore, many institutions are still waiting for stricter regulations to ensure compliance with AML/CFT in private blockchain models. They are waiting for “gray” areas to turn “green”.

2/ Lack Of Infrastructure Maturity And Security

2025 witnessed a new wave of many privacy projects in DeFi, especially those run on Ethereum like Aztec, Railgun, or Miden. However, their infrastructure remains insufficiently mature to attract large-scale institutional adoption.

For example, on Railgun, it takes one hour for a shielded transaction to complete. While such latency may be acceptable for individual users, it poses a serious challenge for institutions -especially investment firms that rely on frequent and time-sensitive strategy execution.

Source: Railgun community

This delay is regarded as a trade-off in the design. Unlike Tornado Cash, which uses fixed-denomination pools (i.e. 0.1 ETH pool or 1 ETH pool) and enables withdrawals within minutes, Railgun allows arbitrary transfer amounts but requires longer settlement times to reduce traceability.

Source: PPOI

Moreover, in fact, current privacy tech still has its loophole as in November 2025, an Upbit hacker successfully laundered over 400 WETH (over $1.2 M) through Railgun and passed their ZK proof of innocence. This incident has raised concerns about privacy protocols’ robustness, which may further hinder participation from traditional organizations.

Furthermore, in terms of transaction fees, secure ZKP-based protocols like Railgun, or even Tornado Cash, generally impose significantly higher transaction fees compared to public blockchains like Ethereum, Polygon, Arbitrum, etc.

Source: Mirador

On average, transaction fee on several current privacy projects like Railgun (blue column) and Veil Cash (purple column) ranked most expensive about 0.25% - 0.3% on volume. In contrast, transferring on other public chains like Ethereum, Polygon or Arbitrum is much cheaper 0.001%-0.005%. With higher costs, traditional organizations will have even more reason to hesitate making private on-chain payments.

Finally, different from TradFi, one of the top concerns for both retailers and institutions when it comes to investing is the cybersecurity (risk of being hacked). Every project in DeFi (including privacy ones) inherently carries risks related to its logic/code, and tokenized RWA is not an exception.

Source: CertiK Report

Look at the chart above, the reality shows that hackers are increasingly interested in RWA assets as this market grows by exploiting loopholes in its logic.

The first half of 2025 already saw a significant loss of RWA assets (nearly $15M), including a $5.8M oracle manipulation attack from Loopscale and an $8.85M vulnerability exploitation from Zoth Protocol. These negative cases reflect an on-chain and operational failures within the core technology of RWA ecosystem.

All of these limitations above can further weaken the case for institutional adoption.

3/ Limited Proof Of Success And Network Effect

Large companies often wait for a clear successful case before deciding to join the game. To create an institutional FOMO, DeFi needs a proof that privacy tech actually works, with no bug or virtual yield.

A typical case in the previous cycle was when the Bitcoin/ETH ETF was approved by SEC in the early 2024, immediately a wave of capital was constantly poured in crypto.

Source: Bitcoin Strategy Platform

However, currently in the privacy tech space, there is no global adoption “leader”: no private blockchain has yet achieved widespread adoption for large volumes of institutional on-chain transactions.

In reality, Canton Network, a permissioned selective privacy chain model, backed by Goldman Sachs, BNP Paribas, and DTCC, now executes over $350B treasury repo flows per day from Broadridge’s Distributed Ledger Repo (DLR).

The report on the collaboration between Deutsche Bank and Nethermind notes numerous pilot deployments: from BIS’s Tourbillon Project, Bank of England – MIT, to Deutsche Bank’s PoC integrating the Privado platform, and “Proof-of-Reserves” systems for Binance, OKX, and others. These are signs that “on-chain privacy” has begun to be tested at the organizational level.

Yet, at present, there are still only a few small groups (Aztec, Dusk, Penumbra, Midnight...) providing privacy technology, and a dominant privacy platform has not yet emerged. Organizations are still waiting for “kicks” like a major leader announcing successful use of privacy liquidity. Realistic proof of great privacy is gradually emerging but is not enough to convince the entire industry to tap in.

CONCLUSION

2026 is a year of privacy tech growth but still not ready for mass institutional adoption. With privacy 2.0 infrastructure, protocols with selective disclosure requires more time for development and maturity, as well as harmonize perfectly between the privacy demand (of both retail users and institutions) and regulatory compliance.

In order to attract higher capital amount from traditional finance, as well as create a FOMO mentality among organizations, the duty lies not only in privacy protocols/chains only but also depends on other external factors.

Disclaimer

This research aims to analyze current trends in privacy technology and assess their implications for institutional adoption, based on observable market developments and existing infrastructure.

Privacy protocols and selective disclosure solutions are still at an early stage. Regulatory frameworks, technical designs, and adoption dynamics may change significantly over time. Any forward-looking views on institutional participation or capital inflows from traditional finance reflect market observations rather than guarantees.

Readers should treat this analysis as a reference for understanding the privacy landscape in 2026, not as investment or legal advice, and are encouraged to conduct their own independent research before forming conclusions.

Leave a comment

INTRODUCTION

“2× leverage” may sound like a simple amplification of exposure, but YieldBasis introduces something different: a continuously rebalanced leverage structure designed to neutralize the AMM’s curvature and transform the LP payoff into a linear one.

In Part I, we looked at impermanent loss from its mathematical roots. In this part II, we examine YieldBasis’ proposed solution: constant 2× leverage to make impermanent loss become zero.

KEY TAKEAWAYS

This article explores how YieldBasis constructs a constant 2× leveraged LP position, explains how it differs from simple static leverage, and then evaluates whether the strategy remains economically viable once costs are taken into account.

• Section 1: Explaining the “2x leverage” introduced by YieldBasis and the difference between constant leverage and simple static leverage

• Section 2: Testing Constant 2× Leverage - Accounting for Borrow and Rebalancing Costs

SECTION 1: THE BASIS BEHIND 2X LEVERAGE – THE CONSTANT LEVERAGE

In the previous article, we explained that impermanent loss in LP positions comes from the fact that price movements follow a √x pattern (where x is the token’s price change, such as WBTC).

The core to eliminate Impermanent Loss is to turn this √x behavior into a linear x movement, so that an LP position can move at the same pace as simply holding the token.

To make it happen, users’ value of a position must be changed without putting in additional capital. Of course, the most natural solution that comes to our mind is leverage.

And to address Impermanent Loss, therefore, Yield Basis’ concept is to apply leverage to effectively square the payoff function:

This is where “2× leverage” comes from.

However, thinking of it as just a 2× position is not enough, and can even miss the core idea. The section below walks through the concept in a clearer and more intuitive way.

a/ Simple static Leverage and Constant leverage

Not just 2X leverage – Must be “Constant” 2X leverage

As you think, simple 2X leverage can not eliminate Impermanent Loss

Simple static Leverage means using initial capital 𝑉_initial to borrow an additional amount D (debt), then create a LP position that is totally worth 𝑉_leverage, with:

The concept here is: You borrow once and never rebalance (debt is fixed).

• At the initial time (t=0), we have:

• When the asset value changes (t=1), debt is still fixed, we have:

Because debt does not change, every dollar change in LP value flows directly into your capital. This can be written as an equation of derivatives:

But since you are holding twice the LP amount relative to your initial capital, it becomes:

Integrating for total value, we have:

This means that the payoff curve is still a square-root (√P) function . It is merely shifted and scaled up in magnitude. It does not become a linear function (P).

Applying static leverage (for example, 2×) only makes the curve bigger, can not change its shape

This means: The higher the price, the slower the position grows.

This can be shown mathematically with a simple derivative:

As price x increases, the derivative decreases, meaning the position becomes less responsive to further price changes.

For example: With L = 2, assume you open a position with 2× leverage:

• Your initial capital: 100

• Debt: 100

• Total assets: 200

When the price goes up (you earn less than expected)

Suppose the asset price increases by 50%:

• Total assets rise from 200 to 300

• Debt remains fixed at 100

• Equity becomes 200

The actual leverage is now:

Now suppose the asset price drops by 30%:

• Total assets fall from 200 to 140

• Debt is still 100

• Equity shrinks to 40

The actual leverage jumps to:

Static 2× leverage does not maintain constant price exposure. It reduces leverage when things go well, and increases leverage when things go badly - the exact opposite of what is needed to hedge impermanent loss.

As a result, your exposure to price movements steadily decreases as the position moves in your favor. Even though the asset keeps going up, each additional price increase has a smaller impact on your equity than before.

In practice, this means leverage stops compounding returns on the upside, the 2x position increasingly behaves like an unlevered one, despite starting at 2×.

As a result, impermanent loss is not eliminated.

To turn a curve into a straight line, you cannot rely on static leverage alone. You must continuously rebalance to keep the position moving at a constant speed.

b/ YieldBasis core concept: Constant 2X leverage

Constant 2× leverage does not mean borrowing once to double your position size, instead, it means continuously maintaining a constant 2× sensitivity to price movements at every price level, not just at entry.

In other words: Static 2× leverage (which we have discussed before) just doubles your initial exposure, while constant 2× leverage keeps your price responsiveness fixed at 2×:

• The position is continuously rebalanced to keep leverage level.

• Borrowing amount is adjusted dynamically.

• The slope of the payoff remains constant at 2.

• The position moves at exactly twice the speed of the underlying price, at all price levels.

In other words, you do not commit to holding a fixed LP amount, instead, you commit to maintaining a fixed growth rate relationship.

For example: If the LP grows by 1%, your capital value must grow by 2%.

That means the position is continuously rebalanced so that your exposure always reacts at twice the speed of price, no matter where price is.

Therefore, instead of looking at absolute dollar changes, we look at relative growth:

This means the percentage growth rate of your capital is always twice the percentage growth rate of LP.

Move to logarithms:

Integrate both sides:

Using log rules:

Now substitute the LP formula of √x, now we have:

The curve now becomes a straight line; impermanent loss is eliminated:

In other words, the leveraged (green) line now perfectly overlaps with the hold (black) line. The position behaves exactly like simply holding the asset.

There is no more AMM curve effect, no divergence from holding. When the price goes up or down, the portfolio moves at the same rate as the asset itself — just like a normal spot position.

Comparison

Now, let’s put the constant 2x leverage line (green line) and static leverage line (red line) on the same graph to see the difference:

Interestingly, this shape looks very similar to the impermanent loss diagram shown in Part 1.

• The Green Line (Constant 2x) is a straight line. It represents continuous rebalancing. Because the position is constantly adjusted, the LP curvature is effectively neutralized. This is the ideal outcome: the structure behaves like a simple leveraged holding, with no visible impermanent loss effect.

• The Red Line (Static 2x) is curved, represents borrowing once and leaving the position unchanged. At the beginning, it closely follows the green line, but as price moves higher, it starts to bend downward and gradually falls behind.

If you compare this chart to the classic “HODL vs. LP” diagram from Part 1, the structure is almost identical.

The widening gap between the green and red lines shows that static leverage does not remove impermanent loss, it is just an amplification, not a transformation.

Through continuous rebalancing, constant 2x leverage actively reshapes the exposure, effectively transforms a liquidity position into a synthetic leveraged spot holding.

However, zero impermanent loss is no free, it refers to the costs: LEVERAGING COST and CONSTANT LEVERAGING COST

SECTION 2: THE COST OF ZERO IMPERMANENT LOSS

Of course, a 2x leveraged position will generate roughly twice the trading fees compared to providing liquidity with the original capital.

But consider a simple example:

Suppose you start with $100 and borrow another $100 to create a 2x LP position worth $200 (temporarily ignore overcollateralization requirements for simple understanding)

If the pool generates a 10% annual fee return, your gross fees would be $20 instead of $10, which looks attractive.

However, if the borrowing cost is 12% per year on the $100 you borrowed, that already costs you $12. On top of that, maintaining constant 2x leverage requires frequent rebalancing. Each rebalance may incur swap fees and slippage, or other additional costs...

While the fee income is amplified, the financing and maintenance costs are amplified as well. The real outcome depends on whether the extra yield is large enough to outweigh those ongoing expenses, or:

1/ Constant 2x leverage on traditional AMM: Negative PnL

To better understand how a constant 2x leverage strategy operates within YieldBasis, specifically how it interacts with a CryptoSwap pool, let’s first run a simple simulation with a traditional pool.

We model what would happen if the same constant 2x leverage logic were applied to a traditional constant-product AMM, here is Uniswap V2.

This exercise is purely illustrative. In practice, such a mechanism does not naturally exist in a standard AMM, the simulation is only meant to help us isolate and observe the mechanics of maintaining constant leverage in that environment.

The analysis breaks performance down into three simple but important drivers:

(1) Fee Income

The model looks at historical trading volume and calculates exactly how much fee income your position would have earned.

It adjusts this based on your real share of the liquidity pool over time, since your ownership percentage changes as the pool grows or shrinks.

So this isn’t theoretical yield. It’s your actual pro-rata fee income.

(2) Borrowing Cost

To maintain 2x leverage, you need to borrow capital.

The model pulls AAVE’s daily variable USDC borrowing rates and calculates the real financing cost over time. Because rates move, this captures the actual volatility of the lending market, not some fixed assumption.

This is the true cost of keeping leverage on.

(3) Rebalancing Cost

This is where most strategies quietly lose money.

To maintain 2x exposure, the position needs to rebalance whenever price moves. The model simulates those rebalancing trades day by day.

By comparing trade size against pool reserves (liquidity depth), it estimates the exact slippage and execution cost you would incur when adjusting back to target leverage.

In other words: how much value leaks out just from staying on strategy.

In most market conditions, the cost of borrowing stablecoins and constantly maintaining a 2x position is higher than the extra trading fees earned from using 2x leverage.

In short, the money paid for funding and rebalancing eats up the additional fee income.

The basic conclusion is that this strategy is not effective when run on a traditional AMM with high slippage. This is the primary inefficiency highlighted by the model, and the structural gap that Yield Basis uses CryptoSwap to solve.

2/ Constant 2x leverage on CryptoSwap AMM through YieldBasis

(1) Borrowing crvUSD

YieldBasis makes an LP position move 1:1 with BTC (eliminating impermanent loss) by pairing WBTC with borrowed crvUSD and maintaining a constant 2× compounding leverage.

• Borrow and pair: BTC is matched with an equal USD value of crvUSD, borrowed through a dedicated CDP. Both assets are then deposited into the Curve BTC/crvUSD pool.

• LP as collateral: The LP token received from Curve is used as collateral for the borrowed crvUSD.

• Target structure (50% debt / 50% equity): With 2× compounding leverage, debt always represents half of the total LP position value. As the LP value changes, the debt is adjusted to keep this ratio constant.

• Compounding leverage mechanism: A specialized “Re-leverage AMM,” together with a virtual pool, creates incentives for arbitrageurs to step in whenever BTC price moves. Their activity automatically pushes the position back to the 2× leverage target

(We will deeply diving into it in part 3 of this article)

As a result, the LP exposure is continuously reshaped so that it behaves like holding BTC spot with 2× leverage, but without the typical impermanent loss curve.

