Two perp DEXs, two architectures, one lopsided market. Hyperliquid runs an on-chain order book and clears roughly $6B a day. GMX runs a liquidity pool and clears roughly $400M. This comparison looks at why the gap exists and what each design is still good for.
Key Takeaways
- Hyperliquid holds over 70% of perp DEX market share with $5.9B in TVL as of June 10, 2026. GMX sits at $162M in TVL per DeFiLlama.
- Fees favor Hyperliquid for active traders: 0.015% maker and 0.045% taker base, against GMX's 0.05 to 0.07% plus borrow costs.
- The two are not actually competing for the same user. Hyperliquid is execution infrastructure. GMX is a passive LP yield product with a trading venue attached.
Background
GMX defined the first generation of perp DEXs: traders take positions against a pooled counterparty (GLP, now GM pools), pricing comes from oracles, and LPs earn the spread plus fees in exchange for taking the other side of trader PnL. It made perps on-chain viable.
Hyperliquid made them fast. A purpose-built L1 running a fully on-chain central limit order book, with matching and settlement at exchange speed. No pooled counterparty: makers and takers meet in the book, the way they do on any serious derivatives venue.
Analysis
Hyperliquid
The order book model wins on everything an active trader prices: tighter spreads, deeper liquidity at the touch, funding that tracks the market rather than pool utilization, and base fees below most centralized venues. Roughly $6B in daily volume and a 70%+ share of the perp DEX market is the result. The builder-fee model also makes it the settlement layer other platforms route to, which compounds the liquidity advantage.
GMX
GMX's pool model lost the active-trader war but kept a real niche: passive yield. GM pool LPs earn real fees from trader activity without running a strategy, and the oracle-pricing model means zero-slippage entry for mid-size positions in quiet markets. The structural cost is that LPs are short trader PnL, and the pool bleeds when traders win. At $162M TVL against a $5.9B rival, GMX is now a yield product more than an execution venue.
Data
| Metric | Hyperliquid | GMX |
|---|---|---|
| TVL | $5.9B | $162M |
| Daily volume (approx.) | $6B | $400M |
| Base fees (maker / taker) | 0.015% / 0.045% | 0.05 to 0.07% + borrow costs |
| Model | On-chain order book | Oracle-priced liquidity pool |
| Counterparty | Other traders | LP pool |
(Source: DeFiLlama, protocol documentation, June 10, 2026)
Implications
For execution, the market has voted and the result is not close. For an allocator, though, the more useful frame is that these venues expose different legs of composable positions: perp funding on one side, passive LP yield on the other, and basis between them. The trades that matter, equity basis, funding capture, cross-venue spreads, need both legs visible at once. No single venue shows you that, and that gap is where the next generation of cross-asset tooling is being built.
Risk Considerations: Perpetual futures carry liquidation risk and funding cost drift. GMX LP positions are short trader PnL by construction. Geofencing applies at the venue level. Check eligibility before sizing.
Data sources: DeFiLlama, protocol documentation. Analysis as of June 10, 2026. Research, not advice.