GMX (GMX)

MegaETH Terminal is shutting down, but we're ending the season with a banger.

+29,000 points and a top #30 rank.

Overall, I lost about $5k USD, with the biggest losses coming from Hitone (-$1k USD), Monster (-$1k USD) and GMX (-$2.5k USD due to liquidation).

So now I'm hoping for at least $0.25 per point, so it wasn't all a complete waste of time.

Overall, there are actually some good apps on there.

I still think shutting down the Terminal is a mistake tho

I've seen many people possibly misunderstand the non‑orderbook perp DEX Jupiter/GMX.

They think that centralized contracts without KYC like Hyperliquid, which correspond to "AMM exchanges," are generational replacements.

But that's completely wrong.

1. First, the execution logic of Jupiter/GMX is not an "AMM curve" (that belongs to the older @perpprotocol).

Instead, an oracle pulls prices from a price source, and the LP pool acts as the counterparty for trades.

In theory, this type of exchange can align with the price source without slippage. For example, CME gold futures; oracle‑based perp DEXs can directly match its quotes.

Whether it's an order book like Hyperliquid or a CFD model like @variational_io, they all need to "arbitrage" prices over, essentially acting as intermediaries moving price differences.

However, GMX adds a "price impact" to guard against timing differences caused by price propagation and block times, preventing arbitrage.

2. What truly makes GMX/Jupiter impressive is their ability to create assets.

Many people don't realize that, in the realm of contract trading, so‑called "real‑world assets" essentially don't exist.

Your opening and liquidation prices have no relation to the liquidity of the underlying assets in the real world; they are provided by the contract's pricing mechanism (order book/oracle, etc.).

So if you like to criticize "lack of underlying value support" as gambling, you can now fire at order book exchanges.

But GMX/Jupiter are different; their liquidity comes from LP pools. LPs, acting as counterparties for all trades, collect all client losses and trading fees and redistribute them proportionally – essentially the house edge of a Macau casino.

This is a truly yield‑generating asset.

Moreover, these pools can be used as counterpart‑risk‑free structured finance, collateral for DeFi nesting, and arbitrage trades.

Jupiter's JLP used to be the largest source of arbitrage trades and profits on Solana. When @DriftProtocol had issues, many didn't understand its purpose—it was then one of the biggest providers of JLP neutral arbitrage strategies.

Such capability is unavailable to order book exchanges. Centralized exchange vaults, including HLP, are essentially active market‑making strategies—like handing your funds to a centralized black box. Using HLP as collateral essentially transfers counterparty risk.

3. Irreplaceable business logic based on the binary nature of "betting against the pool".

Uniswap has already shown that the LP pool mechanism provides an irreplaceable efficiency advantage for issuing long‑tail assets—no order book to date can replace the same mechanism as Uniswap/pump/fourmeme.

This is not an efficiency issue; it's a market‑making cost issue, and I'm not just talking about cold starts—how can an order book exchange recover all counterparty P&L and trading frictions?

If you didn't understand, you can go back and check my tweets on polymarket and propAMM.

The real issue with GMX and Jupiter may be that they consider only the traders' perspective, not the upstream market demand—a fundamental difference from Hyperliquid.

Nevertheless, you can check the DefiLlama rankings; regardless of how top‑ranked order books game the rankings, GMX/Jupiter's volume and revenue are barely affected.

If they were more aggressive, their performance would improve, but this already proves they are irreplaceable.

Fixed. https://t.co/DkaDH9A6QQ