No need to wait until Monday's open; Hyperliquid's HIP-3 is becoming a pricing reference for CME.

Deep comparison of Hyperliquid and traditional futures giant CME in micro-level performance during extreme market conditions and weekend closures.

Author: shaunda devens, Blockworks Research

Translation: Deep Tide TechFlow

Deep Tide Guide: As HIP-3 proposal advances, Hyperliquid is accelerating its expansion from the cryptocurrency space into traditional finance (TradFi) assets. Recent sharp volatility in the silver market provides an excellent stress test for this decentralized derivatives protocol.

This article uses detailed trade, quote, and order book data to deeply compare Hyperliquid and the traditional futures giant CME (CME Group) in terms of micro-level performance during extreme market conditions and weekend closures.

The study finds that, although still unable to match the depth of traditional giants, Hyperliquid demonstrates unique competitive advantages in retail-level order execution and “24/7 around-the-clock price discovery,” even becoming an important reference for pricing at Sunday open.

Full Text:

HIP-3 is driving Hyperliquid beyond the crypto space, with traditional financial (TradFi) tools now accounting for 31% of the platform’s trading volume, and an average daily nominal trading volume exceeding $5 billion. Silver is the most significant part of this capital flow, and last Friday’s intense volatility provided a stress test for HIP-3’s market health.

Using high-frequency trade, quote, and order book data, benchmarked against CME/COMEX Micro Silver futures, we find that for smaller, retail-heavy orders, Hyperliquid Silver offered narrower spreads and better execution before the crash. We also showcase a new 24/7 use case: positioning and pricing for Sunday reopening auctions.

Key Findings:

• Before the crash: Under the main trading volume of perpetual contracts, Hyperliquid’s top-of-book prices are highly competitive. Its median spread is 2.4 basis points (bps), compared to 3 bps for CME; median slippage is only 0.5 bps relative to the benchmark. Retail-sized flows dominate (median trade size $1,200), with robust but relatively limited depth: Hyperliquid’s depth within ±5 bps is about $230,000, while CME’s is approximately $13 million.
• During the crash: Both platforms’ performance declined, but Hyperliquid experienced more significant execution tail risk. Its spread widened by 2.1 times, CME’s by 1.6 times. Market deviations exceeded 400 basis points at times, later reverting via funding rate mean reversion. The main issue was execution quality: 1% of Hyperliquid trades deviated from mid-price by over 50 bps, while CME had none.
• Weekend: Hyperliquid was the only platform trading during CME’s closure, functioning like an exchange-based order-driven silver derivatives platform. Over 49 hours, it processed 175,000 trades with a nominal volume of $257 million, and median spread compressed to 0.93 bps. However, HIP-3’s weekend trading structure is relatively thin, with volume only 31% of weekday levels.
• We believe that pricing for reopening and phased risk hedging before the single-price open are core use cases for 24/7 trading. But based on current Hyperliquid stock-like asset performance, pre-open price prediction does not significantly outperform Friday’s closing price.

HyperLiquid: HIP-3 captures trading volume

Last week, the silver market faced structural liquidity events. As retail, futures, and regional spot market liquidity demands surged simultaneously, silver prices re-evaluated sharply. From peak to trough under high volume, silver fell about 17%; US retail investors injected approximately $170 million into silver ETFs in a single day, reportedly the largest single-day inflow on record, nearly double the peak of the 2021 “Silver Squeeze.” Meanwhile, CME activity surged to multi-year highs, and the Shanghai Gold Exchange prices showed double-digit USD premiums over London benchmarks.

For the crypto industry, a more important development is that these funds did not stay confined to traditional financial platforms. As volatility increased and traditional commodity markets approached the weekend, incremental demand for metal exposure migrated to 24/7 derivatives trading platforms, where position adjustments and risk transfers could continue without trading hour restrictions.

On Hyperliquid, silver perpetual contracts settled hundreds of millions of dollars in nominal volume throughout the week. Under HIP-3, stock and commodity perpetual contracts also hit new highs, with daily volume expanding from $378 million over 66 days to $4.8 billion. By Friday, about 31% of total platform volume was linked to TradFi instruments. Silver became one of the most active contracts on the platform, with a substantial shift: five of the top ten contracts by Friday volume were non-crypto assets.