(2) Rebalancing to keep 2x leverage

Rebalancing basis: Borrow and Repay

In practice, the position is continuously adjusted by either repaying or borrowing additional crvUSD. When the BTC price moves, the value of the LP position changes, which causes the leverage ratio to drift away from its target. To correct this, the system:

• Borrows more crvUSD when leverage falls below the target, using the borrowed stablecoin to increase exposure.

• Repays crvUSD debt when leverage rises above the target, reducing risk and restoring balance.

This process ensures that the position stays close to its intended leverage, most importantly, the 2× target that neutralizes impermanent loss and preserves linear price exposure. The rebalancing mechanism is designed to be mechanical and capital-efficient, relying solely on crvUSD flows without requiring users to manually intervene.

(3) Rebalancing to concentrate liquidity (CryptoSwap pool)

Curve’s CryptoSwap uses concentrated liquidity, meaning capital is focused around a specific price range rather than spread evenly across all prices.

This design improves capital efficiency, but it introduces a challenge: When the market price moves away from the center of concentration, liquidity becomes less effective.

To restore efficiency, the protocol must shift the peak of the liquidity curve toward the new market price.

However, moving the curve is not free.

Re-centering liquidity requires trades against the pool, and those trades create a real economic loss. Therefore, the system needs a budget to pay for this adjustment.

Previously, this adjustment budget came from trading fees generated by the AMM, but it only works when trading volume is strong.

In YieldBasis, it is improved by using crvUSD borrow fee as part of liquidity adjustment budget: Rather than treating borrow interest as isolated revenue, the system redirects part of this continuous cash flow into the liquidity adjustment budget.

The protocol splits borrow fee paid by user:

• 50% goes to LPs/admin

• 50% is allocated as a subsidy for moving concentrated liquidity

This subsidy directly finances the cost of shifting the liquidity curve toward the new market price.

(4) Simulation with total cost

In this simulation, we found the answer for question “If I invest $1 million into the WBTC/crvUSD (YB) pair using 2x leverage, how much real profit would I have made per day over the past 60 days after deducting borrowing interest and all price-related losses?”

Important note: The Curve and Uniswap V2 simulations use different pairs and position sizes.

• Curve test: WBTC/crvUSD with $1,000,000 capital.

• Uniswap V2 test: WBTC/USDC with $10,000 capital.

This adjustment was necessary because Uniswap V2 does not have enough reliable historical data for WBTC/crvUSD, and its liquidity is much lower.

The smaller $10,000 position ensures the backtest reflects realistic market depth and avoids distortions from running large trades in a shallow pool.

This analysis runs a full 60-day historical backtest to measure the actual net profitability of a 2x leveraged liquidity strategy on WBTC/crvUSD.

From the implications discussed in part (c), we construct a simplified simulation model to evaluate the sustainability of constant 2× leverage.

The model combines on-chain data to break performance into three core components:

(1) Trading Fee Revenue (Gross Yield)

It aggregates daily trading volume and TVL to compute the Capital Turnover Ratio, which estimates how much gross yield is generated from trading fees.

(2) Rebalancing & Slippage Cost (Execution Drag)

To measure execution efficiency, the model compares actual trade prices with oracle reference prices (Chainlink/internal) on a minute-by-minute basis.

This captures the “hidden cost” of slippage — in other words, how much value is lost when the strategy rebalances during price divergence.

(3) Borrowing Cost (Cost of Capital)

It tracks the variable on-chain borrowing rates of the collateral assets to calculate the exact interest expense required to maintain 2x leverage.

For clarity, we do not treat the liquidity-movement subsidy as a separate expense. It is already internalized within the borrow cost.

In this Yield Basis test model, fee income is higher than costs in most market conditions.

For the majority of days, trading fees from the 2x position exceed the combined rebalancing and borrowing costs. This means the strategy is generally net positive under normal volatility and volume conditions.

However, there are specific stress periods where costs spike sharply, typically during sudden market moves. In those moments, rebalancing becomes expensive and can temporarily exceed fee income.

CONCLUSION

YieldBasis’ constant 2× leverage demonstrates an important insight: Impermanent loss is not unavoidable, it is structural, and therefore, it can be engineered around.

By continuously maintaining a fixed leverage ratio, the system effectively squares the AMM payoff curve, transforming √x into x. The LP position no longer drifts away from holding, it becomes straight.

In Part III, we will dive deeper into the mechanics of how this transformation is implemented at the protocol level.

Disclaimer:

In this article, we approach YieldBasis’ constant 2× leverage from first principles. Rather than taking “zero impermanent loss” at face value, we examine the mathematical logic behind it and test whether the mechanism truly reshapes the AMM curve as claimed.

Our goal is not to advocate or criticize, but to understand the structure: what works in theory, what depends on assumptions, and where real-world costs enter the picture.

In Part III, we will move beyond equations and explore how this transformation is implemented at the protocol level, and whether the design can remain sustainable in practice.

Leave a comment

INTRODUCTION

Hyperliquid has just introduced its newest improvement proposal, HIP-4, which enables prediction market/outcome trading, illustrating a strategic shift beyond perpetual (perp) futures into broader derivatives primitives.

Yet, this evolution also naturally leads to some questions about the potential of HIP-3, which supports permissionless builder-deployed perps. While HIP-3 promised a new “open marketplace” where builders could earn fees and bootstrap their own perps, the actual economics observed on-chain raise an important question:

Is deploying a HIP-3 perp still attractive for builders today?

This article is to evaluate the current and future economics of HIP-3 builders, quantify the revenue potential against operational and structural drawbacks, and assess whether listing perp markets on HL remains a worthwhile strategy in the evolving competitive landscape.

KEY TAKEAWAYS

This article will analyze the feasibility of HIP-3 in the deployer perspective through the following main sections:

• Section 1: Understanding about HIP-3 deployment

• Section 2: Analyze the economic feasibility of HIP-3 based on the current performance through the estimated revenue and RoSC indicator; along with the reasons behind the current dominance of TradeXYZ.

• Section 3: Identify some potential purposes as well as criteria that a builder often care about listing a perp DEX on HIP-3 under the non-financial perspective.

• Section 4: A forward-looking vision of the paths that subsequent builders might choose to deploy on HIP-3.

  ---

SECTION 1: HIP-3 DEPLOYMENT OVERVIEW

Before HIP-3, listing new perp markets is mainly controlled by Hyperliquid’s core team, which is similar to that of a traditional centralized exchange (CEX) like Binance, and delistings are driven by validator votes.

This approach seems to go against the core values ​​of a decentralized exchange like HL. Therefore, in October 2025, HIP-3 was introduced to allow any builders to deploy a perp DEX on HL’s infrastructure without permission. Since the launch of this proposal, in addition to validator-operated perps, HL now has another type of market called builder-deployed perps.

With the participation of TradeXYZ, the first builder-deployed market on HIP-3 as well as the first equity perp market on HL, HIP-3 is gradually seeing more entries from other builders, such as Felix, Ventuals, HyENA, KM, DreamCash and ABCDEx (coming soon).

After over 3 months of launch, this market is about to approach $40B in the cumulative trading volume, accounting for roughly 1% the all-time trading volume of HL. TradeXYZ market is still dominating at nearly 90% as reflected in the picture below.

1/ HIP-3 Deployer’s Responsibilities

Listing a perp market on HIP-3 is not simply a listing process, but a full market deployment lifecycle.

Unlike the centralized listing market model, HIP-3 requires builders to handle everything from designing specs to monitoring performance, making HIP-3 a system which is not only open but also places builders in a more specialized role.

The deployer of a perp market is responsible for: market definition and market operation.

1. Market definition (Step 1-4)

• Step 1:

In order to deploy a perp DEX on HL, each builder is required to stake at least 500k $HYPE, currently valued at around $15M, which will be locked for at least 30 days. Unlike ETH staking on Ethereum, this stake amount is not used for consensus assurance or block reward earning. Instead, it acts as a performance guarantee, where deployers may lose their stake if their market operations are not good (explained later).

• Step 2+3:

After meeting the staking requirement, the builder can deploy one perp DEX and design it by themselves, with its own underlying assets, collateral (i.e. USDH, USDT), order book system, and other independent contract configurations, including setting oracle prices, leverage limits, and settling the market (if needed).

• Step 4:

The first 3 assets (i.e. BTC/USDH, ETH/USDH) in any perp DEX can be deployed directly, while additional markets (i.e. TSLA/USDH, NVDIA/USDH) require going through a Dutch auction, whose each round lasts 31 hours, the starting price each time is double the previous final price.

This mechanism is like an economic filter to prevent random listing or spam markets, which can lead to fragmented liquidity. In short, only assets that are expected to generate significant trading activity are worth the initial auction cost.

Once the market goes live, the builder is ready to operate it (Step 5).

2. Market operation (Step 5)

After the market is deployed, builders do not interfere with order book system but responsible for market correctness, including oracle system, price verification and risk monitoring.

Like validator-operated perps, HIP-3 perp DEX is operated on-chain and transparently. Therefore, every deployer input can be analyzed by validators. If input is considered irregular, causing downtime/invalid state or making any malicious/incompetent behavior, validators can vote slash stake (up to 100%).

Because builders need to commit a significant stake value while may face the risk of slashing, deployment is not a low-cost investment decision but quite a risky one. This can be understood that HIP-3 is not just about “deployment,” but about transferring risk and market design decisions from the protocol to the builder.

Therefore, understanding the balance between potential benefits and the risk of HIP-3 deployment becomes extremely important.

2/ HIP-3 Deployer’s Benefits and Risks

Under HIP-3, builders can inherit the HyperCore infrastructure including its high performance margining and order books. For example, the API to trade HIP-3 perps is unified with other HyperCore actions.

Above all, listing a perp DEX on HIP-3 comes with the right of receiving half of fee revenue from market performance (fee share is fixed at 50%).

However, as mentioned above, any builders accused of malicious/incompetent behaviors can get slashed by conducting a stake-weighted validator vote (up to 100%), meaning losing all 500k $HYPE staked amount. Even if the deployer has begun withdrawing stakes, the stake is still slashable during the 7-day waiting period.

That’s why HIP-3 deployer is not only a market operator but also must maintain its risk management role.

Though this model offers builders certain flexibility and ownership to their perp products, it also creates a trade-off between capital commitment, operational responsibility, and potential financial returns. To assess whether this trade-off is justified, the following section will examine the financial feasibility of implementing the HIP-3 project from the builder’s perspective.

SECTION 2: FINANCIAL PERSPECTIVE - IS DEPLOYING A HIP-3 PERP STILL FINANCIALLY ATTRACTIVE FOR BUILDERS TODAY?

From a builder perspective, revenue is the first thing to consider when it comes to building a perp DEX, on HIP-3, in this case.

So, is deploying a HIP-3 perp still financially attractive for builders today?

In order to answer this question, we need to consider two main factors: estimated revenue and Return on Staked Capital (RoSC).

And it’s important to note that, though transaction volume and fees generated primarily come from the underlying products and partner protocols, HIP-3 serves as a critical infrastructure enabling these products to be deployed, scaled, and monetized within HL.

Therefore, the performance metrics analyzed below should be understood as results achieved through HIP-3, not independently of it.

1/ Estimated Revenue

NOTE: In order to work out the estimated revenue, we also collected some figure from Hyperzap, and assume that:

• Each deployer’s oracle performance is stable

• The price of $HYPE is at $30

• There’s no prolonged depegging of the quote asset

• Revenue sources mainly come from the trading fee and staking yield

In reality, HIP-3 revenue is also dependent on market growth and sensitive to liquidation dynamics.

Here’s the formula:

For a more accurate simulation, we will choose projects that’s been launched for at least one month. Currently there are 4 main partners on the HIP-3 market, including TradeXYZ, Felix, Ventuals and HyENA. Besides, some remaining builders like KM (markets by Kinetiq) and DreamCash have also joined but we will not consider the performance of these 2 markets as they’re only launched 3 weeks ago.

TradeXYZ, the earliest deployer launched 3 months ago, is sharing nearly 90% trading volume of the whole HIP-3 market, enabling trading with major equity products like XYZ100 (NASDAQ100), TSLA, NVDA, or GOOGL.

According to Hyperzap:

• The cumulative trading fee that TradeXYZ has collected is about $3.34M in 3 months, so on average, its monthly collected fee is over $900k.

• However, according to Hyperliquid HIP-3 docs, from the deployer perspective, the fee share is fixed at 50%. So, the adjusted trading fee that TradeXYZ collected in practice is over $450k.

• Together with the $HYPE staking yield (~ $30k), XYZ market’s monthly revenue is over $480k.

As such, we estimate that this market will gain over $5.8M in the yearly revenue. Based on this figure, TradeXYZ’s payback time will be less than 3 years.

Applying the same methodology, HyENA and Felix, launched nearly 2 months ago, are ranked second and third, and ranked last by Ventuals with the super small trading volume and fee as our metrics above. Specifically:

• HyENA:$1.6M/year payback period of over 9 years

• Felix: $950k/year payback period of over 15 years

• Ventuals: $450k/year payback period of over 30 years

At these levels, achieving capital recovery appears highly challenging for crypto projects operating under current market conditions.

2/ Return On Staked Capital (RoSC)

Because HIP-3 requires capital to be locked up rather than spent, the crucial question for the deployer is not capital recovery, but capital yield. And we decide to use the Return on Staked Capital (RoSC) indicator, which is quite the same as Return on Investment (ROI), a financial ratio used to measure the benefit an investor receives in relation to their investment cost. But in this case, Staked Capital is not lost from the start.

RoSC is used to answer the question: “With the same 500k HYPE locked (~$15M), is deploying HIP-3 more worthwhile than not deploying at all?” In simple terms, RoSC stated how effectively each USD of $HYPE staked is “working“.

Below is the formula of RoSC:

Where:

• Annual Fee Revenue: the estimated fee income per year as calculated above

• Staked $HYPE value: the value of staked $HYPE as calculated above

Our goal is to compare the fee returns of deploying a perp DEX on HIP-3 with other investment strategies with the same amount of money (~15M).

If not deploying a perp DEX on HL, this capital could be allocated to other yield strategies across DeFi.

• Currently, the staking APY of HYPE is only around 2.26%. The supply yield on some popular DeFi lending protocols like AAVE or Compound is often around 4%.

• By CAIA, ETH staking yields are around 2–4% depending on the protocol and chain, while SOL staking yields are around 6–7%.

• However, CoinDesk stated that lending or restaking strategies sometimes achieve returns of 10–25% in environments that encourage incentives.

So, we believe that:

• If RoSC < 5%, deploying on HIP-3 is financially ineffective (Red)

• If RoSC 5-10%, deploying on HIP-3 is strategic (Yellow)

• If RoSC > 10%, deploying on HIP-3 is highly profitable (Green)

Based on the above formula and these assumptions, below is our result of every market.

