400,000 wallets trading Hyperliquid perps through retail apps. $280M lost all-time. Fees ate more than half of it. Not because users traded badly. Because every app stacks its own builder fee on top of Hyperliquid’s base rate. Invo users deposited $12.4M and traded $5.8B. 467 turns on capital. At that frequency, fees compress any edge to zero. Monthly active users tripled. But only one in seven new traders is still around three months later. Growth is running on new wallet supply, not retention. I’ve seen this structure many times in this industry. Distribution scales faster than user experience. The builder code model is perfect for the exchange — new users, volume, fees, all incoming. The cost sits with the user. Hyperliquid solved something real: it outsourced user acquisition while offloading the downside. That is genuinely smart infrastructure design. But a growth model dependent on a continuous supply of new capital coming in has a ceiling. The question is whether any of these apps figures out retention before they hit it. #Hyperliquid #DeFi #DigitalAssets #M55Capital