GMX: Introducing Global Hedge Vault (GHV) for Eliminating Impermanent Loss

Proposals
1

Background

In the current GM/GLV structure, LPs earn trading fees and borrowing fees but also bear directional PnL risk from traders. This behaves similar to impermanent loss:

  • When traders are profitable overall, LP returns decrease or even turn negative.
  • When traders lose overall, LP returns increase.

The great success of the Single-Sided BTC Pool ($66m), Single-Sided ETH Pool ($50m) confirms the existence of LP segments with different risk appetites. By separating GM/GLV fees income from PnL risk, we can better match different LP preferences, attract more liquidity, and significantly improve competitiveness - especially as we are about to expand to ETH mainnet.

Objective

Introduce the Global Hedge Vault (GHV) as a separate, investable hedge vault to absorb PnL volatility from GM/GLV:

  1. Provide GM/GLV LPs with risk-neutral returns by removing PnL exposure.
  2. Allow higher risk preference LPs to earn PnL-driven returns via GHV.
  3. Enable structured risk products without altering any GM/GLV architecture.

Mechanism

  1. The GHV account is protocol-whitelisted, trading with 0 fees to minimize hedge costs.
  2. Each time a trader opens a position, GHV opens an equal-sized opposite position to maintain OI balance (eg. BTC Long = $50m, BTC Short = $48m → OI Balance = $2m (long bias) GHV opens $2m short → BTC Long = $50m, BTC Short = ($48m + $2m)).

Impact

  1. GM/GLV LPs’ directional risk is entirely shifted to GHV. GHV LPs bear the PnL volatility in exchange for a risk premium.
  2. As for market arbitrage interaction:
    a. If professional arbitrageurs exist, they only need to understand GHV’s mechanism and identify GHV’s position, then they can still open a $2m short for arbitrage, with the short increasing from $50m to $52m and then quickly returning to $50m.
    b. If arbitrageurs are absent, GHV holds the exposure and absorbs related PnL.

Return & Risk Allocation

  1. GM/GLV LPs: Earn risk-neutral returns (trading fees + borrowing fees) with no need to self-hedge.
  2. GHV LPs: Earn PnL-driven returns, historically positive and stable. GHV significantly lowers risk for GM/GLV LPs, attracts stronger liquidity, creates a new product class for risk-seeking investors, and provides a sustainable risk-transfer and yield-splitting mechanism to support multi-chain growth.

image
image2902Ă—1300 345 KB

  1. In GMX V2, the largest cumulative trader loss was $16m, while the largest cumulative trader profit was $7m. Since the deployment of the adaptive funding fee mechanism, OI balance has been well maintained, resulting in a more stable PnL curve.

image
image2516Ă—1810 322 KB

Structured Product Potential

With GHV, GM/GLV can be decomposed into (referencing Pendle’s terminology for easier understanding):

  1. PT-GM/GLV: It is only exposed to backed assets and automatically accrues fees from leverage trading and swaps for the corresponding market.
  2. YT-GM/GLV: Holds only PnL risk exposure, potentially positive or negative.

image
image2380Ă—610 83.7 KB

In this case, PT-GM/GLV is the current GM/GLV token, and YT-GM/GLV is the GHV token.

PT-GM:BTC-BTC APY 4.86% / YT-GM:BTC-BTC APY 3.84%
PT-GM:ETH-ETH APY 7.39% / YT-GM:ETH-ETH APY 5.08%

Implementation Cost & Compatibility

  1. GHV acts like an external trader/MM and does not require GM/GLV core changes which means almost zero implementation cost.
  2. Full compatibility of working with existing architecture; deployable to ETH mainnet or other new chain seamlessly.

Risk Assessment

  1. In extreme cases, GHV’s NAV could go to zero, but GM/GLV would continue operating as before.
  2. It carries concentration risk. Large single-market exposure in GHV needs diversification controls.

Extension Possibility

With LP risk significantly reduced, a small adjustment in fee allocation could strengthen GHV: eg. Reduce LP fee share from 63% → 50%, redirect 13% to the GMX Treasury for GHV replenishment.