NOTE: These assumptions are not meant to be precise forecasts, but rather a reasonable baseline for comparative analysis.

Our result shows that RoSC of TradeXYZ is over 36% (Green), followed by HyENA at 8.6% (Yellow) and Felix at 4.05% (Red), whereas the remaining market (Ventuals) is only 0.78% (Red).

• Ventuals’ performance has not contributed significantly, accounting for only 0.78% (RoSC). It’s clearly shown that the main revenue of this market is from $HYPE staking yield.

• Felix’s fee revenue only accounts for about 4%, reflecting that the profitability of this DEX is still quite challenging.

• HyENA’s fee revenue sharing ratio is 8.6%, reflecting an effective deployment strategy.

• Notably, TradeXYZ’s total yearly income accounts for 38.7%, indicating that capital stakes are heavily leveraged through volume, and trading fees become the primary source of revenue (36.44%), rather than a secondary one.

Taken together, the estimated revenue and RoSC analysis indicate that:

HIP-3 remains financially attractive for a small set of dominant builders, but capital-inefficient for those unable to capture a significant market share.

3/ Why Is TradeXYZ Dominating The HIP-3 Market?

The first and foremost reason is that XYZ is the first mover on HIP-3 market, which is a superior advantage for this DEX itself. This is also the pioneering DEX allowing equity perps trading on HL, which has never appeared on this platform before. Many traders joining in the crypto world also have engaged in macro and equity trading, but lacked access to equity derivatives in a non-custodial, blockchain-based environment. TradeXYZ has tapped into this unmet need and quickly gain the market share advantage.

The first product on HIP-3 is XYZ100/USDC, which was launched on 13th October, 2025, recorded about $13M in trading volume on that day. Day by day, this figure together with other equity products, continues to surge, helping the XYZ cover the majority of HIP-3 market share (about 90% as the metric below).

Another reason for the current success of this deployer is that XYZ markets currently include equities and commodities quoted in USDC – the most popular stablecoin used as collateral and the quote asset on HL.

On the other hand, Felix and Ventuals also enable equity/commodity and pre-IPO valuations trading pairs but only quoted in USDH. This is a new fiat-backed native stablecoin of HL just launched 3 months ago; therefore, user confidence in this situation is not robust enough to facilitate the trading volume. That’s one of the most plausible reasons why these 2 markets’ performance appear to be inferior to the others’.

In short, TradeXYZ, with the first-mover advantage and a smart choice of the quote asset, leads the HIP-3 market with flying colors. Looking at the current situation, much of HIP-3 market’s success depends on TradeXYZ’s performance, reflecting an imbalance market competition.

SECTION 3: NON-FINANCIAL PERSPECTIVE - WHO IS HIP-3 FOR?

However, if we only look at the mentioned-above financial factors to determine whether a builder should deploy a perp DEX on HL or not, that’s a relatively narrow perspective. Only considering the estimated revenue or RoSC, we can’t firmly decide that Felix, Ventuals or other smaller builders shouldn’t list their perp markets on HL.

From a deployer or a project perspective, in addition to economic view, other factors are also important, even more important than the short or long-term profit.

HIP-3 is a suitable marketplace for the following target builders.

1/ Those With Product-Market Fit But Lacks Of Distribution

Distribution fit refers to a builder’s ability to reach real users with actual trading demand and capital.

For many new builders - already equipped with product-market fit and solid infrastructure (oracles, risk engines, funding logic) - this has become a primary concern.

At launch, Hyperliquid had not yet listed real-world asset (RWA) perps such as equities or commodities. On HyperCore, crypto-native markets like BTC, ETH, SOL, and memecoins already enjoy deep liquidity and mature validator support. As a result, HIP-3 builders have limited competitive advantage when launching additional crypto-native perpetuals.

Therefore, a new path is needed to create a differentiation on HIP-3 market. RWAs, with the base of a stably high-quality volume from traditional finance (over $1 trillion) and little volatility (less manipulation risk), appear to be a better option – a product-market fit, especially for TradFi-curious and macro traders.

Moreover, HL has many benefits that attract early deployers like XYZ, such as user quantity (nearly 900k), high liquidity and volume (top 1 perp DEX by Open Interest - $7B), a CEX-like UI built on HL Layer-1 blockchain with a smooth trading experience and low latency (currently about 10k TPS).

Considering these factors, HIP-3 is a distribution channel for financial products that fit the market.

2/ Those Who Want To Monetize Their Core Asset

HIP-3 is also a suitable sales channel for a project to list their perp market, in order to amplify the value of the project’s core product.

A typical example in this case is Ethena. On 9th December, 2025, HyENA DEX (the collaboration among Based, Ethena and Hyperliquid) went live on HIP-3 perp market. Most trading products on this market are crypto-native perps, but quoted in USDe – a stablecoin of Ethena launched in the early 2024, such as BTC/USDE, ETH/USDE.

The main purpose of Ethena listing a perp market on HIP-3 is not just to earn trading fee, but creating indirect demand for their core asset, USDe.

Within over 50 days of launch, the HyENA market recorded over $1B in trading volume and generated more than $380k in trading fees. With volumes continuing to trend upward, HyENA is beginning to close the gap with established HIP-3 deployments like TradeXYZ.

The more trading volume this market gets on HIP-3, the higher USDe exposure and usage will be. The metric below reflects the increase in the USDe supply since the launch of HyENA. Currently, over $75M USDe has been used on Hyperliquid and this number is still growing, showing the potential of a USDe-based margin trading market.

3/ Those Who Want To Seize The First-Mover Advantage And Trust In Hyperliquid

HIP-3 is a fit to the builders with a strategic vision, especially those who can grasp the opportunity of being first on the battlefield.

Take Felix and Ventuals as an example. Even though the trading performance of these two markets remains modest, only about 0.3-2% HIP-3 market share, what they are betting is the future of their products. Both protocols focus exclusively on RWA-linked products (stock indices like TSLA, MAG7, OPENAI...) but quoted in USDH.

While USDH, Hyperliquid’s native fiat-backed stablecoin, is still early-stage and has yet to be fully stress-tested, its future adoption could materially reshape this segment of the ecosystem. Like Ethena and its stablecoin, which was launched two years ago, now has become one of the largest stablecoin in the crypto world.

USDH is just 3 months old, and still has a long journey ahead, together with HL ecosystem. If USDH usage increases further in the future and RWA trends improve, Felix and Ventuals will benefit even more as early movers, and vice versa.

Notably, competition in USDH-denominated markets is already intensifying, specifically with the entry of KM (markets by Kinetiq). KM has processed transactions totaling $680M in just over 3 weeks, underscoring the latent demand in this market.

Anyway, this will be a long-term journey that only large builders can survive.

SECTION 4: LOOKING AHEAD - WHAT CAN LATE ENTRANTS DO ON HIP-3?

TradeXYZ is a pioneer in RWAs perp trading, Felix and Ventuals are pioneers in RWAs trading products quoted in HL’s native stablecoin, and Ethena (HyENA) is pioneer in monetizing their core asset – USDe.

So, a strategic question is: What can late entrants do on HIP-3?

HIP-3 is not completely closed. It’s only closed to small builders with short-term vision, but not the opposite side.

Large builders can also choose the same path as Ethena (HyENA) to route the volume and exposure to their core assets. Currently, most of trading pairs on HL are quoted in USDC, USDH and USDe; and builders can choose to list HIP-3 perp products quoted in other top stablecoins. For example, Sky can deploy a DEX to boost their DAI/USDS stablecoin exposure.

Besides, product-market fit is always an essential factor to consider before deciding whether to list a perp market on HIP-3. HIP-3 needs novel markets that meet the trader demand.

One of the promising primitives that remains unexplored on HIP-3 is prediction market–based perps. At present, RWAs dominate HIP-3 deployments but there’s still no prediction market perps listed to track probabilistic or event-driven outcomes. This market is highly potential because of its high trading volume.

As can be seen in the figure below, every week Polymarket and Kalshi – the leading prediction markets today - attract over $2B in notional volume. This figure shows that the demand for speculation and hedging based on events is very high.

Also, as predicted by 21shares, by 2026, prediction markets, like Polymarket and Kalshi, are expected to surpass $100B in annual trading volume, onboarding millions of users as real-world events in geopolitics, sports and beyond are monitored, traded and registered onchain.

If we look at the weekly volume of this market, the total volume has just exceeded $6B, reflecting a high trading demand.

With HIP-3’s permissionless model, deployers can try transforming “event-based” prediction markets into 24/7 leveraged perp markets. This could be a suitable direction for late entrants as they don’t need to compete directly with equity markets, but can instead focus on product innovation.

In fact, with the recent introduction of HIP-4, which aims to build upon HIP-3 by enabling outcome and prediction contract primitives, if builders are afraid of being a HIP-3’s late entrant, becoming an early mover on HIP-4 with prediction market is also worth considering.

Besides, large segments of traditional derivatives markets remain largely untapped in crypto-native perp venues. Looking at the daily trading volume of the global forex market (nearly $10B), but perp products linked to FX are still very rare on on-chain platforms. So, listing a perp DEX like this on HIP-3 can also be a potential path.

Hyperliquid in general, and HIP-3 in particular, requires niche and differentiated markets to drive expansion. And from the project’s perspective, this product aligns well with current trading demand.

CONCLUSION

To sum up, HIP-3 is not a money printer for every builder but only for those with great resilience and capital to survive a long-term journey.

To remain competitive under current conditions, builders must not only meet Hyperliquid’s listing requirements but also deliver products that align with prevailing market narratives.

For some participants like Ethena, HIP-3 functions less as a profit center and more as strategic infrastructure.

And for those who place long-term trust in the growth of Hyperliquid, HIP-3 is still an open gateway for them to expand in the future.

DISCLAIMER

This article is an independent research-based analysis of the HIP-3 framework and its incentive structure for builders on Hyperliquid. It aims to evaluate HIP-3 from an economic and strategic perspective, rather than to promote any specific deployment or project. From there, we hope readers can form their own perspective on the associated opportunities and risks.

All views presented are based on publicly available information and reflect the state of the market at the time of writing. Outcomes for individual builders may vary significantly depending on execution, capital strength, risk management, and broader market cycles. This content is for research and educational purposes only and should not be interpreted as financial or investment advice.

Leave a comment

INTRODUCTION

HIP-3 is transforming Hyperliquid from a single perps DEX into a permissionless derivatives marketplace. Among the first HIP-3 deployments, HyENA stands out as a perps DEX built on Hyperliquid with USDe margin, backed by the strategic collaboration between Ethena, Hyperliquid and Based.

Despite going live for just over one month, HyENA has already attracted notable attention due to its rapid growth and unique margin design.

This article provides a detailed breakdown of HyENA’s economic design and trading mechanics, and evaluates whether it offers a compelling value proposition for traders.

KEY TAKEAWAYS

In this article, we explore the inner workings of HyENA perps DEX and examine its potential under the trader’s perspective through the following main sections:

• Section 1: Explore the operating model of HyENA

• Section 2: Explore the benefits that users and protocols parties will gain

• Section 3: Examine the potential of trading on HyENA from the user’s perspective

SECTION 1: HYENA OPERATING MODEL

HyENA is a perp DEX deployed on HIP-3 (Builder-Depl oyed Perpetuals Improvement Proposal of Hyperliquid) standard, with the combination of 3 main layers powered by 3 following protocols: Hyperliquid, Ethena and Based.

1) Hyperliquid HIP-3 – Infrastructure Layer

Before HIP-3, only the core team can list perpetual markets on Hyperliquid. In order to solve this limitation, HIP-3 was introduced as a framework born to allow any builders to deploy their perps DEX permissionlessly, as long as they meet certain listing requirements.

One of the most notable requirement is related to staking amount, which is currently 500k $HYPE, worth roughly $15M. It means HyENA was launched only after Ethena successfully staked the required 500k $HYPE, serving as a safeguard against malicious or low-quality deployments.

Hyperliquid permits HyENA to be deployed as a separate perp DEX while inheriting the Hyperliquid’s core infrastructure, including orderbook, matching engine, and low latency of this protocol. Moreover, HIP-3 allows every deployer to customize their own market, collateral, fees, and incentives.

Thanks to HIP-3, HyENA users enjoy the same trading experience as on the Hyperliquid DEX. Hyperliquid has HLP vault acting as the protocol’s liquidity layer through multiple market making strategies, performing liquidations, and accruing platform fees. Likewise, HyENA also has HLPe vault (managed by Upshift), which fulfills an equivalent role, but the only difference is that HLPe vault is used for market making strategy, it is not directly involved in liquidations or Auto-Deleveraging (ADL).

From a structural perspective, HyENA can be viewed as a “mini Hyperliquid DEX”, inheriting the same execution and liquidity architecture while operating as an independent HIP-3 deployment.

2) Ethena – Asset Layer

Most trading products on HyENA are native tokens quoted in USDe, like BTC/USDe, ETH/USDe. So traders are required to deposit USDe as maintenance margin when opening a position on this DEX.

In return, traders can earn a certain extra USDe reward (APR) on their full spot and perps margin balance powered by Ethena. Currently, the base yield rate applied to traders with available USDe is 4.5%/year. It means that in addition to the trading PnL, traders can earn an extra 4.5% APR, which will be applicable to all idle balances and balances used to margin positions.

With this incentive, margin yield can partially offset trading costs such as fees and funding, depending on position direction and funding conditions.

How can USDe keep its peg and earn yield for stakers?

USDe (Ethena USD) is a yield-bearing synthetic dollar launched in February 2024, not backed 1:1 by cash or short-term US Treasury bills like USDC but via holding liquid stables like USDC and USDT and the delta-neutral hedging mechanism, where Ethena:

• Long spot ETH (or LSD ETH)

• Short perpetual ETH with equivalent volume

For every dollar of USDe, Ethena will hold an equivalent amount of crypto as collateral while going short on CEXs like Binance. Or:

1 USDe = $1 of ETH + Short 1 ETH / USD Perpetual

By this way, Ethena neutralizes ETH price volatility, creates a value of approximately 1 USD for its stablecoin.

Moreover, because perpetual markets typically have a positive funding rate, Ethena collects funding fees plus the basis spread from that short position. This is the main source of revenue to pay for USDe stakers, who will receive sUSDe (Staked USDe) after staking as a receipt to gain yield afterwards.

3) Based – Execution Layer

The final party in the development of HyENA DEX is the Based team, which is also the leading Hyperliquid application in revenue. Based acts as an execution layer, which provides front-end trading and enhances UI/UX for HyENA traders.

SECTION 2: TRADERS AND PROTOCOLS PERSPECTIVE - WHAT WILL THEY BENEFIT?

1) What Will Traders Benefit?

The biggest difference between HyENA’s users and Hyperliquid’s or non-HyENA’s native users is the lazy yield generated from USDe margin.

At its core, USDe doesn’t generate any yield from the stablecoin itself. On Ethena, sustainable yield is earned only when USDe is staked into sUSDe, with returns primarily sourced from protocol revenue such as funding and basis trades. So theoretically, if you only hold USDe without locking or staking it, you won’t earn any yield.

