Hyperliquid order-book depth matches Binance, Bybit and OKX median on four of five assets, Castle Labs finds

Castle Labs' order-book study hands Hyperliquid a result that matters for institutional trading: on four of five crypto assets, the onchain perpetuals venue held as much resting liquidity as the median of Binance, Bybit and OKX, and cost less to trade, according to The Defiant. Matching the median puts Hyperliquid level with the middle of that range, not ahead of it, and on the four assets where it matched, the evidence still leaves the deepest of the three books at least as deep. The quantity measured is depth — resting liquidity — which is what a large order has to be sized against; the coverage does not say which asset missed, how wide the shortfall ran, or over what period the books were sampled.

The onchain case thins where the study still finds a difference: pricing stocks and commodities while their underlying markets are closed. Venues quote those instruments when the cash markets behind them are shut, so the residual gap sits where there is no live reference price to match, and the cost of being wrong about where an asset reopens falls on whoever is quoting. Depth on crypto pairs has not been shown to carry across to those books.

Venue competition for institutional business has lately been settled on collateral terms: four venues on one bank's platform turned segregated off-exchange collateral from a marketing line into a cost of doing institutional business in September. Execution quality is the other half of that pitch and the harder half to evidence, because a book measured at rest is not a book measured under a large order. Professionalization has come mainly through wrappers, listing standards and platform approvals, which govern what can be held and where; an order-book comparison answers a different question, whether the venue can absorb size. The three competitors are meanwhile buying reach elsewhere: Binance's $100 million stake lifted USDC quote markets to 329 pairs on one exchange, as this publication reported.

The study as described carries no volume figures and names no user. Adoption would look like a fund, a pension or an RIA platform routing perps through an onchain venue; the coverage reports no such mandate.

Depth at the median of three exchanges answers an objection; it does not yet record a migration, and it rests on a single study's snapshots. The narrow thing to watch is whether the equity and commodity books, the one place the study still finds a difference, converge the way the crypto books did.