All discussion are welcomed.

Q

7 Likes
2

This is what DeFi should focus on: structuring, building blocks, and composability.

GHV, in a nutshell, acts as the counterparty to the market’s long and short positions, maintaining a consistent long-short balance.

This ensures that GM/GLV LPs are not exposed to market volatility.

GHV can then pursue a split strategy on Pendle.

With multi-chain deployment (including Ethereum L1), GHV, and Pendle, GMX could catch up to Pendle’s market capitalization if it gets the right direction.

6 Likes
3

I’m always a proponent of risk localization! Very cool proposal, thanks Q!

The only thing that concerns me now in this story is the inevitable complication of our Menu :slight_smile: which could lead to a repelling effect for unsophisticated users. But this shouldn’t be a stop factor! Just at some point I suspect we’ll need to start discussing introducing a separate UI mode toggle: Main mode \ PRO mode :slight_smile:

giphy

2 Likes
4

Great point, Q! From my perspective, the most critical factor here is pricing the risk of trader PnL. At the very least, it should account for:

  1. The additional trading costs incurred from hedging orders;
  2. The fact that GHV also needs to hedge on behalf of the risk-neutral pool’s funds, meaning the fee distribution ratio should dynamically adjust based on the capital ratio between the two pools.
1 Like
5

Good idea, GHV is very similar to Hype’s HLP.

One question, GMX is using price impact to rebalance the long/short ratio, should we rebuild it to adjust GHV to avoid excessive risk exposure?

1 Like
6

Mechanism still unclear for me. So I need to deposit my GM/GLV to GHV vault and choose one of two risk profiles- neutral or PnL driven? Will I still have the option of doing nothing with my GM/GLP and get current risk profile?

7

You still have the option of doing nothing for being the LP of GM/GLV. GHV is an independent option similar to GLV. If you want to be the LP of GHV you should deposit USDC rather than GM/GLV cuz GHV should be backed by USDC.

1 Like
8

So I will deposit USDC to GHV (and get GHV tokens back?) and then what? How to choose if I want exposure only to fees or only to PnL?

9

Any data points on trader pnl for pools such as ETH/USDC and BTC/USDC over a year period without accounting for Trading fee’s, i believe most chart is factoring fees too thus this is hard to guage if it be a solution to split? wish to see if the % would make it a viable option for LP…

1 Like
10

Correct, if you want to get exposure only to PnL you can deposit USDC to get GHV while if you want to get exposure to fees you can deposit USDC or other backed assets to get GLV.

GHV → PnL
GLV → Fees

Of course you can custom your a GHV : b GLV cuz there is more flexibility.

11

Have forwarded to Chaoslabs for further estimation, I believe they will provide the data points soon.

1 Like
12

Thanks Q for this response, i really appreciate the effort and thought in clarifying the qns. Do wish to have more statistics to determine the viability of this proposal.

13

First, let me commend you — you’ve contributed many meaningful proposals.

As a long-term GM LP for more than 3 years, I’d like to share some thoughts from my perspective. My main concern is not whether GM is fully risk-neutral, but rather its long-term return profile and the security, especially when v2 has reduced significantly my exposure to the trade’s pnl.

In the long run, I know I will earn some profit from traders’ losses, particularly in the BTC and ETH markets. If trading profit were stripped out, GM’s yield would inevitably decline, which in turn would reduce the overall GM size. Since GM size directly determines the protocol’s maximum trading capacity, at this stage, any change that might diminish it should be considered with caution now.

I believe the most critical bottlenecks today lie in the high slippage and funding rates faced by traders when opening large positions. If, after the introduction of the net OI mechanism, the same GM size can support two or even several times the current position capacity, then the GHV proposal would become highly meaningful — as it would instantly keep net OI close to zero over the long term, amplifying this multiplier effect.

My suggestion, therefore, is to prioritize launching and iterating on the net OI mechanism until it reaches sufficient maturity before introducing the GHV mechanism.