Besides, on HyENA, USDe holders are not required to stake USDe to earn yield. Instead, USDe used as margin becomes effectively immobilized within the trading system, creating a form of sticky supply. And this reward is also derived from Ethena’s revenue. Ethena designs yield as an incentive layer for traders with USDe margin, aligned with Ethena’s objective of securing circulating supply while encouraging USDe adoption in derivatives trading.

After one month of launch, HyENA has allocated over $400k to the eligible USDe holders. Currently with nearly 9k users, on average, each user is qualified for receiving $45. This is just a hypothetical averaging calculation since in practice, USDe rewards are highly proportional to traders’ volume.

Some traders may earn a relatively huge amount of incentives, like in the picture above, the highest USDe reward ($16.3k) currently belongs to a whale’s wallet address ***9482 with the total spot and perp balance of over $3M (by Flowscan) while other smaller fishes are earning less. As rewards are distributed on a daily cadence to user’s HyperCore balances, based on rolling average user balances, this amount can grow more.

About the rolling average user balances, it can be understood by the way that the system takes multiple balance points over time and calculates the average to determine the “bonus balance” for that day, instead of looking at the balance at just one point in time.

Example:

If you hold USDe inconsistently throughout the day, assume your balance record looks like this:

• 00:00 - 0 USDe

• 06:00 - 1,000 USDe

• 12:00 - 1,000 USDe

• 18:00 - 0 USDe

Then, the average balance = (0 + 1000 + 1000 + 0) / 4 = 500 USDe

At the current 4.5% base APR, your USDe reward would be 22.5$/year

Even if you have 1,000 USDe at one point, the bonus will be “almost the same” as if you held an average of ~500.

However, if your balance stays at 1,000 USDe every day, the average would be 1,000 USDe

And with 4.5% APR, your USDe reward would be 45$/year

NOTE: The HyENA’s documentation only mentions “rolling average user balances” but doesn’t specify the timeframe or formula, so this is just a hypothetical example for readers to understand the basic mechanism behind reward calculation.

So, what traders benefit from HyENA is the bonus that can be passively earned from the USDe margin, and how much it is depends on their perps and spot balance.

2) What Will Ethena Benefit?

Aside from the collected trading fee, thanks to the launch of HyENA and its positive performance, the USDe adoption and supply has increased.

Currently, the total trading volume gained on HyENA HIP-3 market is over $900M, resulting in the USDe supply increase on Hyperliquid, nearly $80M in over 1 month and counting.

If the quantity of USDe margin is getting higher, it creates a sticky demand for USDe. As mentioned above, when traders use USDe as collateral, this amount will be locked in the trading system. ~$80M USDe on Hyperliquid means that it’s not available for sale on the CEX, not sitting idle in the wallet, and not easily withdrawn if HyENA traders still keeps their positions.

As a result, from a strategic view, this not only improves Ethena’s hedging efficiency at scale, but also accelerates USDe’s transition from a passive stablecoin into a productive on-chain collateral asset. And as long as USDe continues to be used as collateral on Hyperliquid or HyENA, the risk of mass redemption will be reduced.

Moreover, Ethena is gaining distribution without needing to build a DEX from scratch. HIP-3 in general, and HyENA specifically, inherits the high performance margining and order book system of Hyperliquid. As such, Ethena can gain access to deep trader liquidity without bearing the cost and complexity of building its own derivatives infrastructure.

3) What Will Hyperliquid Benefit?

Not to mention the collected trading fee, one of the biggest issues that Hyperliquid can solve is the capital inefficiency in the idle USDC locked as collateral. Most of financial products on Hyperliquid are trading pairs quoted in USDC, whose total supply on this protocol is around $4B. However, as only used as the maintenance margin without bearing any lazy yield, this huge amount is considered inefficient in usage.

With USDe on HyENA as yield-bearing margin, with the current base APR around 4.5%, not to mention the boosted reward at the first initial month (over 7%), this can be an effective solution for Hyperliquid to handle the capital inefficiency.

As seen in the hypothetical metric above, if USDC margin also goes with 11% APR rewards as the model of USDe, the all-time reward would be over $500M. And even only with the base APR (4.5%), the reward would be around $196M – the amount that should have gone to the USDC holders’ wallet but was missed out on.

In short, between Ethena (with higher USDe exposure) and Hyperliquid (with enhanced capital inefficiency), HyENA is a product of a win-win relationship.

SECTION 3: IS HYENA A PROMISING INVESTMENT CHANNEL FOR TRADERS? WHICH POTENTIAL RISKS SHOULD A TRADER BE AWARE OF?

To answer the question: Is HyENA a promising investment channel for traders?

We are to examine the profitability of trading on HyENA HIP-3 from a trader’s perspective.

1) Profitability

Obviously, profit is what traders care most when it comes to trading.

Not to mention the trading PnL (as this depends on the profit-making potential of each individual), is HyENA financially attractive enough for traders to join the game?

At present, according to HyENA team, USDe reward is configured at the base 4.5% APR, the same as sUSDe staking rate on Ethena. Specifically:

• Holding USDe as spot on HyENA earns a ~4% APR.

• Opening a short position on USDe-margined perpetual pairs on Based/HyENA earns a ~4% APR on margin.

• Opening a qualifying long position on USDe-margined perpetual pairs on Based/HyENA earns a ~12% APR on margin (this is boosted reward during HyENA’s initial month only).

USDe reward is a great bonus for USDe holders on HyENA, but more attractive for large holders. All HyENA’s users automatically earn native USDe rewards on their full spot and perps margin balance powered by Ethena. The more your spot/perp balance, the higher reward you will receive. This also means that rewards are inherently size-biased, favoring large balance traders and institutions over smaller retail users.

As the metrics above, many traders with small spot/perps balance only collect a small amount of USDe rewards. In the unfavorable cases where the trading PnL is negative, USDe can’t compensate for the traders’ losses.

USDe reward is basically an incentive for users to trade on HyENA which increases USDe’s exposure for Ethena protocol. If large holders or traders in general know how to take advantage of this mechanism, they can benefit more than others.

2) Other Potential Concerns

Beyond the reward, as a trader, you should also be aware of other potential issues that might arise during the trading process. The two main concerns to take into account is the depegging risk of USDe and the oracle risk of HyENA.

Depegging Risk

One of the biggest concerns for traders when deciding whether to trade on HyENA is the risk of losing the USDe’s peg.

USDE previously had a history of losing it peg. On 10/10, USDe price on Binance dropped to $0.65, causing many positions related to this stablecoin to be liquidated.

However, as stated by Guy Young, founder of Ethena Labs, this was not due to problems with the collateral or Ethena itself, but a failure of the internal Binance oracle, which fetches data from its own orderbook rather than an external price feed.

Binance’s oracle design flaw was the major vulnerability and also the root cause in this case. On the same day, this CEX announced a promise to switch to external price sources by October 14.

About the risk of depegging caused by the Ethena itself, USDe maintains its peg 1:1 via delta-hedging mechanism as mentioned above (i.e. long spot and short perps). Other than that, this mechanism also helps Ethena to earn positive funding rates from short perps, which is also Ethena’s revenue.

Certainly, the future is unpredictable, but since its launch in early 2024, USDe has not experienced any depegging incidents due to insolvency risks stemming from the protocol itself.

Oracle Risk

About perpetual contracts on HyENA, USDe is used only as collateral of maintenance margin, while the oracle price is still quoted in standard USD. By not converting USDe to USD in the oracle layer, HyENA cleanly separates pricing risk from collateral risk.

Price fluctuations in USDe on HyENA only affect the collateral asset class, impacting traders’ margin capacity rather than shaping market prices. Since USDe is not referenced in the oracle system and plays no role in calculating the reference price, oracle-induced price distortions cannot trigger or amplify USDe price breakouts. Therefore, the risk that USDe depegs due to the oracle price won’t exist.

Also, to power the trading infrastructure, HyENA perps are primarily fetched from Edge – oracle protocol of Chaos Labs to compute funding rates and is also a component of the mark price calculation. Unlike leading oracle providers like Chainlink or Pyth, Edge is a new decentralized oracle protocol just launched for over 2 months so it needs more time to test its security.

So, as a risk mitigation measure, HyENA is also running with a backup oracle provider as failsafe and will reuse oracle prices from Hyperliquid, which is a weighted median of CEX prices. For instance, the oracle price of BTC-USD perp on HyENA will match that of the BTC-USD perp on Hyperliquid.

In short, on HIP-3, a deployer has the full control over their oracle selection. While HyENA’s use of Chaos Labs’ oracle enables rapid deployment, the oracle’s limited track record in live perp markets introduces an additional risk factor that traders should be aware of.

If you are a risk-averse person, trading on HyENA HIP-3 should be taken into consideration. But vice versa, this perp DEX can be a suitable option for risk-takers.

CONCLUSION

Based on the analysis above, HyENA should not be viewed merely as another perp DEX on Hyperliquid, but as a real-world stress test for the HIP-3 framework and USDe-based margin design.

Its capital-efficient structure shows how perps can evolve beyond pure speculation into yield-generating infrastructure. Yet, this efficiency comes with new layers of risk - collateral stability and oracle reliability - that traders can no longer afford to ignore.

Whether HyENA becomes a blueprint for future HIP-3 deployments or a cautionary example will depend not on narrative, but on how these risks unfold in real market conditions.On the other hand, from the trading perspective, besides the profitability, traders also need to pay attention to some potential concerns related to the collateral’s depegging and oracle risk.

Disclaimer

This article is an independent educational analysis of HyENA and its deployment within the HIP-3 framework on Hyperliquid. Our goal is to help readers understand how HyENA’s capital-efficient design, USDe-based margin, and oracle dependencies function in practice, and from there, form their own perspective on the associated opportunities and risks.

All concepts are presented in a clear and approachable manner, without assuming deep prior knowledge of HIP-3 mechanics or advanced perpetual trading infrastructure. The content is provided for informational purposes only, and readers are encouraged to conduct their own research before engaging in any trading or DeFi activity.

Leave a comment

INTRODUCTION

Impermanent loss is one of the most confusing concepts in DeFi. Many people hear the term, experience the result, but never fully understand where it comes from and whether LP trading fee can cover this loss.

In part I of this series, we take a step back and look at impermanent loss from its mathematical roots. From there, we explain why Yield Basis uses a carefully designed 2× leveraged structure to neutralize this effect and how protocols keeps this leverage overtime.

KEY TAKEAWAYS

This part I explains impermanent loss from first principles and shows how Yield Basis redesigns the payoff structure to eliminate it at the source.

• Section 1: Explaining how impermanent loss naturally arises from traditional AMMs and examining the coverage ability of tradining fee.

• Section 2: The basis of YieldBasis innovation to solve impermanent loss at its core

In traditional finance, idle capital usually leads to two familiar choices: speculation (such as buying gold and waiting for price appreciation) or earning yield by depositing it in a bank.

Crypto follows the same logic. With idle tokens, you generally choose between:

(1) Holding them to gain from price movements or

(2) Providing liquidity to earn trading fees.

This trade-off causes opportunity cost, it is also where impermanent loss comes in.

SECTION 1: IMPERMANENT LOSS – A RESULT OF BONDING CURVE

Impermanent loss doesn’t mean you lose tokens. In practice, it represents the opportunity cost of providing liquidity: When prices move, the AMM of liquidity pool automatically rebalances your position. This rebalancing can leave your liquidity position worth less than if you had just held the assets outside the pool. That gap is what we call impermanent loss.

To better understand how “your liquidity position worth less than if you had just held the assets outside the pool”, we start by building a graph that compares two strategies: holding assets versus providing liquidity, over different price levels

• The x-axis represents the price change of the asset (for example: BTC) relative to the starting point

• The y-axis represents the position’s total value change, showing how much the position is worth relative to the starting point

Each strategy gives different relationship of x and y.

Strategy 1: Just holding a 50/50 portfolio of assets (for example: BTC and USDC)

This strategy means initially, the total value of BTC holding is equal to the total value of USDC (50:50 ratio).

Therefore, initial total value of portfolio (V_initial) is defined:

Suppose that BTC price increase x times, we have new total value of portfolio (V_new) is:

since USDC stay the same, we have

Therefore:

Because

after transformation, we have

or

(linear line)

Strategy 2: Liquidity providing with the same 50:50 proportion (for example: BTC and USDC) in traditional AMM X*Y=K.

Important note: In the formulas below, we deliberately exclude trading fees, incentives, and rewards. This is done to isolate the pure effect of AMM rebalancing and price movement. Fees and incentives vary across protocols and market conditions, and including them would obscure the core mechanism behind impermanent loss.

Consider a constant-product AMM with:

• A = quantity of BTC in the pool

• B = quantity of USDC in the pool

The AMM enforces the invariant:

Therefore,

BTC Price in terms of USDC =

Then:

From (1) and (2), we have:

The total value of the liquidity position (V) is the sum of the BTC and USDC components:

Replace (2), we have:

When asset price increases x times:

Therefore,

(curve)

Now, let’s perform both strategy on the same graph:

• Blue dashed line (Holding a 50/50 portfolio): This represents the case where you do NOT provide liquidity. You simply hold 50% of one asset and 50% of the other in your wallet (for example, 50% WBTC and 50% USDC).

  This line is linear (for example, if the price of $BTC doubles, the value of your $BTC portion also doubles). It’s very straightforward and intuitive.

• Black solid line (LP value in a X*Y=K AMM): This represents the case where you provide liquidity to a standard AMM pool (such as Uniswap V2), which follows the X*Y=K formula.

  This line is curved (it follows a square-root shape).

Then, it’s time to consider the “difference” between the black line and the blue dashed line. Theoretically, this difference can be either a loss or a gain, depending on how the two lines are positioned relative to each other.

To examine the difference between the two curves, we can apply the AM–GM inequality, which can be simply written as:

With a=x (x is price ratio so x ≥ 0 under all conditions), b=1, we have:

<=>

This means that the black line (LP) always stays below the blue dashed line (HODL) as the price moves further away from the starting point. The difference, therefore, turns into the impermanent loss that we have mentioned before.

No matter whether the price goes up or down, keeping your funds in a liquidity pool results in a lower total value than if you had just held the same assets in your own wallet. Moreover, this loss becomes increasingly pronounced as the asset price continues to rise.

This gap is exactly what we call Impermanent Loss, and essentially, it should be understand as the opportunity cost due to selling assets too early. The main cause of it is the automatic rebalancing mechanism – the core concept of AMM.

Core concept: In a classic AMM, the core formula is:

or

Amount of Token A (x) × Amount of Token B (y) = Constant (k)

Assume you have a liquidity pool made up of BTC and USDC. The automatic rebalancing mechanism always tries to keep the total value of BTC in the pool equal to the total value of USDC (50:50 ratio).

When BTC’s price on the open market increases, the BTC price inside the AMM pool is temporarily lower.

