Hyperliquid once built a moat by relying on deep liquidity in long-tail assets, but the median USD depth has shrunk by 84.6% within 5 bp in the past year. Trader matteo compared the data of Binance and OKX, dismantling the rebate advantages of large market makers and the L1 congestion quota mechanism, and how to squeeze the living space of small market makers step by step. (Preliminary briefing: AI makes code worthless? Multicoin dismantles the three moats that make DeFi truly profitable) (Background supplement: GitBook was hit by an SSO vulnerability all the way through the production environment! Hyperliquid perpetual contract platform TradeXYZ announced its move) Trader matteo posted on the 84.6%, compared with 61% of Binance’s similar assets and 78% of OKX’s, the liquidity loss is significantly greater. He compared 86 samples of long-tail altcoins and found that even if the factor of currency price decline is excluded, the number of pending orders at the core of the market has been significantly reduced, and small market makers are facing structural elimination. gossip (read) priority fees on hyperliquid are broken currently the team that is benefitting the most is wintermute (insert fist shaking red emoji guy meme) this is not publicly known but can be determined with certainty by mapping out transfers to labeled wallets from the top… pic.twitter.com/K20Eclsl5x — Loris (@0xLoris) October 23, 2026 incredible things are happening in the hype discord pic.twitter.com/AykXNdKsFi — Ludwig (@0xvanbeethoven) November 4, 2026 Hyperliquid’s long-tail asset liquidity has deteriorated sharply over the past year. Across comparable markets, median USD depth within 10 basis points fell 82%, versus 61% on Binance and 78% on OKX. For the same tokens, Hyperliquid’s median 5-basis-point depth relative to… — Coincamps (@coincamps) November 11, 2026 Deep shrinkage of 84.6%, Hyperliquid’s loss relative to Binance intensified matteo Screened 86 assets that were listed on Binance, OKX and Hyperliquid at the same time between November 1, 2025 and November 1, 2026, and provided leverage below 10 times as long-tail samples. Among the 38 markets comparable in the 10 bp range, the median USD depth of Hyperliquid fell by 82%, while Binance and OKX were 61% and 78% respectively. The difference in decline is quite significant. To strip away the impact of the broader market, he divided each asset’s depth in Hyperliquid by its corresponding depth in Binance and compared the ratio before and after a year ago. At 5 bp depth, the median ratio drops by 57.5%, showing that Hyperliquid is losing a lot of liquidity relative to Binance even when comparing the exact same token. Pending orders at the midpoint of the market dropped sharply, but the top 20 orders increased against the trend. In response to the possible shrinkage of the US dollar caused by the drop in currency prices, matteo simultaneously tracked the US dollar depth of the market quotation and the number of tokens on pending orders. Taking November 2025 as the base period of 100, one year later the US dollar index dropped to about 13 in the 5 bp range, and the token quantity index fell to 28, which means that after excluding the currency price factor, there is still a large loss of pending orders at the core price. However, if the observation scope is expanded to the first 20 trading positions, the US dollar index stops at 61, and the token quantity index rises to 127. This means that the order book is not without volume, but that traders really want to close the deal close to the market position, the scale of pending orders has been greatly reduced, and the liquidity distribution has obviously shifted outward. The spread suddenly diverged after catching up with CEX, and the end of October became a turning point. In the past year, Hyperliquid's bid-ask spread originally continued to catch up with Binance and OKX, but it suddenly diverged recently. Since October 30, as the liquidity close to the transaction price has become more concentrated, the price spread has widened significantly, which is highly consistent with the market structure centralization problem that market makers have recently complained about. Comparing October 8-14 with October 31-November 6, the median depth share occupied by the largest account group in each market within 25 bp rose from 34.5% to 38.6%, and the median quote spread also expanded from 2.13 bp to 3.55 bp. However, matteo reminded that this period was mixed with the Binance security incident on October 29, and the fluctuations will not be directly attributed to priority fees or network congestion based on the curve alone. Market-making competition has not weakened, but there have indeed been fewer participants. matteo looks back at the underlying account structure and classifies the main account and the explicitly associated sub-accounts into the same account family to avoid double counting. Comparing the data of the same batch of 80 markets for a full week a year ago and a recent full week, the median share of pending order volume of the largest account family dropped from 25.2% to 14.6%, ruling out the idea that a single giant dominates the entire market. But the number of active market makers did decline. In the same batch of 72 markets, the median number of account families with orders of at least $1,000 each on both sides within 25 bp dropped from 7.64 to 6.39; even if converted at a fixed token price, there are still 58 markets where the number of active market-making accounts declined. The fee structure is congested with L1, and small market makers are double-attacked. The fee system has structural disadvantages. Statistics on the historical order handling fees of 54 account families show that the lowest quartile of trading volume has to pay a handling fee of 0.961 bp, while the highest quartile of groups can receive a commission of 0.236 bp. Not counting hedging costs and unfavorable transaction losses, small market makers have a disadvantage of about 1.20 bp at the starting line. In the overall price difference of only a few basis points... Hyperliquid shrank deeply by 84.6%: Why did small market makers retreat en masse? 〉This article was first published in DongZu BlockTempo "DongZu DongTen - the most influential blockchain news media".