Verify the proof, ignore the hype.
Hyperliquid’s protocol revenue just dropped 43% year-over-year. $357 million in Q3 2025 down to $202 million in Q2 2026. The HIP-3 fee split is the culprit. I’ve seen this pattern before. In 2020, I stress-tested MakerDAO’s liquidation cascades. The math was unforgiving then. It’s unforgiving now.
Context: The HIP-3 Mechanism
Hyperliquid is a Layer 1 built for decentralized perpetuals. HIP-3 is a permissionless market deployment system. Anyone can stake 500,000 HYPE—roughly $28 million—and launch a perpetual market. The builder keeps 50% of the trading fees. The protocol gets the other 50%. Simple in theory. Brutal in practice.
Kain Warwick, founder of Synthetix and Infinex, called it unsustainable. He’s been through this. At Synthetix, external builders max out at 30% fee split. His argument: Hyperliquid can unilaterally slash the builder’s cut at any time. The builder has no recourse. The platform holds all the power. The data backs him up.
Core: The Revenue Chain Collapse
Total trading fees on Hyperliquid remain robust. Volumes haven’t crashed. But the split has shifted. RWA perpetuals—tokenized stocks and commodities—now account for 50% of all volume, up from 2% in early 2026. That’s a $3.6 billion open interest, surpassing Bitcoin perpetuals. The growth is real. The problem is where the fees go.
Here’s the chain: - Total fees: Assume $400 million in Q2 2026 (rough estimate based on prior data). - 50% goes to builders: $200 million. - Protocol retains: $200 million. But the reported revenue is $202 million. Close. - 99% of protocol revenue goes to the Assistance Fund for buybacks: $200 million. - Actual buybacks reported: $149 million. The rest goes to other fund uses. - Buybacks down from $290 million peak. That’s a 48% drop. - HYPE price: $57.66, down from $76.67 high. That’s a 24.8% decline.
The buyback engine is stalling. The token’s deflationary narrative is losing steam. Traders are still trading. But the fee flow has shifted from the protocol to the builders. The protocol is paying for growth it can’t capture.
Concentration risk compounds the issue. One builder, trade.xyz, controls over 90% of HIP-3 open interest. That’s a single point of failure. If trade.xyz leaves or gets hacked, the protocol loses half its business. The platform has no diversification. It’s a hostage situation dressed as a partnership.
Contrarian: The Hidden Stability
Some argue the 50% split is a feature, not a bug. It attracts deep-pocketed builders who can create liquid markets. The $28 million HYPE stake acts as a bond. The builder won’t walk away easily. Even if the split is cut to 30%, the sunk cost keeps them locked in. That’s the theory.
But the data says otherwise. Revenue is down. Buybacks are halved. The token is bleeding. The builder might stay, but the token holder suffers. The platform’s unilateral control is a double-edged sword. It can cut fees, but it can’t force the builder to stay if the economics sour. trade.xyz has no counterparty. It’s a monopoly. And monopolies exploit their position.
Warwick’s point is simple: the current split is a temporary market-share grab. Once Hyperliquid owns the dominant position, it will renegotiate. The builder has no leverage. The protocol has all the cards. But the market is already pricing in this risk. HYPE is down 25% from its peak. The question is how much further it falls.
Takeaway: The Vulnerability Forecast
Code is law, but bugs are reality. HIP-3 is a smart contract mechanism. It’s technically sound. But the economic design is fragile. The platform owns the settlement layer. The builder owns the liquidity. Both need each other, but the power dynamic is asymmetric. The protocol can change the rules. The builder can only leave.
I’ve been through this. In 2022, I reverse-engineered Arbitrum’s fraud proofs. The lesson: trust the protocol guarantees, not the fee split. When the revenue drops, the protocol will optimize for itself. HYPE holders will see buybacks shrink further. The only fix is a governance change to reduce the split. That will trigger a builder exodus. The short-term pain is inevitable.
HYPE’s current price may not fully reflect the buyback decay. If the split stays at 50%, revenue continues to slide. If it drops, builder concentration becomes a crisis. Either way, the token is in a structural decline. The market hasn’t priced in the full magnitude of the revenue chain. The next quarterly report will confirm the trend.
Verify the proof, ignore the hype. The math is clear. Hyperliquid’s HIP-3 is a brilliant mechanism. But it’s a mechanism that’s already broken. The only question is when the protocol pulls the trigger.