This creates an arbitrage opportunity: Traders notice that BTC in your pool is cheaper than the market price, so they step in, sell USDC for your pool, buy BTC at that a lower price, and sell it elsewhere for a profit.

Therefore, the pool’s balance changes: number of BTC goes down, and number of USDC goes up, while their value is equal.

Similarly, whenever the price of either token in the pool goes up or down, new arbitrage opportunities are created. These arbitrage trades are exactly how the AMM continuously rebalances the pool, without the need for an order book or any intermediary.

The problem is: An AMM does not wait until the price reaches the top to sell. Instead, it sells gradually at every price step as the price goes up. That is exactly why LPs can never achieve the maximum profit compared to simply holding the asset and selling it at the peak.

For example: You start providing liquidity in an AMM pool at the BTC price of $10,000. After a period of time, BTC price rises four times to $40,000 (x=4). Considering the same amount of BTC and USDC at the initial point, this increase in BTC price causes:

• 50/50 holding portfolio increases 2.5 times

• LP position increases 2 times only.

Theoretically, AMM liquidity providers earn trading fees from pool activity. In reality, liquidity provision is most attractive in stable market environments, where fees are sufficient to offset impermanent loss and generate positive returns.

Whether trading fees can make up for impermanent loss depends on how much prices move and how much trading happens in the pool at a given time. Many studies have looked at whether trading fees can cover impermanent loss for liquidity providers, and most of them test this by running repeated simulations.

In this article, we build a simple assumption-based model that estimates the daily performance of being an LP of $5000 WBTC and $5000 USDC in the Uniswap v2 WBTC/USDC pool over two years, compared to simply holding the same 50/50 WBTC–USDC portfolio.

With

• Price move metric (x): each day it takes the WBTC price in USD and computes

meaning how many times WBTC has moved versus the start date.

• TVL: for each day, it takes the latest pool reserves before the day ends and converts them to USD to get TVL/day. TVL is a “stock” number (what’s in the pool at a point in time), and it’s used as the denominator for fee yield.

• Volume (pool-level, from Swap events): for each day, it sums all swaps that actually happened inside this specific pool using the pool’s own Swap events.

• Daily fee yield: it estimates daily LP fees as %fee/day = 0.3% × (volume / TVL) (0.3% is Uniswap v2’s fee rate). Fees are then accumulated using compounding.

• Impermanent loss: with a price move , Uniswap v2’s no-fee LP value scales like , while a 50/50 hold scales like . So:

So , the model’s final comparison is

If net < 0, estimated fees don’t offset estimated IL; if net > 0, they do.

In this test case, we can see that as the volatility factor x increases (meaning the BTC price rises further from the starting point), trading fees become less effective at covering impermanent loss, and the resulting loss grows larger.

SECTION 2: HOW TO SOLVE IMPERMANENT LOSS AT ITS CORE? - YIELDBASIS SOLUTION

As showing in the model above, we can see that in traditional AMMs, your position is continuously rebalanced between BTC and stablecoins. That makes your value only grows as √x when BTC price increases x time, which is slower than x itself.

For example, if the price of BTC increases four times (x = 4), the value of your position increases by only two times (√4 = 2) , resulting in an underperformance relative to holding 50:50 portfolio (increases 2.5 times) and just holding the same amount of BTC only (increase completely 4 times).

Not token incentive, impermanent loss can only be solve at the root if position value can grow linearly with BTC price, especially, the BTC value in position increase (1:1) with BTC price (y=x line).

Including trading fee (calculated by %fee * LP value), we have position value change function in traditional AMM is:

To reach zero impermanent loss, the square-root term in x must be eliminated, or in other words, protocols need transform the payoff from √x to .

To address this, Yield Basis’ concept is to apply leverage to effectively square the payoff function:

Keep going, we have:

Fee rate is usually small so the quadratic term is negligible, allowing this approximation.

By squaring the square-root term, the position returns to linear price exposure. In other words, the asset value moves (1:1) with the price: when the price increases by a certain factor, the position value increases by the same factor, just like holding the asset outright.

How does this concept work in reality?

In simple terms, 2× leverage means starting with one unit of capital, borrowing one additional unit against it, and depositing the full two units into the liquidity pool. This leverage removes the square-root effect of AMMs, allowing the position to track BTC price movements more closely while continuing to earn trading fees.

This intuition is exactly what Yield Basis is designed to capture in practice. By using the initial capital as collateral to borrow stablecoins ($crvUSD) and using the original capital together with the borrowed stablecoins to provide liquidity to the pool, the protocol effectively squares the LP payoff function.

We will deeply break down Yield Basis concept to resolve impermanent loss in part II of this article.

CONCLUSION

Impermanent loss is not caused by fees, incentives, or bad market timing. It is a structural result of the constant-product AMM curve, which forces liquidity positions to grow with the square root of price changes rather than linearly.

Yield Basis addresses this problem at its root. By maintaining a constant 2× leverage through automated borrowing, repaying, and rebalancing, the protocol transforms the LP payoff from √x back into x. The result is a position that behaves like simply holding the asset, while still earning trading fees.

DISCLAIMER

In this article, we focus on breaking down impermanent loss from its mathematical foundations and explaining why it naturally arises in traditional AMM designs. Our goal is to help readers clearly understand the structural trade-offs faced by liquidity providers, and from there, build their own intuition about what it would take to eliminate impermanent loss at its root.

We do not avoid formulas or quantitative reasoning, but every step is presented in a simple and accessible way, supported by clear assumptions and intuitive comparisons, so that the underlying mechanics are easy to follow.

In Part II, we will apply this framework to Yield Basis, unpacking how its leveraged and automated rebalancing design reshapes the AMM curve. By grounding the discussion in first principles, we aim to rigorously examine whether Yield Basis can truly neutralize impermanent loss, and under what conditions this approach holds in practice.

Leave a comment

INTRODUCTION

DeFi transactions settle quickly and protocol metrics update continuously. RWAs do not operate the same way. Real-world assets are not repriced every block, their valuations are typically published on a schedule (NAV reports), and their transferability can be restricted by compliance rules such as allowlists and settlement requirements.

Aave Labs’ Horizon (launched March 2025) combines these two systems in one lending market. The stablecoin side behaves like a normal on-chain money market. The RWA collateral side depends on off-chain custody, legal structures, and periodic valuation updates. Because these components update at different speeds and under different constraints, key DeFi expectations are not guaranteed. Withdrawals may be delayed by off-chain processes, liquidations may depend on stale or delayed valuations, and price inputs may reflect reported values rather than real-time market discovery.

This article analyzes Horizon as a mechanism. It focuses on how yield is generated and how collateral value is defined and maintained under these constraints.

KEY TAKEAWAYS

• Section 1 : Stablecoin Supplier: Supplier yield = borrower interest, driven by utilization/rate model; incentives only adjust the displayed APY.

• Section 2 : Borrower: Borrowing is allowlisted and NAV-priced; liquidations can be NAV-event-driven and collateral may be transfer-restricted.

• Section 3 : Issuer/SPV/Custody: Collateral value comes from an off-chain NAV pipeline (issuer/SPV/custody → NAV → oracle).

OVERVIEW: WHAT HORIZON REALLY IS

Horizon is Aave’s RWA lending market: anyone can supply stablecoins, but only allowlisted institutions can borrow, because the collateral is permissioned tokenized (Real-World Asset tokens with transfer and access restrictions).

In the simplest terms:

Horizon is like a “DeFi bank”: deposit options are open to everyone, but collateral and borrowing options are only open to verified users.

HORIZON SPLITS THE MARKET INTO TWO HALVES.

The first half is RWA collateral. This side is controlled. Only institutions that have been approved and allowlisted by the issuer can supply RWA tokens as collateral.

The second half is stablecoin liquidity. This side is open. Regular users can supply stablecoins such as USDC and earn yield.

The intersection is borrowing. You can only borrow from the stablecoin pools if you are part of the group that has supplied valid RWA collateral.

We will explain more detail in next part, but totally, in plain terms, Horizon lets anyone deposit, but borrowing against RWA collateral is restricted to verified users.

So, from here on, we use borrowers to refer to allowlisted institutions that supply RWA collateral and borrow stablecoins from Horizon.
We use stablecoin suppliers to refer to users who deposit stablecoins (e.g., USDC) into the pool to provide liquidity and earn yield.

HOW HORIZON DIFFERS FROM A NORMAL AAVE MARKET

1) Collateral access is restricted

On Aave Core, most collateral is crypto-native (e.g., ETH, wBTC, major native tokens) and anyone can use it.

On Horizon, collateral is RWA ( e.g., VBILL, JAAA ) , so only approved participants can supply it.

2) Collateral price is NAV-based, not purely market-based

Aave Core collateral is typically priced by market-reflective feeds. ( “Market-reflective feeds” are price oracles that track prices formed in active markets (DEX/CEX) and update frequently based on real trading. In other words, the collateral price is tied to continuous market discovery, not periodic NAV-style reporting.)

Horizon RWAs often depend on NAV (Net Asset Value) updates reported by an issuer/NAV agent and published on-chain via an oracle. That means the system depends on an off-chain valuation and reporting cadence, not just on-chain trading.

3) RWA collateral positions may be non-transferable

To keep the market permissioned, the receipt token that represents an RWA collateral position can be made non-transferable. This means it can’t be freely sent or traded like a normal token. It’s not typical DeFi behavior, but it ensures the collateral position stays inside the allowlisted group and can’t be passed to unapproved wallets.

Overall, we use this diagram to summarize Horizon’s end-to-end flow before going into details.

Off-chain, an issuer creates the RWA token through an SPV (Special Purpose Vehicle) that legally owns the underlying assets, while a custodian holds and safekeeps those assets. A NAV provider produces valuation updates, and an oracle publishes NAV on-chain so the protocol has a usable collateral “price truth.”

On-chain, borrowers supply the RWA token as collateral and borrow stablecoins from the Horizon market. In parallel, suppliers can deposit liquidity permissionlessly and earn interest that is ultimately paid by borrowers. Governance and operations are split: Aave DAO sets risk limits, while Aave Labs coordinates day-to-day execution and partner operations.

To dive into Horizon’s internal mechanics, in the next section, we will analyze it through three perspectives: the stablecoin supplier, the borrower, and the RWA issuer. This approach helps readers understand how the system works end to end, where yields and risks actually come from, and which parts depend on on-chain logic versus off-chain processes.

SECTION 1 : PERSPECTIVE 1 - A STABLECOIN SUPPLIER

From a supplier’s view, Horizon feels like a normal Aave.

You supply USDC, receive an aToken, and see an APY on the screen. But the key point is simple: as a stablecoin supplier, you are not earning Treasury coupons. You are earning interest paid by borrowers, exactly like any lending pool.

So what should you actually watch?

• Utilization (pool usage): if borrowing demand is low, your supplied liquidity sits idle and yield drops.

• Borrow rates: when utilization rises, borrowers pay more, and that is what flows back to suppliers.

• Liquidity on withdrawal: when you withdraw, you are withdrawing from the pool’s available liquidity, not from an SPV selling assets for you. If the pool is heavily utilized, withdrawals can feel tighter for the same reason they can on Aave.

Therefore, to better understand how interest is calculated, let’s break down exactly how stablecoin suppliers earn yield in Horizon.

UTILIZATION

Utilization (U) measures how much of the supplied stablecoin liquidity is currently borrowed.

U = Total Borrowed / Total Supplied

Utilization over time for the Horizon USDC pool

Interest rates react to utilization through two operating zones on the rate curve:

• Normal zone (below the kink): ~0% to ~90% utilization.
  In this range, borrow rates rise gradually as utilization increases. In the chart, the pool’s Current U is 81.14%, which is still below the Optimal ~90%, so it is operating in the normal zone.

• Protection zone (above the kink): > ~90% utilization.
  Once utilization moves past ~90%, borrow rates ramp up much faster (the curve steepens sharply). The goal is to protect remaining liquidity and push utilization back down by making borrowing more expensive.

• Kink / breakpoint: ~90% utilization on this chart.
  This is the threshold where the interest-rate model switches from “gradual increase” to “aggressive increase.”

Because Horizon only allows allowlisted institutions to borrow, its utilization reflects borrowing demand from that restricted group—not from the broader user base.
When utilization rises, it means approved borrowers are actively borrowing stablecoins against their RWA collateral. Because Horizon only has a small, restricted set of borrowers, utilization can move sharply depending on when a few large borrowers borrow or repay, rather than changing smoothly with broad retail activity. When one or two large borrowers become active, add collateral, or receive updated NAV-based borrowing power, utilization can shift noticeably in a short period.

BORROW APY AND SUPPLY APY: THE REAL MECHANICS BEHIND THE NUMBERS

BORROW APY

Borrow APY (Annual Percentage Yield ) is the interest rate borrowers pay to the pool when they borrow USDC (or another stablecoin). It is not manually set by a team. It is the output of the pool’s interest-rate model, which reacts to how “tight” liquidity is.

• When liquidity is abundant, borrowing is cheaper to encourage demand.

• When liquidity becomes scarce (high utilization), borrowing gets more expensive to protect liquidity and rebalance the pool.

• Borrow APY depends on Utilization

Where is the pool’s interest-rate curve (it increases as utilization rises).

SUPPLY APY

Supply APY is what suppliers earn for depositing USDC into the pool. The baseline source of supply yield is simple: borrowers pay interest, and suppliers receive a share of it.

Supply APY is usually lower than Borrow APY for two reasons:

1. Only the portion of liquidity that is actually borrowed generates interest (utilization matters).

2. The protocol keeps a portion of borrower interest as reserves (reserve factor).

This explains why numbers “make sense together.”
Example intuition: if borrow APY is ~7–8%, utilization is high, and reserve factor > 0, base supply can naturally land around ~5–6%.

HOW INTEREST ACCRUES IN HORIZON

Horizon records interest the same way Aave does: through aTokens (for suppliers) and debt tokens (for borrowers), not through manual interest payments.

If you supply: you receive an aToken that represents your deposit. Your yield accrues automatically through the aToken mechanics.

• Example: supply VBILL => receive aHorRwaVBILL.

If you borrow: the protocol mints a debt token that represents what you owe. Your debt grows over time as interest accrues.

• Example: borrow USDC => you receive USDC, and your USDC debt token balance increases over time.

APY AND APR

Aave’s UI can show both APR and APY, which is easy to confuse.

APR is a simple annual rate that does not include compounding, while APY is the effective annual rate after compounding.

In practice, Aave interest accrues continuously: lenders’ aTokens grow over time and borrowers’ debt grows over time. So the most accurate way to read the UI is: stablecoin suppliers earn Supply APY, and borrowers pay Borrow APY. If APR is also shown, treat it as a simplified, non-compounded version of the same rate.

WHY THE UI CAN SHOW BORROW APY < SUPPLY APY ?

This is the confusing part. The interface can show different types of numbers for borrowing vs supplying because of incentives (rebates) such as Merkl.