We have always viewed HIP-3 as a scalable Delta-one (first-order sensitivity) wrapper. Its returns are linear, contracts have no expiration, and holding costs are reflected through funding rates and basis, rather than option-like time decay.

From an investment perspective, trading platforms that span beyond crypto can add differentiated, cycle-insensitive revenue streams. This is significant because Hyperliquid’s protocol revenue ranks high in volatility, with weekly revenue volatility around 40%. Additionally, implied analysis suggests that capturing even a small portion of TradFi derivatives flow could double income, offering a feasible path for stepwise growth.

However, whether these markets can scale depends on three implementation constraints: continuous and resilient oracle design, sufficient order book depth to maintain price integrity, and reliable hedging paths when underlying reference markets are intermittent. In this framework, the silver event is Hyperliquid’s first meaningful stress test of linking to TradFi perpetual contracts, benchmarked against CME.

This report evaluates performance across three phases (pre-deviation, sell-off, weekend), measuring price integrity, liquidity resilience, and risk engine behavior when external markets are impaired or closed. We compare Hyperliquid’s pricing, basis, and liquidity metrics during overlapping trading periods with CME, then analyze the transition from “weekend to reopening” to quantify rebound behavior.

Our ultimate goal is to answer: Is Hyperliquid’s HIP-3 product suitable as a trading venue for perpetual stock/commodity exposure? Has Hyperliquid created a high-performance 24/7 stock and commodity market?

Data

We use tick-by-tick trade, quote, and order book data from Hyperliquid’s silver perpetual contract (XYZ100), comparing it with CME’s near-month silver futures (SILH6).

For Hyperliquid, we rely on the TradeXYZ deployment, as it consistently carries the highest HIP-3 trading volume.

We compare Hyperliquid with CME Micro Silver (SILH6), as its unit size better matches the retail-to-midscale order distribution on perpetual contracts. On crash day, SILH6 traded 641,926 lots (about 642 million ounces; at $78–$120/oz, nominal value roughly $50–$77 billion), while macro contracts (SIH6) had deeper depth but performed modestly on spreads and slippage before the crash. Since this report focuses on execution quality at typical perpetual contract sizes, SILH6 is the most relevant CME benchmark.

The dataset covers January 30 to February 1, including 540,000 Hyperliquid trades and 1.3 million depth snapshots, benchmarked against 510,000 CME trades and full 10-level order book data on crash day. We analyze in three phases: pre-crash (Friday 12:00–17:00 UTC), sell-off (17:00–22:00 UTC), and weekend (from Friday close to Sunday reopen).

Market before the crash

Starting from the baseline before the crash, when CME and Hyperliquid traded normally with external reference intact.

On the surface, Hyperliquid’s silver perpetual market appears quite mature: quotes are consistently tight, activity high. The top-of-book bid-ask spread averages 2.7 bps (median 2.4 bps), with 90% of observations at or below 5 bps.

The natural benchmark is CME near-month silver futures (SILH6), which is the most liquid tradable reference during the overlap. It’s important to note that CME’s depth is structurally deeper, remaining an institutional-grade liquidity venue. Our goal is not just comparison but testing whether Hyperliquid can reliably track the underlying benchmark and provide dependable price integrity and execution for mainstream order sizes.

Between 12:00–17:00 UTC, CME’s nominal trading volume was about $85.5 billion, while Hyperliquid’s was $679 million. Despite the scale gap, median spreads are very close: CME averages about 3.1 bps, Hyperliquid slightly narrower.

Nevertheless, CME’s spread distribution is tighter in the tails, with 96% of observations within 5 bps, compared to 90% for Hyperliquid, consistent with the deeper, more stable passive liquidity typical of major futures platforms. Hyperliquid’s narrower quotes should be understood in context: its capital flows are easier to manage and skewed toward retail (average trade ~$5,000, median $1,190), which mechanistically reduces “toxicity” at the order book top.

