Hook
Kain Warwick, the founder of Synthetix and Infinex, just dropped a sobering bombshell on Hyperliquid’s HIP-3 mechanism: the 50% fee split for external market builders is not sustainable. As someone who has spent years building decentralized derivatives protocols, Warwick’s words carry weight. But the data tells an even more unsettling story. Over the past 12 months, Hyperliquid’s protocol revenue has dropped from $357 million to $202 million — a 43% decline. Token buybacks, once the engine of HYPE’s deflationary narrative, have nearly halved from $290 million to $149 million. Yet, total trading volume remains strong. The money is just flowing to different hands. This is the paradox that defines Hyperliquid’s current state: a platform that is more active than ever, yet less rewarding for its token holders.
Context
Hyperliquid is a Layer 1 blockchain designed for on-chain perpetual futures trading. In early 2026, it introduced HIP-3, a mechanism that allows anyone to deploy an “unlicensed” perpetual market by staking 500,000 HYPE (approximately $28 million). External builders — like trade.xyz — keep 50% of all trading fees generated by their markets. The remaining 50% goes to Hyperliquid’s protocol, which funnels 99% of it into buybacks of HYPE tokens. This created a powerful incentive for professional market makers to build on Hyperliquid, and it worked: RWA (real-world asset) perpetuals grew from 2% to 50% of total platform volume in just a few months. But Warwick, who has navigated similar fee splits at Synthetix (where external builders get at most 30%), argues that 50% is an unsustainable equilibrium. He warns that Hyperliquid can “unilaterally cut builder fees or absorb their markets” at any time, leaving builders with no guarantees.
Core
Let’s break down the numbers. Hyperliquid’s total protocol revenue is simply the sum of all trading fees multiplied by 50% (the platform’s share). The other 50% goes to builders. In Q3 2025, protocol revenue peaked at $357 million. By Q2 2026, it had fallen to $202 million — a 43% drop. But total trading volume barely declined, if at all. The disconnect is clear: more fees are being captured by external builders, not the protocol. This is the direct consequence of HIP-3’s generous split. The buyback mechanism, which uses 99% of protocol revenue, followed suit: from $290 million down to $149 million. HYPE’s price, which hit $76.67 in early 2026, now trades around $57.66 — a 25% decline. The deflationary narrative is losing steam.
But here’s where it gets tricky. The 50% split attracted a single dominant builder: trade.xyz, which now accounts for over 90% of all HIP-3 open interest. That’s an enormous concentration risk. If trade.xyz decides to leave or scale back, Hyperliquid could lose nearly half its volume overnight. The platform’s control over the fee split — as Warwick pointed out — is a double-edged sword. It can lower the split to 30% tomorrow, which would boost protocol revenue by 40% and restore buyback power. But it would also risk alienating the very builders that made HIP-3 a success. This is a classic prisoner’s dilemma: both sides benefit from cooperation, but the platform holds the ultimate power.
From a technical perspective, HIP-3 is a marvel of permissionless innovation. But the economic asymmetry — low barrier to deploy, high barrier to earn — creates a fragile ecosystem. The 500,000 HYPE stake acts as a sunk cost, locking builders into the platform even if fee splits are cut. Yet, as I learned during my 2020 DeFi audit of OpenYield, trust is earned in drops, but lost in buckets. Once a platform changes the rules, the damage to builder confidence is hard to reverse.
Contrarian
Most critics focus on the fee split being too high. But what if the 50% split is actually a strategic feature, not a bug? Consider this: Hyperliquid’s real competition is not Synthetix or dYdX — it’s centralized exchanges like Binance and Bybit. Those platforms charge 0% to 0.02% maker fees, and they often share rebates with market makers. To attract institutional-grade liquidity, Hyperliquid had to offer something comparable. The 50% split is effectively a market maker rebate paid in protocol fees. It’s a cold-start mechanism that worked brilliantly: RWA open interest hit $3.6 billion, surpassing Bitcoin perpetuals. The problem is that this mechanism is unsustainable in the long run — not because it’s too high, but because it creates a single-point-of-failure with trade.xyz.
The contrarian angle is that the real risk is not the fee split percentage, but the lack of builder diversity. If Hyperliquid can attract 5-10 builders with 10-20% of OI each, it could lower the split to 35% without losing market share. The platform’s unilateral power is actually a tool to engineer that diversity. As I wrote in my 2022 bear market solidarity series, “Hold through the noise, build through the silence.” The noise is the fee split debate; the building is the diversification of builders.
Takeaway
Hyperliquid stands at a crossroads. It has the technology, the volume, and the market share to dominate decentralized derivatives. But the current fee structure is a ticking time bomb for token holders. The path forward requires a delicate balance: reduce the split to restore buyback health, but do it gradually to retain builder trust. Education is the antidote to exploitation — and in this case, the community must understand that governance decisions are not just about price, but about the sustainability of the entire ecosystem. The future belongs to those who teach together, especially when the market is sideways and the noise is loud. Code is law, but humans are the protocol. We built trust in the chaos, not despite it. Now, we must build sustainability in the silence.