Additional suggestion, if the net OI mechanism proves effective, we could consider reducing — or even waiving — trading fees for traders who open short positions using BTC/ETH as collateral solely to capture funding fees. This would help net OI converge to zero more quickly, and it would be a simpler approach than introducing GHV.

10 Likes
14

Thanks for sharing your thoughts, Hugo. They echo what I was thinking about as well: a substantial share of GM LPs are concerned more with total returns than with being fully risk-neutral. (We saw this with GLP as well; people will stomach risk and sizable trader PnL-based drawdowns if the APR is ultimately very healthy).

Removing trader PnL will lower GM performance, and that may well reduce the size of the GM pools. This should be carefully considered.

That said, there are significant advantages to a GHV-type hedging instrument as well. I appreciate the thought that’s gone into this.

2 Likes
15

Hi Hugo,

This is a brand new idea and quite constructive. I fully agree your understanding of Net OI and the potential impact of it. I will continue to research it and also see how Chaoslabs evaluates it.

Thank you!

3 Likes
16

Hi, Q,

Glad you replied my post. This is Hugo, the anti-spam robot seems locked my previous account, so I have this new account to reply.

Following your GHV proposal, I’ve got some extra thoughts beyond the GM LP yield angle I mentioned in my last post.

Basically, GHV acts like a real-time counterparty: whenever someone opens a position, GHV steps in and keeps OI balanced. For funding fee farmers, this kind of kills their edge—if they try to jump in when OI is already balanced, they break the equilibrium and end up paying extra costs, so most of the time they just sit out. In other words, GHV eats their opportunity, and over time, it might become the main counterparty risk taker.

The problem is, if GHV isn’t big enough, it caps overall trading volume and turns into a bottleneck. That’s why you’d need some sort of growth mechanism—without it, GHV probably shouldn’t even launch.

Another issue: since GHV is always in the market, its price swings a lot. Anyone participating basically has to act like a trader, constantly watching the charts and making calls, which is super costly. And in a one-sided market, they’ll just pull out, which again limits how much OI can be opened.

Just to be clear, I’m not really against GHV. What you’re doing—moving risk into a separate bucket so it can be hedged—that’s a classic move in financial product design. And honestly, for GMX it’s a pretty inspiring thought experiment. It made me realize something: in GMX’s LP system, counterparty risk never easily fully disappears—it always ends up sitting somewhere. If you want a closed system to scale, you need to bring in external energy to reduce entropy.

That’s why I think v2.0 was genius: it offloaded a lot of GLP’s counterparty risk and passed part of it to funding fee farmers. Even though they sometimes lose money in negative rate periods, they make enough overall to stay active, which helps keep OI balanced.

And about Hyperliquid—early on, GMX community thought they were wash-trading, but in reality their insane TPS just makes it super easy to plug into traditional market maker liquidity. Their HLP is more of a side liquidity provider. That’s why they can push $20–30B daily volume—like 100x GMX.

So if GMX could design things to make external liquidity easier to plug in, and combine that with Net OI mechanism, I think volumes could easily grow 10x.

I don’t know the exact answer, I’m not a finance pro. But maybe:

  • make funding fee farming cheaper so farmers earn more,

  • or set up a dedicated vault so they can farm funding fee passively without watching the market 24/7.

Anyway, just throwing ideas out there—hope it’s useful.

3 Likes
17

hi newhugo! thanks for your contribution
i have remove the “silenced” from your older account.
u may want to reuse back ur old account (or stick to this new one, up to u)

thanks again and apologies for the inconvenience caused by the bot.

2 Likes
18

10x volume is music to our ears.
i think tradfi funds is a huge boost and a hard nut to crack.

1 Like
19

Personally I will vote against, exposure to trader wins/losses is a significant reason I LP here vs. anywhere else. Take the ETH/ETH pool, you want to cut it from 12.5% to 5%?.

13% of 63% is 20%, that’s not a “small fee reduction”.

I completely understand why some investors would want this, but there isn’t a single positive point about the proposal for me personally.

20

Much better proposal IMO

1 Like