While bid-ask execution is comparable, depth is not. CME supports about $1.98 million within ±2 bps, Hyperliquid about $30,000; within ±3 bps, CME supports ~$5.45 million, Hyperliquid ~$83,000; within ±5 bps, CME ~$13 million, Hyperliquid ~$231,000. For retail orders crossing spreads, narrower quotes are a real advantage. For trades over $50,000 nominal, depth differences determine final execution costs.

Even so, for a platform without designated market makers, Hyperliquid’s displayed depth is not insignificant. Order book is roughly symmetric, with buy/sell depth ratios near 1, expanding from about $231,000 within ±5 bps to $814,000 within ±10 bps, and roughly $1.5 million within ±25 bps.

However, compared to traditional futures venues, Hyperliquid’s depth shows weaker “solid execution” performance. Matching occurs on on-chain CLOBs with block-level order sorting, where cancellations are processed before same-block limit orders. Therefore, execution priority depends partly on transaction type rather than just arrival time, weakening the correlation between “visible depth equals guaranteed volume” as in CME’s continuous matching engine.

Execution quality provides additional info beyond spread and depth. Using best bid-ask at trade time, median slippage from mid-price is 1.5 bps on CME, 2 bps on Hyperliquid. CME’s execution is very tight, with 99% within 5 bps of mid. Hyperliquid’s distribution is broader: 83% within 5 bps, 96% within 10 bps, with occasional trades deviating over 20 bps, consistent with intermittent order book gaps and thinner capacity.

Slippage slightly increases with trade size (CME about 1.5 bps at 1 lot, 1.6 bps at 2–5 lots); on Hyperliquid, the slope is steeper, rising from about 1.9 bps below $1,000 to 2.8 bps above $50,000. Notably, the execution gap between platforms is much narrower than the quote depth difference suggests. For median Hyperliquid trades (~$1,200), execution cost differs from CME median by only about 0.5 bps, despite CME’s much larger typical trade size.

Finally, interpreting execution also depends on oracle and mark price design, since traders may transact in a deep but non-representative mark order book. Under HIP-3, the oracle is published by the deployer as a non-trading reference with fixed cadence and clamps; the mark price for managing funding, margin, and liquidations is a robust aggregation of oracle and local order book signals, also constrained to prevent sharp moves. This separation allows the trade price to maintain a persistent premium or discount without mechanically forcing immediate liquidation, anchoring risk management to a slowly moving reference price, while still enabling continuous order book-driven price discovery.

In the pre-crash window, Hyperliquid maintained about 29 bps of persistent premium over CME. This premium decomposes into: oracle vs CME component (~18 bps, reflecting differences between oracle’s basket of underlying assets and near-month futures) and perpetual vs oracle premium (~9 bps, reflecting net long demand and funding pressure in perpetual contracts). The premium is very stable, rarely inverted.

Comparing execution with the oracle shows a median premium of about +9 bps.

In summary, the pre-crash benchmark indicates that, given its scale and participant base, the platform provides clear settlement for retail and medium-sized flows. Despite shallower depth across all tiers, its smaller capital flows still deliver highly competitive spreads, with median trade differences from institutional benchmarks within 0.5 bps.

Depth gaps do exist and have material economic implications for large trades and extreme scenarios. But considering typical market sizes, Hyperliquid was already operating at surprisingly high market quality levels before decoupling began.

Market during the crash

Around 17:00 UTC on Friday, January 30, reports emerged that Trump intended to nominate former Fed governor Kevin Warsh to succeed Powell as Chair. Widely viewed as hawkish on monetary policy, silver prices re-evaluated sharply, experiencing the largest single-day decline since March 1980. Silver dropped about 31% from Thursday’s near $120/oz high to around $78 intraday. Futures, ETFs, and perpetual contracts with leverage faced margin pressures, and forced liquidations became a key part of the market.

For perpetual platforms, this feedback loop can be self-reinforcing. As the reference price declines, market makers short perpetual contracts, and losing positions are forcibly liquidated into the order book. If liquidity withdrawal outpaces liquidation settlement, transaction prices can jump multiple tiers, widening the basis and increasing tail slippage.

