Two years ago, if you asked which decentralized perpetual exchange was winning, the answer was GMX. Through 2023, GMX dominated decentralized perpetual trading on Arbitrum with peak TVL above $700 million and daily trading volume in $1-2B range. The GLP liquidity provider model was elegant — single liquidity pool earning trading fees from perpetual traders without the impermanent loss complexity of standard AMMs. GMX token had real value capture through fee distribution. The protocol felt like the answer to "how do you do decentralized perpetuals at scale."
Hyperliquid wasn't really on most people's radar in early 2023. Sub-million daily volume. Sovereign L1 in early development. Hyperliquid mainnet had just launched. Anyone evaluating decentralized perpetual sector for investment thesis would have rationally focused on GMX with maybe dYdX as alternative. The Hyperliquid bet looked uncertain.
Through Q1 2026, the picture inverted completely. Hyperliquid does approximately $4.2 billion in daily perpetual volume — orders of magnitude larger than GMX V2's current ~$50-150M daily. GMX V2 (the upgraded version launched in 2023 attempting to address V1 limitations) holds about $150M TVL versus the V1 peak of $700M+. GMX token sits at market cap meaningfully below 2023 peaks while Hyperliquid HYPE became one of the top performing tokens of the cycle.
This piece is about the dethroning. Specifically: what happened, what GMX did right, what it did wrong, and what's left of the protocol that genuinely still serves specific use cases.
Going back to GMX's strengths in the original V1: GLP (GMX Liquidity Provider) was a genuinely innovative model. Liquidity providers deposited basket of assets (BTC, ETH, USDC, others). Perpetual traders opened positions against the basket. LPs earned trading fees plus benefited from trader losses (since aggregate trader PnL is roughly zero-sum minus fees). GLP yields ran 15-30% APY consistently in good periods.
The model worked because it solved a real problem. Standard AMMs don't work well for perpetual liquidity provision — impermanent loss makes LP economics adverse. GLP avoided this through basket structure and fee accumulation. LPs were willing to provide liquidity at scale because GLP returns were genuine.
Where GMX V1 had structural limitations: limited market coverage (initially 4-6 markets, expanding gradually), order book limitations (oracle-priced rather than true order book), specific UX friction for active traders, GMX as governance token had limited direct value capture mechanism.
Hyperliquid attacked these limitations directly. Sovereign L1 architecture supporting true on-chain order book with sub-second settlement. Comprehensive market coverage at launch (50+ perpetuals, expanding to 180+). HYPE token with aggressive buyback program creating genuine value capture mechanism. Better UX for active traders. Lower funding rate spreads versus CEX equivalents.
Each Hyperliquid advantage compounded. Active traders chose Hyperliquid for better execution. Market makers concentrated liquidity there for better economics. Volume concentration created network effects. By 2024, Hyperliquid had clearly displaced GMX as decentralized perpetual leader.
GMX team responded with V2 launched in 2023. V2 changes included: separate liquidity pools per market (replacing single GLP basket), better trader UX, expanded market coverage, isolated market risk model. Technically competent upgrades. Strategically reasonable response.
The realized adoption of V2 has been disappointing. V2 TVL grew gradually but never returned to V1 peak. Daily trading volume on V2 hasn't matched V1 peak either, much less competed with Hyperliquid scale. The structural problem: V2 fixes what V1 lacked, but Hyperliquid fixed it better through fundamentally different architecture (sovereign L1) that GMX can't match while remaining on Arbitrum.
GMX team has remained on Arbitrum strategically. Building on Arbitrum means inheriting Arbitrum ecosystem benefits (fee structure, ecosystem integrations, established user base). It also means accepting Arbitrum performance constraints versus Hyperliquid's purpose-built L1. The strategic choice is reasonable but bounded.
Q1 2026 GMX V2 specific metrics:
Total TVL: ~$150M Daily perpetual volume: ~$50-150M Active markets: ~30+ across BTC, ETH, SOL, AVAX, and various altcoin perpetuals GLP-equivalent (GM) yield: 8-15% APY for liquidity providers GMX token market cap: ~$280-450M depending on day
Compare to peak V1: $700M TVL, $1-2B daily volume, GLP yields 15-30%.
