The Hyperliquid matching engine is drawing attention after OKX founder Star Xu described its design as more favourable to market makers than the sequencing used by conventional centralized exchanges.
In reported remarks from the New Era Finance podcast, Xu linked cancellation priority to the platform’s apparently stronger order books. He also suggested that the design offers another way to support assets with limited liquidity.
The podcast listing dates the episode’s publication to October 5, 2026, and says it was recorded on August 25. The circulating commentary therefore concerns an existing design—not a newly announced matching-engine upgrade.
Hyperliquid’s documentation supports the central sequencing point. It also provides a more precise explanation than the shorthand that the platform simply matches orders “block by block.”
How the Hyperliquid matching engine orders actions
HyperCore maintains an order book for each asset. Its documentation says orders match in price-time priority: price determines precedence, followed by time for orders at the same price.
Separately, actions are grouped and sequenced within consensus batches. There are usually one or two such batches in a block.
Processing order | Action category |
First | Actions that do not submit GTC or IOC orders |
Second | Cancellations |
Third | Actions submitting at least one GTC or IOC order |
Within each category, actions follow the order proposed by the block proposer. Modifications are categorized according to the replacement order they submit.
This distinction matters: action sequencing and order-book matching priority are different rules. Cancellation priority does not mean a maker can reverse a completed trade or always cancel before any incoming order.
Readers can find background on trading terminology in the Bitnxt crypto glossary.
Why market makers value cancellation priority
Market makers post buying and selling quotes. When market conditions change, an old quote can become unattractive before the maker manages to update it.
Hyperliquid explicitly describes its execution logic as protecting makers against “toxic flow.” Its latency documentation says cancellations and add-liquidity-only orders submitted at the same time will almost always execute before IOC and GTC orders. The stated objective is tighter spreads and deeper liquidity, including during volatility.
The economic logic is understandable: reducing the risk of being filled at a stale price may encourage a liquidity provider to quote more competitively.
That is a plausible benefit. It does not establish that every trader receives a better outcome on every market.
Order types change the execution experience
Hyperliquid’s API documentation distinguishes three instructions:
Instruction | Behaviour |
ALO — add liquidity only | Cancels rather than immediately matching |
IOC — immediate or cancel | Executes available quantity and cancels the unfilled remainder |
GTC — good until cancelled | Can remain on the order book until filled or cancelled |
Consider a hypothetical buyer who sees a sell quote and submits an IOC order. If that quote is cancelled before the buying instruction executes, the buyer may receive a smaller fill—or none within the specified price limit.
The maker avoids an unwanted execution. The buyer encounters a different result from the earlier book snapshot. Both follow from the same sequencing design.
For users researching automated tools, Bitnxt’s crypto trading bot directory provides product context. Actual venue support and order behaviour still require checking with each provider.
A deeper-looking book needs an execution test
Xu’s explanation should be evaluated through trading results.
Displayed depth captures orders visible at a moment in time. Execution quality concerns what happens when someone attempts to trade against them.
Useful comparisons would examine the same assets, order sizes and time windows across venues, including:
Average fill price after fees.
Filled quantity and unfilled orders.
Changes between the observed quote and execution.
Performance during volatility.
No controlled cross-exchange benchmark reviewed for this report establishes that Hyperliquid consistently outperforms centralized venues on those measures.
The distinction also matters as more asset classes reach blockchain trading systems. Bitnxt’s tokenized-stock market analysis explores why trading, settlement and ownership structures need separate examination.
You might also like: Hyperliquid’s Singapore Base Puts Regulatory Oversight Back in Focus.
Bitnxt view: judge the trade, not just the book
Our assessment is that maker protection can improve the conditions for supplying liquidity. Whether that benefit reaches takers depends on actual fills, costs and reliability.
Xu’s comments identify a meaningful design choice. The next useful evidence is a repeatable comparison of execution under equivalent conditions.
For traders, the decisive question is practical: when an order is submitted, how much executes, at what price, and at what total cost?