Both platforms’ spreads worsened, with Hyperliquid’s tail response being more pronounced. Its median spread expanded from 2.4 bps pre-crash to 5.1 bps during the crash (2.1x). The 95th percentile (P95) rose from 6.0 to 18.2 bps, with only 49.5% of observations within 5 bps (vs. 90.5% pre-crash).

In the worst 5-minute window around 18:20 UTC, median spread hit 17 bps. CME also widened, from 3.0 to 4.8 bps median (P95 12.7 bps), maintaining a tighter overall distribution. Its worst 5-minute window reached 10.1 bps at 18:20.

Similarly, depth shrank reflecting liquidity withdrawal. On Hyperliquid, within ±5 bps, depth fell from about $231,000 pre-crash to $65,000 during the peak, with median at zero—mainly because spreads had already widened beyond ±5 bps.

At wider tiers, liquidity persisted even under high stress: within ±25 bps, about $542,000; within ±50 bps, roughly $1.07 million. CME also showed similar mechanical patterns at narrow tiers (±2 and ±3 bps often zero during peaks), but absolute capacity remained an order of magnitude higher. Under peak stress, CME still held about $1.16 million within ±5 bps, Hyperliquid near zero.

Execution quality at median levels declined on both platforms, but tail behavior diverged. Hyperliquid’s median deviation from mid-price increased from 2.0 to 4.1 bps (~2x), CME from 1.5 to 2.7 bps (~1.8x). CME maintained tight execution, while Hyperliquid exhibited heavy tails: about 1,900 trades (~1% of crash volume, nominal $21 million) deviated over 50 bps from mid, while CME had none.

Due to lower liquidity and forced liquidations, Hyperliquid’s mark price ultimately deviated from the oracle. The HL-CME basis peaked at 463 bps at 18:30 UTC, but the excess over 400 bps lasted only 95 seconds, falling back below 50 bps within 19 minutes. The spread followed a similar trajectory.

Overall, Hyperliquid’s spreads widened more, and its distribution showed heavy tails, consistent with the thinner order book during forced liquidations. But these deviations did not persist. During unprecedented volatility, Hyperliquid maintained continuous tradability, with prices anchored near benchmarks, with major damage concentrated in tail executions during rapid moves.

Market close

At 22:00 UTC Friday, CME closed, halting the traditional institutional silver reference cycle. Hyperliquid remained open. For HIP-3 perpetual contracts, this is a special phase where external oracle updates are unavailable, and the platform shifts from “externally anchored” to “constrained internal guidance” reference.

Deployers continue publishing indices, but use impact prices derived from order book derivatives, filtered through slow EMA (exponential moving average). The management of margin and liquidation mark prices is a robust blend of the index, short-term basis filters, and local order book signals, constrained by maximum leverage limits (about 5% for silver).

The weekend mechanism, in theory, enables price discovery outside trading hours. When external oracles resume on Monday, the internal price is pulled back to the external reference, but this intermediate window allows traders to position based on Friday’s levels before the open auction.

Throughout the weekend, trading continuity remains high: 175,000 trades, $257 million in volume. Participant composition skews significantly toward retail compared to normal periods. Median trade size drops to $196 (from $1,245), with the 99th percentile at $18,100.

In this dimension, quote liquidity tightens notably. Weekend bid-ask median spread is 0.93 bps, compared to 2.40 bps on normal days. Depth declines but remains stable and balanced. The median bid-ask depth within ±10 bps is $358,000. Execution follows similar patterns.

Using mid-price at trade time, weekend slippage median is 0.87 bps, versus 1.98 bps during normal hours. In other words, crossing spreads during weekend flows incurs lower costs than on weekdays, despite weaker absolute capacity.

Price behavior shows significant volatility over the weekend, not static. Silver prices fell from $85.76 to $83.70 before reopening, providing a real-time reference for continuous trading.

The Globex reopening provides the clearest cross-platform check: at 23:00:00 UTC, CME’s first trade price was about 97 bps above Hyperliquid’s mid-price. By 23:00:01, the gap compressed to about 10 bps. Hyperliquid’s continuous weekend market produced a price level very similar to CME’s opening auction. Notably, the internal price on Sunday was closer to Monday’s open than Friday’s close.