The contraction is real. The protocol still operates at meaningful scale but at substantially smaller scale than peak. Daily volume comparable to mid-tier centralized exchange perpetual products rather than competing with major venues.
Where GMX V2 still serves users:
Arbitrum-native traders who don't want to bridge for perpetual trading. Operating within Arbitrum ecosystem cohesion has real value for some users.
GLP-equivalent yield seekers. GM token holders capture meaningful APY (8-15%) from perpetual trading fees. Higher yield than most stablecoin yield sources, with different risk profile.
Specific market coverage that Hyperliquid doesn't list. GMX V2 supports some markets that Hyperliquid doesn't, providing access for specific trading needs.
GMX token positioning for users with conviction in protocol recovery or yield generation continuing.
Where GMX V2 doesn't serve users well anymore:
Active high-frequency perpetual traders who care about execution quality. Hyperliquid wins decisively on this dimension.
Users wanting comprehensive perpetual market coverage. Hyperliquid's 180+ markets exceed GMX V2 substantially.
Users wanting decentralized perpetual sector token exposure. HYPE captures more upside per dollar of allocation than GMX given Hyperliquid's stronger value capture mechanism.
The realized GMX trajectory through 2024-2026 reflects a competitive market where being good isn't enough when the competition is structurally better. GMX V1 was good. GMX V2 is good. Hyperliquid is structurally better through architecture choices that GMX can't replicate without changing fundamental approach.
For my own positioning: I had GMX position during 2022-2023 GMX V1 peak. Captured meaningful upside on both GMX token appreciation and GLP yield positioning. Exited GMX positions through 2023-2024 as Hyperliquid trajectory became clear and rotated to HYPE positioning. Currently zero GMX exposure.
For users still positioned in GMX or considering position: the realistic situation is that GMX represents stable but bounded protocol with established yield generation mechanism for liquidity providers. Token positioning depends on whether you believe protocol recovers competitive position (unlikely given structural Hyperliquid advantages) or maintains current scale as specialized infrastructure.
For users wanting perpetual trading on Arbitrum specifically: GMX V2 works fine. Execution is reasonable. Costs are competitive within Arbitrum context. The trade-off is accepting somewhat worse execution versus going to Hyperliquid sovereign L1.
For users wanting decentralized perpetual yield generation: GLP-equivalent positioning on GMX V2 captures real yield. The yield isn't going anywhere structural — perpetual trading on GMX continues generating fees that flow to liquidity providers. Sized appropriately, GM token positioning provides reasonable risk-adjusted yield.
Forward speculation through end-2026: GMX likely continues operating at current scale (~$100-200M TVL, ~$50-150M daily volume) without recovery to peak nor catastrophic compression. The protocol is stable infrastructure that lost dominant position but maintains specialized functionality. GMX token economics support continued operation but not meaningful price appreciation absent specific catalyst.
The broader lesson from GMX trajectory: in competitive sectors, protocols that establish dominant position based on first-mover advantage get displaced when competitors emerge with structurally better approaches. The displacement isn't always quick — GMX V1 maintained substantial volume through 2023 even as Hyperliquid scaled. But sustained competition with structurally better alternative typically leads to displacement.
For users tracking decentralized infrastructure broadly, GMX is a useful case study. Strong execution. Real innovation. Good team. Outcompeted by competitor with different fundamental approach. The lesson isn't that GMX failed — it's that even good execution may not be enough when competitor has structural architectural advantages.
Sourcing notes: GMX TVL, volume, and token figures from GMX dashboards, DefiLlama, Arbitrum ecosystem analytics through April 2026. Daily volume fluctuates ±25% across the quarter. GLP yield calculations approximate. The competitive comparison with Hyperliquid uses publicly available metrics. Personal positioning observations reflect my own historical and current allocation. GMX token economics depend on real-time market dynamics and protocol performance. Decentralized perpetual sector continues evolving with competitive dynamics that may shift further. None of this is financial advice.