Borrow: The UI may show the “net cost” after rebates

Borrow can have two layers:

• Protocol Borrow APY: the raw interest rate produced by the pool

• Borrow incentives (rebates): extra rewards for borrowers (e.g., via Merkl) that reduce the effective cost

Merkl incentives are third-party rewards designed to push specific behavior in the market, such as borrowing or supplying. They are separate from Horizon’s core lending engine: Merkl does not modify the pool’s protocol interest-rate curve or how interest accrues inside the pool. Instead, Merkl adds an external reward layer on top of the protocol rates. When borrowers receive incentives, their effective cost can be lower than the protocol borrow rate (because rewards offset part of the interest). When suppliers receive incentives, the APY shown on the interface can be higher than the base supply yield (because rewards are added on top)

So what the borrower effectively feels is:

Example : In the screenshot, the borrower’s effective rate is:

Supply: The UI may show “base + incentives”

Supply can also include incentives:

• Base Supply APY: from borrower interest (the lending engine)

• Supply incentives: extra rewards paid to suppliers (campaigns, subsidies, emissions)

So the displayed number may be:

Example : In the screenshot, the borrower’s effective rate is:

Therefore, Horizon’s APYs are not contradictory: the protocol sets borrow and supply rates through utilization and reserve factor, while Merkl incentives can rebate borrowers or boost suppliers. When Borrow APY looks lower than Supply APY on the UI, it’s usually because incentives are being mixed into the displayed numbers, not because the lending math is broken.

SECTION 2: PERSPECTIVE 2 – THE BORROWER

On the borrower side, Horizon is not a “supply and earn” product. It is a gated borrowing workflow. Institutions first go through onboarding steps such as KYC (Know Your Customer - identity verification), legal agreements, issuer approval, and allowlisting (being added to an approved-access list).

Only after that can they deposit permissioned RWA token as collateral and borrow stablecoins from the pool.

Once collateral is deposited, the position follows the same lending risk logic as Aave, but the key inputs behave differently. Borrowing power is defined by LTV (Loan-to-Value, the maximum borrowable percentage of collateral value) and liquidation thresholds (the point where a position becomes eligible for liquidation).

The difference is the collateral “price.” For many RWAs, value is not discovered on a DEX in real time. It is driven by NAV (Net Asset Value, the reported per-token value of the off-chain asset portfolio) updates that are produced off-chain and then published on-chain.

So the next step is to break down the borrower mechanics that actually move Horizon’s risk: how LTV is set, adjusted and what events push a position toward liquidation (forced debt repayment + collateral seizure).

LTV ( LOAN-TO-VALUE)

In Horizon, the borrower (an allowlisted institution) starts by depositing an RWA token as collateral and then borrowing stablecoins. That is why LTV is one of the first parameters that matters on the borrower side.

LTV in Horizon is not calculated differently from Aave Core. It is the same concept and the same basic logic: maximum borrowable value as a percentage of the collateral value.

Max LTV is the initial borrowing cap: at position opening, it defines how much the borrower can borrow relative to the collateral’s value (as determined by the oracle price, which usually reflects the latest NAV update).

Example

VBILL shows Max LTV = 88% , Oracle price = $1.00. If an allowlisted borrower deposits VBILL worth $10,000, the maximum initial borrow is approximately:

LIQUIDATION IN HORIZON: SAME TRIGGER, DIFFERENT CONSTRAINTS

After LTV, the next borrower-side mechanism that matters is liquidation. Horizon uses Aave’s standard risk logic, but the inputs and constraints are different.

1) The trigger is the same: health factor below threshold

In both Aave Core and Horizon, a position becomes liquidatable when its health factor falls below the liquidation threshold. This can happen because:

• Debt grows over time from interest, or

• Collateral value drops (in Horizon, this is often driven by NAV updates rather than real-time market prices).

2) The key difference is “price truth”

Aave Core (crypto collateral): collateral pricing is market-reflective and updates frequently via oracles that follow active trading venues. Health factor moves continuously with the market.

Horizon (RWA collateral): many collateral types are valued using NAV updates that are computed off-chain and then published on-chain. That changes liquidation dynamics:

• A position can look healthy under the previous NAV, then become liquidatable immediately after a new NAV is posted.

• The dominant risk is not normal volatility; it is bad NAV reaching the chain (late, incorrect, or outlier values).

3) Execution is more constrained because collateral may be permissioned

In standard DeFi liquidation, collateral is typically a freely transferable ERC-20, so the loop is simple: liquidator repays debt → receives collateral → sells on a DEX → recovers liquidity.

In Horizon, RWA collateral (or its receipt token) can be transfer-restricted for compliance. Practical implications:

• Not every address can liquidate or receive collateral. Liquidators may need to be eligible/allowlisted.

• Even after receiving collateral, the liquidator may not be able to “dump” it immediately on a DEX. Exit may depend on restricted venues, redemption flows, or off-chain processes.

4) Who gains and who loses in liquidation (economic clarity)

• Borrower (loses): collateral is taken at a discount relative to the repaid debt. This is the effective liquidation penalty. The borrower benefits only in one sense: their debt is reduced, preventing the position from becoming insolvent.

• Liquidator (gains): earns the liquidation bonus (collateral received worth more than the debt repaid), but takes on additional constraints (eligibility, exit friction, operational risk).

• Protocol / lenders (benefit): liquidation’s purpose is solvency protection—it prevents bad debt and protects suppliers. The protocol’s “gain” is risk containment and system stability, not necessarily direct profit.

So, we can see overall liquidation process through this diagram :

SECTION 3 : PERSPECTIVE 3 - THE ISSUER, SPV, AND CUSTODY STACK

This is the part the interface does not show you.

Off-chain roles (high level):

• Issuer : designs the product, sets eligibility rules (allowlist/whitelist), and is responsible for ensuring the token is backed by legally owned assets.

• SPV (Special Purpose Vehicle): the legal entity that owns the underlying assets. It is typically structured to be bankruptcy-remote so the assets are ring-fenced from the issuer’s balance sheet.

• Custodian (custodian): the regulated safekeeping layer that actually holds the instruments (T-bills, cash, credit positions) and handles settlement and segregation.

• NAV provider (NAV calculation agent): produces valuation reports for the portfolio on a defined schedule and methodology.

• Oracle : publishes the NAV value on-chain so the protocol can use it as the collateral reference price.

NAV IS THE COLLATERAL PRICE, BUT IT UPDATES ON A SCHEDULE

RWA collateral in Horizon is often priced by NAV, not by continuous on-chain market discovery. At a high level:

Where:

• AUM (Assets Under Management): total market value of what the SPV/fund holds (cash, T-bills, bonds, short-duration notes, private credit).

• Fees: costs charged at the structure level (management, custody, audit, admin).

• Liabilities: unpaid expenses or other obligations at the vehicle level.

• Tokens outstanding: total RWA token supply currently in circulation on-chain.

Two pricing systems run in parallel

Horizon runs two “clocks” at the same time:

1. NAV prices the RWA collateral side on a reporting cadence.

2. The Aave interest-rate model prices stablecoin borrowing in near real time based on utilization.

These two clocks can move at different speeds, which is why the market can look normal on the lending side while the collateral side is operating on periodic updates.

NAV APY is not the supplier’s yield

NAV APY is just an annualized view of NAV movement. It describes the RWA portfolio performance, not what stablecoin suppliers earn. Supplier yield still depends on borrowing activity and the lending engine.

The key risk to track

The main failure mode is bad NAV reaching the chain (late, wrong, or outlier). If the protocol consumes incorrect NAV, it can misprice collateral and trigger incorrect risk outcomes. This is why RWA markets rely on validation, bounds, and emergency controls around NAV publishing.

NAV is an off-chain number, so the protocol cannot “verify” it using on-chain data alone. Instead, RWA markets reduce NAV risk through a controlled publishing process and strict guardrails. Off-chain, NAV is typically calculated and reconciled against custodian statements (often with a fund admin/attestation layer). On-chain, the protocol only accepts NAV updates from authorized signers, and it applies validation rules such as freshness checks (reject stale NAV), deviation bounds (reject large jumps vs. the last NAV), and outlier filters. If an update looks abnormal, emergency controls (pause/freeze) can temporarily stop borrowing or liquidation to prevent mispricing from causing incorrect risk outcomes

A quick example:

A fund holds $2,000 of T-bills and mints 2,000 tokens. NAV starts at 1.00.
After interest accrues, the portfolio becomes $2,010. The new NAV is:

No new tokens are minted. Value increases because each token becomes worth more.

CONCLUSION

We see Horizon as Aave’s attempt to integrate RWAs into a lending market in a way that can work operationally, using permissioned collateral and NAV-based valuation alongside open stablecoin liquidity.

Whether it succeeds is less about the APYs and more about execution quality: the reliability of the off-chain NAV pipeline, the clarity of intervention boundaries, and how these constraints affect liquidity, liquidations, and user expectations under stress.

DISCLAIMER:

This article is intended as a conceptual and educational breakdown of Aave Horizon’s RWA lending design. The focus is on mechanism-level intuition, how rates, utilization, collateral valuation, and liquidation constraints work rather than exhaustive smart contract implementation details.

It is structured as a three-perspective framework (stablecoin supplier, allowlisted borrower, and issuer/SPV/custody) to clarify which parts are on-chain logic versus off-chain operations. A deeper contract-level walkthrough and transaction-level proofs can be presented separately in a follow-up part.

Leave a comment

INTRODUCTION

If you’ve ever used Hyperliquid or other decentralized exchanges (DEXs) such as dYdX, GMX, you may notice that most trading pairs are primarily quoted in USDC on these platforms. Hyperliquid also relies heavily on USDC as collateral for both trading and staking as reflected in the metric below.

Why is this the case? Why not USDT, the largest stablecoin in the entire crypto market?

In this article, we will examine the reasons behind this trend, discuss the potential risks that Hyperliquid may face as a result, and whether Hyperliquid can reduce its high dependency on USDC in the future.

KEY TAKEAWAYS

This article will explain in detail why Hyperliquid DEX chose USDC as its primary quote asset and collateral, via the following main sections

• Section 1: Explore why Hyperliquid mainly uses USDC as the quote asset and collateral but not USDT – the largest stablecoin in the crypto world

• Section 2: The most critical risk that USDC holds, which can pose a threat to Hyperliquid, in reality

• Section 3: Examine whether Hyperliquid can reduce its dependency on USDC through the launch of its native stablecoin (USDH)

BEFORE WE BEGIN

1/ Why Hyperliquid Doesn’t Use ETH Or BTC As The Quote Assets?

Definitely because perps require a collateral of high price stability.

ETH, BTC or other non-stable tokens are widely known with a high volatility, which shouldn’t be used as collateral or quote assets. Therefore, perps DEXs prefer safer quote assets, and stablecoin (like USDC or USDT) is the most suitable option.

2/ USDC & USDT Background

USDC (USD Coin) was launched in 2018 by Circle, a US publicly listed company. As widely known, fiat-backed stablecoin like USDC is redeemable 1:1 for US dollars. If you want 1 USDC, you need to give Circle 1 US dollar, then 1 USDC will be minted and given to you, and vice versa. Currently, this is the second-largest stablecoin with the circulating supply of over $70B.

Meanwhile, USDT (Tether USD) was launched in 2014 by Tether, a fin-tech company headquartered in El Salvador. This is the pioneer in the concept of fiat-backed digital currency pegged 1:1 to traditional one (initially the US dollar). USDT is the largest stablecoin with the circulating supply of over $170B, more than twice that of USDC.

SECTION 1: WHY HYPERLIQUID MAINLY USES USDC AS THE QUOTE ASSET AND COLLATERAL BUT NOT USDT?

There are several reasons behind the decision to choose USDC as the quote asset for most trading pairs, as well as collateral on Hyperliquid and other DEXs. And transparency is the first and foremost factor. Why?

A historical failure like Terra-Luna ($TerraUSD depegging event) demonstrate that even sophisticated token mechanics cannot compensate for a lack of transparency in its reserve structures. The missing of $3.5B in its reserves which were supposed to back the cryptocurrency not only seriously affected related investors but has also shaken confidence in the crypto market as a whole.

Therefore, it is important for e-money issuers to prove their transparency and accountability in order to prevent similar incidents in the future.

1/ Transparency (Main)

A collateral or quote asset must be capable of proving itself to be highly, or more precisely, absolutely stable.

The core problem is not that Hyperliquid itself needs a highly stable asset, but that its participants - traders, LPs, and market makers - do. If the quote asset loses its peg, their positions can quickly become eligible for liquidation. As a rational response, LPs would withdraw liquidity, leading to deteriorating orderbook depth and higher slippage. Ultimately, such a scenario could cripple liquidity and confidence, destabilizing the entire platform.

That’s why a fiat-backed stablecoin must be proven with high transparency to secure trust for both the protocols and users. And in this game, USDC appears to do it better.

The transparency is reflected via two major factors: reserve backing and regulatory compliance.

1.1/Reserve Backing

USDC’s reserve structure is simpler and more transparent because it is almost entirely backed by cash and short-term US Treasuries (main) - two of the most liquid and lowest-risk assets available.

Meanwhile, USDT is backed by more sources, including short-term US Treasuries (main), cash, and other financial products like Bitcoin, corporate bonds, commercial paper, etc as in the picture above. A broader mix does not automatically mean higher risk, but in USDT’s case, the quality, liquidity, and transparency of some asset classes were significantly lower.

According to S&P, Bitcoin accounts for 5.6% of all USDT in circulation, which surpasses the 3.9% reserve implied by Tether’s reported collateral ratio of 103.9%. Because these assets are more volatile than cash or treasuries, a sharp decline could potentially undermine their collateral value.

Therefore, the market perceives USDT as having higher risk not because it has more types of collateral, but because parts of its collateral are less transparent, less liquid, and more volatile compared to USDC’s reserves.

Also, as shown in the S&P Global’s analysis, in the normal condition of market on the weekdays, USDT showed a slightly higher depegging frequency and magnitude than USDC the 24-month data through mid-2023. The increasing holding of high-risk assets is also one of the reasons for this volatility.

1.2/ Regulatory Compliance

Circle is widely known as a regulation-first company. Registered and operating under the financial regulatory framework of the US, Circle has been publishing audits per month to prove that their reserves are always sufficient to fully back USDC. Currently, these attestations are being performed by Deloitte, the largest Big Four accounting firm in the world.

About Tether, operating in El Salvador - a regulatory environment that differs from US and EU frameworks – leads to fewer mandatory disclosure and reporting obligations than regulated stablecoin issuers like Circle. Despite being launched in 2014, Tether has long faced skepticism due to the lack of sufficient financial attestations regarding its reserves.

To address these concerns, Tether has begun publishing its Reserves reports since 2017 and quarterly since 2021, but the reporting frequency remains lower than Circle’s, and so far, Tether has not yet completed a full audit of its reserves, in fact.