Among all HIP-3 markets, despite continuous trading, weekends are structurally low-participation periods. Analyzing 32 xyz markets with 5-minute K-line charts (defining weekends as Friday 16:00 to Monday 09:30 ET), we find: when weighted by weekday nominal value, the notional value per 5-minute candle drops to 0.31x of weekday levels (a 69% decline); when equally weighted, to 0.33x (a 67% decline).

Volatility also shrinks, but less than volume. When weighted by notional, 5-minute realized volatility drops to 0.75x of weekday levels (a 25% decline), with median at 64%. A small subset shows limited contraction, or even higher volatility during weekends, mainly due to differences in underlying trading hours and the dynamic of Sunday reopenings, which are still counted as “weekend” in definition.

The silver market fully conforms: xyz:SILVER’s 5-minute notional value drops 72%, while realized volatility declines only 21%. Narrower spreads and stable median execution coexist with lower overall participation and depth reduction away from the order book. In other words, weekend trading is optimized for continuity and small order execution, not institutional capacity. Despite volume drops, Hyperliquid still provides tight execution for the dominant small flows during this period.

Outlook for 24/7 round-the-clock trading

Given this structure, one of the most practical use cases for Hyperliquid’s 24/7 perpetual contracts is pricing the CME Sunday reopening auction. CME’s Sunday open is a single-price call auction: orders accumulate during pre-open, a reference opening price is published, and a brief non-cancellable window locks the order book before continuous matching resumes. The opening price aims to maximize tradable volume, then minimize residual imbalance, with tie-breaking based on previous settlement or other reference points. This method is effective for clearing backlog but concentrates information, hedging needs, and stop-loss flows at a single discrete price point.

A continuous trading venue like Hyperliquid changes the execution challenge because it allows participants to express and transfer risk before the auction compresses spreads. Traders need not passively accept the final auction price but can build positions in real-time during the weekend at the order book prices. In effect, Hyperliquid offers a tradable weekend reference price and a path-dependent execution plan for size, timing, and price limits, which are unavailable during CME’s order book closure.

As shown in the oracle handoff diagram, traders can trade internally over the weekend before the Sunday reopening. When external reference prices resume, the anchor is pulled back to the oracle, creating an economic incentive for traders to price the gap. The subtlety is that this advantage depends on weekend liquidity and nominal capacity, but when these constraints are met, Hyperliquid demonstrates execution benefits.

Across all xyz stock HIP-3 markets, we tested whether internal period pricing provides incremental price discovery before the Sunday oracle reopening. The method compares the volatility of the oracle’s opening relative to Friday’s close, with the observed mid-price volatility 15 minutes before reopening (all relative to Friday’s close).

In the current sample (23 markets; 191 weekend samples, 146 with valid pre-open mid-price snapshots), evidence of incremental weekend price discovery at the oracle level is weak. In 50.7% of observations, the pre-open mid-price is closer to the oracle’s opening than to Friday’s close, with a median increase of about +0.4 bps, which is negligible.

In other words, for these markets, the oracle’s opening price remains largely anchored to Friday’s close, and the deviation of internal mid-prices during the internal period does not persist when the external oracle resumes. This suggests that, at least for current HIP-3 configurations and liquidity mechanisms, weekend trading has not yet produced a robust, highly informative tradable benchmark. Nonetheless, as liquidity and depth build, we expect the internal period to become a more reliable pre-reopening price reference.

Conclusion

Hyperliquid’s HIP-3 silver perpetual contract successfully navigated unprecedented volatility without halts, providing tight order book pricing for retail and medium-sized flows. Market quality declined as expected under stress, especially in tail executions, but this imbalance was temporary; the basis quickly reverted to mean, and prices remained anchored near institutional benchmarks.

The limitation of HIP-3 lies in capacity; the platform handles small to medium orders well, but large orders face substantive constraints relative to CME’s depth.

Beyond the standard advantages of perpetual contracts, the weekend mode is the product’s most strategic value. Hyperliquid offers a continuous price path during traditional market closures, transforming what would be a discrete reopening gap into a tradable reference, creating an advantage in pre-open repositioning and opening price discovery.