In the past, this was also one of the reasons why USDT failed to meet several requirements for E-money tokens under the European Union’s Markets in Crypto-Assets Regulation (MiCA) framework, including reserve composition, redemption rights, and regulatory supervision within the EU. As such, exchanges in Europe had to delist USDT, effectively limiting its use within certain European markets. Even Binance, Kraken and Crypto.com planned to delist USDT in the EU market.

The delisting means EU users can’t trade USDT anymore. Though the EU region represents a relatively small volume of USDT trading, regulatory compliance is important for a stablecoin to prove their transparency and get wider adoption. If e-money law like this applied in the Asia, where USDT accounts for 45% of global volume, this would impact Tether’s and users’ overall business.

On the other hand, USDC is MiCA compliant and continues to improve its credibility among users and regulators.

Hyperliquid is a DEX; they may not need to care about these complex laws, but their users can. From a risk management perspective, regulatory compliance helps mitigate the uncertainty about reserve treatment and information disclosure frequency. For large-scale LPs and market makers, these factors determine whether a stablecoin is accepted as collateral and affect how risk assumptions are designed when using it as margin, regardless of on CEXs or DEXs.

Hence, from a legal standpoint, Circle manages to prove its transparency and accountability, whereas Tether remains rather opaque.

In addition to the transparency, another factor to consider is the USDC adoption rate in DeFi (network effect).

2/ Network Effect: USDC Adoption in DeFi

Since 2020, the heart of DeFi summer, USDC has appeared more frequently on DeFi protocols like AAVE or Compound (the most deposited and borrowed stablecoin) and offered greater APR for LPs. USDC was considered the leading stablecoin for collateralization demands, even surpassing USDT and DAI.

When it comes to Hyperliquid or other DEXs, USDC continues to become the main character in their DeFi game. AAVE and Compound prioritizes USDC, GMX and dYdX also opts for USDC, and so does Hyperliquid. It’s just a network effect because every protocol must be optimized for large liquidity ecosystems.

IF USDC IS SUCH TRANSPARENT, WHY CEXS PREFER USDT?

It’s easy to notice a clear pattern in the market: DEXs tend to prefer USDC, while CEXs lean heavily toward USDT. Even though USDC is generally viewed as more transparent and MiCA-compliant, major CEXs like Binance and Bybit still favor listing trading pairs quoted in USDT rather than USDC.

According to Cryptorank, Binance’s spot market features more than 400 USDT-quoted pairs - roughly double the number of USDC pairs. Similarly, on Bybit, nearly 500 spot trading pairs use USDT as the quote asset, while USDC-quoted pairs remain relatively limited at around 100.

Why is this the case?

To answer this question, we need to look back at history.

USDT (2014) has a longer history, born earlier than USDC (2018). CEXs were also born earlier than DEXs. Since the initial launch, USDT has partnered with many pioneer CEXs in the world like Bitfinex to expand their influence and adoption. Time passes by, users on CEXs are accustomed to trading USDT, especially via OTC or P2P mechanism.

And so, gradually, USDT becomes a fiat-backed stablecoin with a higher liquidity and market share.

To be specific, USDT’s total supply is roughly double that of USDC, giving it over 60% of the entire stablecoin market. CEXs prefer USDT since its global liquidity is the biggest, which is important to maintain a rich order book and support high trading volume.

So, with the first-mover advantage, USDT remains the preferred choice of CEXs to this day.

SECTION 2: IS USDC STILL THE BEST CHOICE?

However, with all these above advantages, USDC also comes with a certain potential risk that could affect the performance of Hyperliquid and other DEXs. The most critical risk is related to depegging.

Depegging risk

Deviating significantly from the pegged value is what people in the crypto world fear most when it comes to stablecoin. In the past, USDC used to be depegged and drop to $0.87 in 2023.

The collapse of Silicon Valley Bank (SVB) is the largest bank failure in the US since the global financial crisis in 2008. Prior to this incident, Circle had a portion of its reserve kept in SVB, accounting for about $3.3B. On March 10th, 2023, SVB was suddenly closed and the news that Circle deposits were temporarily stuck and could not be withdrawn spread widely in the crypto world. At once, investors withdrew their USDC’s holdings, transferring to USDT or fiat currency, causing the price of USDC on most exchanges bottom at $0.87 (~13%), recording a severe depegging event of a stablecoin.

Fortunately, on March 12th, US Treasury Department announced that all SVB deposits will be 100% protected, and provide assurances for Circle. It means that $3.3B deposit of Circle could be refunded. This announcement immediately saved USDC and made it bounce back to the pegged value.

Despite being trusted for its transparency and strong compliance standards, USDC still experienced a depegging event, a serious one indeed.

Also, if we look at the price levels of these two stablecoins in 2 years since June 2021, it shows that USDT is more stable than USDC in terms of overall volatility range. USDT often deviates by about 0.05% from its threshold, while USDC, in reality, once faced the most severe depegging event caused by the SVB collapse.

As can be seen in the picture above, Hyperliquid’s inflow of USDC (via Arbitrum bridge only - also the main network on Hyperliquid) has been increasing rapidly from its initial launch despite some ups and downs, notably with the peak of about $6B in September 2025.

If such a depegging event were to recur, it would have a substantial impact on Hyperliquid.

Overall, this incident raised concerns about the stability of stablecoins in general and USDC in particular, as well as requires certain approaches from DEXs.

SECTION 3: CAN HYPERLIQUID REDUCE ITS DEPENDENCY ON USDC?

With its ambition to become the “House of All Finance,” Hyperliquid has been steadily rolling out new pillars of this ecosystem, including lending, yield products, staking, bridging, and most recently, its native stablecoin (USDH) launched in September 2025. This stablecoin’s introduction also serves as a strategic hedge against the depegging risk of USDC, which accounts for 95% stablecoin deposits on this platform.

USDH is also a fiat-backed stablecoin but issued on HyperEVM. Similar to USDC, this native stablecoin is fully backed by cash and US Treasury equivalents. Unlike USDC whose yield flow goes to its issuer (Circle), USDH directs 50% of earnings to $HYPE buybacks and 50% to ecosystem development funds.

Since the launch of USDH, the dominance of USDC on HyperEVM has significantly declined. At its peak in the beginning of December 2025, USDH supply ($60M) ranked second with nearly 18%. However, this number soon decreased to about 8%, leaving USDC as the dominant stablecoin over HyperEVM, currently holding over 55%.

Note: This figure only considers stablecoin supply on HyperEVM as USDH is launched on HyperEVM. Meanwhile, the amount of USDC ($4B) still lies on HyperCore Arbitrum bridge now.

Moreover, in the medium term, the dominance of USDC on HyperEVM is predicted to increase further in the future as more and more USDC on HyperCore is being bridged to HyperEVM due to the link between them. But in the long term, USDH still shows positive prospects for wider adoption as we look at the increasing USDH HyperEVM supply in the metric below.

In general, as just launched recently, the stability and transparency of this token still need more time to test. Hyperliquid still needs a long time to prove the security of its native stablecoin in order to gain wider adoption and reduce its dependence on USDC.

CONCLUSION

To sum up, DEXs like Hyperliquid prefer USDC as collateral and quote asset for trading pairs because of its greater transparency and stronger regulatory compliance compared to USDT. Additionally, due to a network-effect dynamic, Hyperliquid continues to adopt USDC as its primary stablecoin after early DeFi partners helped establish it as the standard.

However, the past depegging event revealed that USDC is not entirely risk-free for DEXs to rely on. Hyperliquid’s native stablecoin, USDH, was introduced to reduce this dependency, but it will still require time and real-world stress-testing before proving its resilience.

Disclaimer

This article is an independent educational analysis of stablecoin selection on decentralized exchanges, with a focus on why protocols like Hyperliquid currently prefer USDC over USDT as a quote asset and collateral. Our goal is to help readers understand the underlying design considerations - including transparency and network effect - and from there, form their own view on the trade-offs in stablecoin selection.

All concepts will be explained clearly and understandably, without requiring in-depth knowledge of stablecoin mechanics or exchange infrastructure.

Leave a comment

INTRODUCTION

AAVE V4 introduces Risk Premium as a core component of its lending architecture, fundamentally changing how borrowing costs are determined in a shared-liquidity system.

This article takes a deep dive into AAVE V4’s Risk Premium mechanism, explaining how borrowing costs are adjusted under the hood and uncovering the less visible flows and accounting logic that are not immediately apparent from the user interface.

KEY TAKEAWAYS

This article focuses on the mechanics behind Risk Premium in AAVE V4, breaking down how it is calculated, how it evolves over time, and how it turns into actual interest accrual. The main flow will be:

• Section 1: Collateral risk in previous AAVE models and the motivation Risk Premium in AAVE V4

• Section 2: A deep breakdown of the core Risk Premium algorithm, covering its key components, logical steps, and how position-level risk changes final borrowing costs.

  ---

SECTION 1: PREVIOUS AAVE AND COLLATERAL RISKS

1/ AAVE previous lending models: High-level design

a) Collateral rehypothecation (Collateral Reusing)

Previous versions of AAVE use the collateral rehypothecation model to improve capital efficiency, which means: instead of locking collateral in escrow, assets deposited as collateral can also be lent out to other borrowers at the same time. This allows the same unit of capital to support both borrowing and lending activities.

When collateral is held in escrow, it means the borrower’s collateral assets are locked in a smart contract and are not used for anything else during the loan period.

This setup is convenient since the collateral is always available for liquidation or pay back and withdraw. However, escrowing collateral is inefficient in terms of capital usage. While the collateral is locked, it does not generate any yield or serve any productive purpose, meaning borrowers earn no return on the assets they have pledged.

The most effective form of collateral rehypothecation is one that maximizes capital efficiency for both users and the protocol: the protocol can reuse user collateral as lending liquidity, while users are not capital-locked and still retain full economic ownership of their assets.

This design is enabled on AAVE by collateral tokenization, where deposited collateral is converted into aToken - a yield-bearing rebasing tokens that represent both the underlying asset and its accrued interest.

When users deposit assets into AAVE, the protocol mints corresponding aTokens at a 1:1 ratio. The underlying assets are used for lending, generating yield from borrowers’ interest; while user can claim their yield, redeem them at any time for the underlying asset at a 1:1 ratio, use them as collateral to borrow, transfer them to other wallets,...

The yield from borrowing is distributed automatically by increasing users’ aToken balances over time, while the aToken price remains pegged to the original asset.

For example: You deposit $100 WBTC into AAVE, then protocol mints $100 aWBTC and sends them to your wallet.

At this point, you have:

• Your balance: $100 aWBTC

• Each $1 aWBTC is redeemable for $1 WBTC

Over time, borrowers pay interest to the pool. Instead of increasing the price of aWBTC, AAVE distributes this yield by increasing your aWBTC balance. And after some time, your balance becomes $105 aWBTC.

When you decide to withdraw: AAVE burns your $105 aWBTC and you receive $105 WBTC in return. This means the extra $5 WBTC comes from interest earned while your assets were supplied to the protocol.

b) Collateral Clusters

Because Aave allows collateral to actively be reused as liquidity for other borrowers, the protocol groups multiple assets into a single shared lending market, forming what is theoretically known as a “collateral cluster”. And in Aave, each market functions as a collateral cluster. A market defines a group of assets that share the same liquidity pool, interest rate logic, and liquidation system, all of which can be used as collateral and borrowed against one another.

However, rehypothecation also means that assets are economically connected. If one collateral asset becomes risky, its impact can propagate through the shared pool. As a result, collateral clusters inherently socialize risk across the market. To manage this trade-off, AAVE restricts clusters to highly liquid, well-tested assets and applies conservative risk parameters to preserve system stability.

2/ Collateral risk problems

Because each AAVE market functions as a single collateral cluster, all assets within the market share the same pool of liquidity. As a result, borrowing cost (Borrow APY) is based on the supply and demand of the borrowed asset within that shared pool. Or in other words, interest rates are determined by utilization: when demand for a particular asset increases relative to available liquidity, its borrowing rate rises, and vice versa.

This means that borrowers with very different collateral risk profiles often pay similar interest rates as long as they borrow the same asset.

Collateral Risk Profiles managed by AAVE V3

Of course, previous versions of AAVE, like AAVE V3, do not completely ignore this risk. Key parameters such as the maximum loan-to-value (LTV), liquidation threshold, and liquidation penalty are all determined by the collateral asset used to back the loan.

For example: Both Alice and Bob want to borrow WBTC on AAVE V3. While Alice uses ETH - a well-established asset, Bob uses AAVE token - a less proven asset, leads to different parameters like this:

The key is: Even though both Alice and Bob borrow the same amount of WBTC, Alice can borrow more and faces safer liquidation conditions because ETH is a more established and liquid collateral. Bob, using the AAVE token as collateral, is limited by a lower maximum LTV and faces a higher liquidation penalty due to the higher risk profile of his collateral.

But it is not enough to protect system

Although adjusting parameters such as maximum LTV and liquidation penalty based on the collateral asset helps limit downside risk, this mechanism alone is not sufficient to fully protect the system. These parameters only restrict how much a borrower can take and when their position is liquidated, but they do not price the ongoing risk that a borrowing position introduces to the shared liquidity pool.

Borrowers using riskier collateral are allowed to borrow less, but they still pay the same interest rate as safer borrowers as long as they borrow the same asset. As a result, higher-risk positions do not have significant opportunity cost on shared liquidity of a specific market on AAVE, while the cost of their risk is absorbed by the entire pool. This creates a structural mismatch between risk and pricing, leaving the protocol risky despite conservative collateral parameters.

This mismatch may not cause significant issues when assets are carefully curated and segmented into separate markets. However, it becomes a much more serious concern in the context of AAVE V4 development, where the protocol transitions to a hub-and-spoke architecture and concentrates liquidity into a unified liquidity layer.

In such a shared liquidity environment, riskier borrowing positions can no longer be allowed to access the same capital at the same price as safer ones without consequences. To address this, Aave V4 introduces an additional component in its interest rate calculation known as the “Risk Premium”, which dynamically adjusts borrowing costs based on the risk profile of each position’s collateral.

SECTION 2: RISK PREMIUM - OPPORTUNITY COST OF CHOOSING COLLATERAL ASSETS

In AAVE V4, borrowing costs are no longer determined solely by the supply and demand of the borrowed asset. Instead, the protocol extends its interest rate model to account for the quality of the collateral backing each position. This marks a fundamental shift from asset-level pricing to position-level risk pricing.

This risk premium is applied on top of the asset’s base borrow rate, ensuring that users who introduce more risk into the system pay more for accessing shared liquidity.

Under this design, a user’s debt interest is directly influenced by the risk profile of the assets used as collateral. Higher-quality, more liquid, and less volatile collateral results in a lower additional borrowing cost, while riskier collateral introduces an extra charge.

1/ Risk level

• Asset-level risk (Collateral Risk Point): When an asset is listed in a Spoke, governance assigns it a Collateral Risk score

• User-level risk (User risk premium): Reflecting how risky a user’s borrowing position is. If a user supplies multiple types of collateral, the protocol computes a value-weighted average of the Collateral Risk scores of the assets actually backing the user’s debt.

• Spoke-level risk: Each Spoke tracks the average User Risk Premium of its active positions, capturing the overall risk environment within that Spoke.

Because different spokes serve different use cases and user profiles, users’ risk levels can vary even though in AAVE V4, they share the same underlying liquidity hub.

Spoke-level risk is managed by 2 level inside: Asset-level and User-level.

a) Asset-level risk: Collateral risk points (CR_i)

Collateral Risk (CR) is the foundational parameter in AAVE V4’s new interest rate design and should be the first concept to understand when analyzing the protocol’s borrowing mechanics. Unlike previous versions, where collateral risk primarily affected liquidation thresholds and loan-to-value limits, AAVE V4 directly incorporates collateral quality into interest rate formation.

Collateral Risk point (CR) is a risk score assigned to each collateral asset, expressed in basis points (BPS), ranging from 0 to 100,000 (meaning 0% to 1000%).

The closer the value is to 0, the safer the asset is considered; the closer it is to 1000, the higher the risk associated with using that asset as collateral.

CRᵢ is the Collateral Risk point of the asset i.

For example: CR_ETH = 0 BPS, in contrast, a new asset - XYZ may have CR_XYZ = 100 BPS.

This means ETH is considered the safest form of collateral within a given Spoke. Using ETH as collateral therefore introduces no additional collateral risk to the position.

Compared to ETH, XYZ asset is more likely to experience sharp price movements or liquidity stress.

As a result, a position backed entirely by ETH carries nearly no risk premium, whereas using XYZ asset as collateral introduces an additional risk premium that the user must pay.

We will deeply examine how this risk premium is calculated and applied in detail in the following sections.

Importantly, as we have discussed earlier, CRᵢ is not a fixed or universal property of an asset. It is a configurable risk parameter governed at the Spoke level, and its value is set and adjusted by the Governor based on the risk profile and design goals of each Spoke. This means that the same asset can be assigned different Collateral Risk values across different Spokes, depending on how it is intended to be used.

b) User-level risk: User risk premium (RP_u)

In practice, AAVE V4 allows users to back their debt with a portfolio of different collateral assets, rather than relying on just one asset. The composition of this collateral portfolio can vary significantly from one user to another, both in terms of asset types and relative weights.

In addition, risk profile can also vary over time for the same user’s position. Even if the set of collateral assets remains unchanged, the risk profile of a position is inherently dynamic. Even with an unchanged collateral set, a position’s risk profile can evolve over time due to price movements, debt accrual, and updates to risk parameters.

Therefore, assessing risk cannot stop at the asset level alone. Two users may hold the same collateral assets, but in very different proportions, leading to materially different risk profiles. To capture this position-specific risk, AAVE V4 introduces the concept of User Risk Premium (RPᵤ).

User Risk Premium (RPᵤ) represents the aggregated quality of the collateral backing a user’s borrowing position. It is a dynamic value that depends on several evolving parameters:

• Collateral amount (Cᵤ,ᵢ) for asset i: The amount of asset (i) supplied as collateral by user (u). This value can grow over time as collateral may earn yield when being borrowed within the protocol.

• Asset price (Pᵢ) for asset i: Asset prices fluctuate continuously, and different assets exhibit different levels of volatility, directly affecting the risk profile of the position.

• Collateral Risk (CRᵢ) for asset i: the parameter of asset-level risk that we have discussed in the previous part.

Taken together, these factors allow AAVE V4 to compute a single, position-level risk metric that accurately reflects how risky a user’s borrowing position is and also reflects the changes in one position over time.

While RPᵤ defines what the risk level of a user’s position is, the next section explains how this value is computed in practice by identifying which collateral actually backs the debt and aggregating risk accordingly.

2/ Risk Premium algorithms

The Risk Premium algorithm connects the asset-level risk parameters defined earlier with position-level interest pricing in AAVE V4.

Collateral Risk (CRᵢ) provides the initial risk signal for each asset, but because a single position can be backed by multiple collateral assets, risk must be evaluated at the position level through User Risk Premium (RPᵤ).

At its core, this algorithms’ flow identifies the minimum set of collateral, prioritized from lowest to highest risk, that is strictly required to cover the user’s outstanding debt, and then computes a value-weighted average of their Collateral Risk to capture the position’s marginal liquidation risk.

Step 1 - Sort the collateral assets by risk level: starting with the safest assets (lowest CR) and ending with the riskiest ones (highest CR).

Consider a borrowing position backed by a portfolio of three collateral assets: ETH, WBTC, and wstETH.

Suppose the protocol assigns the following Collateral Risk (CR) scores within a given Spoke:

• CR_ETH = 0 BPS (lowest risk)

• CR_WBTC = 200 BPS

• CR_wstETH = 150 BPS

Based on these risk scores, the collateral assets are sorted from lowest to highest risk in the following order:

1. ETH (CR = 0 BPS)

2. wstETH (CR = 150 BPS)

3. WBTC (CR = 200 BPS)

This ordering defines the priority in which collateral is considered when assessing the risk of the position. Lower-risk assets (such as ETH) are assumed to back the debt first, while higher-risk assets (like wstETH and then WBTC) are only taken into account if the safer collateral is insufficient to cover the outstanding debt.

Step 2 - Calculate total debt: Calculate the user’s total debt value (interest included) in base currency (totalDebt).

Once the collateral assets are ordered by their risk level, the next step is to determine how much debt the position needs to be backed.

This is done by calculating the user’s total outstanding debt, including accrued interest, and expressing it in a common base currency. This value, referred to as totalDebt, represents the amount of collateral that must be covered by the ordered collateral assets.

Step 3 - Collateral selection for debt coverage

Starting from the safest collateral, the protocol walks through the ordered list of assets and gradually uses them to cover the user’s total debt.

For each collateral asset, the system checks whether the remaining debt can be fully covered by that asset alone. If it can, only the necessary portion of the asset is taken and the process stops. If it cannot, the entire value of the asset is used, and the algorithm moves on to the next, riskier collateral.

This process continues until the total debt is fully covered, ensuring that only the minimum amount of collateral required, starting from the lowest-risk assets, is considered.

For example: Suppose a user has the following collateral portfolio:

• $1,000 ETH (CR_ETH = 0 BPS)

• $800 wstETH (CR_wsETH = 150 BPS)

• $700 WBTC (CR_WBTC = 200 BPS)

• Total outstanding debt (including interest) = $1,600

1/ ETH can covers $1,000 of the debt at maximum.

Therefore, remaining debt = $600.

2/ wstETH has enough value to cover the remaining $600.

Therefore, only $600 worth of wstETH is used and the process stops, while $200 wstETH and $700 WBTC in users positions are still untouched.

Step 4 - Compute weighted average of Collateral Risk for the included collateral assets and its amounts.

Once the set of collateral assets (and portions of assets) that actually back the debt has been identified, the final step is to aggregate their risk into a single position-level metric.

This is done by computing a value-weighted average of the Collateral Risk scores, using the amounts of collateral that were included in the debt coverage.

The formula to calculate RPᵤ can be written as:

With:

• Cᵤ,ᵢ : Collateral amount for asset i

• Pᵢ : Asset price for asset i

• CRᵢ : Collateral Risk for asset i (BPS)

By weighting Collateral Risk by collateral value, the formula ensures that RPᵤ reflects the effective risk exposure of the position, not merely the presence of risky assets in the user’s portfolio. In other words, a risky asset only increases RPᵤ if it meaningfully participates in securing the debt.

For example (continuing): After performing the Debt Coverage Selection step, the effective set of collateral that actually backs the user’s debt is as follows:

Therefore, only ETH and wstETH are included in the risk calculation, applying the formula, we have:

So that, the Risk Premium in this case is 56.25 BPS.

Although the user supplied three different collateral assets, only ETH and a portion of wstETH are required to secure the debt. As a result, RPᵤ is calculated as a value-weighted average of the Collateral Risk scores of those assets alone.

The higher-risk WBTC does not affect the User Risk Premium at all, because it does not participate in backing the outstanding debt.

However, there are two practical questions remaining unanswered: when RPᵤ is updated, and how it is operationally applied to affect borrowing costs.

When is User Risk Premium (RPᵤ) updated?

In principle, User Risk Premium should continuously reflect the latest state of a user’s position, since it depends on collateral composition, asset prices, and governance-defined risk parameters. In practice, however, continuous updates are not feasible on EVM-based blockchain, as they would require constant on-chain computation and state changes.

Instead, in AAVE V4, RPᵤ is recalculated only when the user executes an action that can change the risk profile of the position. These actions include supplying or withdrawing collateral, borrowing or repaying debt, or explicitly triggering a position update.

If the user remains inactive, RPᵤ is left unchanged, even if prices move or risk parameters are updated in the background. Exceptionally, the Governor retains the ability to forcibly update the User Risk Premium of a given user to match the most recent risk parameters of its collateral assets, even in the absence of user interaction.

This design strikes a balance between risk accuracy and on-chain efficiency, ensuring that RPᵤ is always correct when it matters, without incurring unnecessary computational overhead.

How does User Risk Premium (RPᵤ) operate in practice?

Once RPᵤ is determined, it must be translated into actual borrowing costs without directly modifying the principal debt. AAVE V4 achieves this by introducing virtual debt shares, known as premium shares, together with a corresponding premium offset.

In AAVE V4, premium shares do not represent additional borrowed assets. When a user borrows, they receive exactly the amount shown in the interface, and this amount is recorded as base debt, which is the only real principal in the system.

The purpose of premium shares is purely accounting-based: they exist to measure and accrue the extra interest associated with collateral risk over time.

To ensure that no extra cost is charged upfront, the protocol creates a premium offset at the same time as the premium shares. Initially, the value of the premium shares and the offset are equal, meaning the user owes no risk premium at the moment of borrowing. As time passes, interest accrues on the premium shares while the offset remains fixed. The difference between the two represents the accumulated risk premium that the user gradually pays for using riskier collateral.

This design allows AAVE V4 to price collateral risk dynamically through interest, without increasing the borrowed principal or giving the impression that the user has taken on additional debt.

At any point in time, a user’s accrued premium debt is calculated as the current asset value of their premium shares, minus the premium offset.

When the user performs an action (such as borrowing, repaying, or adjusting collateral), the premium accumulated so far is locked in and moved to a “realized premium” balance. The system then resets both the premium shares and the premium offset, because the user’s Risk Premium may have changed and the premium accounting must be recalibrated to reflect the new risk profile.

The difference between the two represents the accumulated risk premium owed by the user.

Therefore, at its core, the Premium Offset serves as a baseline reference in AAVE V4’s risk-aware interest accounting. Its purpose is to ensure that risk premium accrues only over time, rather than being charged immediately at the moment a position is opened.

3/ Final Interest Rate calculating

As we have mentioned earlier, in AAVE V4, interest on a borrowing position does not come from a single source. Instead, it is split into two components:

• Base rate: based on utilization of borrow assets in Hub

• Premium rate: reflecting how risky the user’s collateral is.

These two components grow separately in the system, but from the user’s perspective, they are combined into one total debt balance.

a) Base rate

The base rate represents the cost of using liquidity from the Hub.

When a user borrows an asset, the borrowed amount is recorded as Base Debt. This is the real principal the protocol provides on the user’s behalf.

Overtime, Base Debt accrues interest according to the Hub’s base borrow rate. This rate depends on liquidity utilization, not on collateral quality.

b) Premium rate

The premium rate represents the extra interest charged because of collateral risk.

The premium rate is where the Risk Premium algorithm becomes economically meaningful. All of the previous mechanisms exist for a single purpose: to determine how much additional interest a user should pay on top of the base rate.

For example: Suppose the base borrow rate for GHO is 5% (500 BPS).

If you use WETH as collateral, which has a Collateral Risk of 0 BPS, no premium is applied. Your interest accrues purely at the base rate of 5%.

But if you instead use LINK as collateral, and based on the user’s collateral composition and the Risk Premium algorithm discussed earlier, the protocol determines that your position has a Risk Premium of 300 BPS (for example).

Your effective borrow rate is now therefore:

In other words, the user pays a higher interest rate not because more liquidity is borrowed, but because the collateral backing the position is assessed as riskier.

CONCLUSION

Risk Premiums make borrowing in AAVE V4 fairer and clearer. Users who provide safer collateral can borrow at lower interest rates, while riskier positions naturally pay more. This encourages better collateral quality and helps keep the system stable.

At the same time, the protocol can support many different assets and lending styles without forcing everyone into the same risk bucket. By pricing risk more accurately, AAVE V4 improves market efficiency and creates a healthier, more sustainable model for both users and the protocol.

Importantly, this accounting design ensures that users always borrow exactly the amount they intend to, while collateral risk is priced gradually through interest rather than upfront charges. By separating principal debt from risk-based accrual, AAVE V4 avoids hidden costs, improves transparency for borrowers, and aligns borrowing costs more closely with the true risk a position introduces to shared liquidity.

DISCLAIMER

In this article, we focus on analyzing the Risk Premium mechanism in AAVE V4, explaining how collateral risk is incorporated into borrowing costs at the position level. Rather than evaluating performance or making protocol comparisons, our goal is to break down the core design logic and on-chain mechanics behind AAVE V4’s risk-aware interest model.

Throughout the article, we do not avoid discussing technical concepts, smart contract flows, or mathematical formulations that may initially appear complex. However, all models, examples, and calculations are intentionally simplified and explained step by step, with the goal of making AAVE V4’s risk-aware interest model clear, intuitive, and accessible to readers with different levels of DeFi experience.

Ahead of the AAVE V4 launch, we will return with further deep-diving articles exploring other key components of the new architecture.

Leave a comment

INTRODUCTION

Last week, we took a quick look at some potential AMMs for fixed-yield trading in the part 1 of Pendle AMM, including Constant Geometric model, YieldSpace’s and Notional’s.

In this Part 2, we will deeply dive into the AMM model behind Pendle V2 to explore the pros and cons of its pricing mechanism, together with our detailed mathematical explanation and swapping transaction via Pendle’s AMM in reality.

KEY TAKEAWAYS

This article is a detailed explanation of how Pendle V2 works via the following main sections:

• Section 1: Understand the essence of Pendle V2 AMM’s pricing mechanism through the mathematical breakdown and Flash Swap mechanism

• Section 2: Examine the strengths and weaknesses of Pendle V2 AMM

• Section 3: Explore how Pendle V2 AMM works in practice through a swapping transaction on-chain

  ---

BEFORE WE BEGIN

Pendle V1 enabled trading PT and YT through two distinct liquidity pools, thereby causing capital fragmentation. To address this, Pendle’s new AMM was launched in V2, which only allows LPs to add PTs and the yield bearing token or SY tokens instead of YTs in V1.

These two tokens are closely correlated, with a high expectation of causing no impermanent loss for LPs. On top of this, PTs and YTs can be traded just inside a single pool of PT liquidity, via flash swaps mechanism (explained later), boosting capital efficiency.