What we detected
Between 12:26 and 14:24 UTC today, Glass tracked an unusually coordinated operation on Hyperliquid. 153 entity clusters, controlling 156 wallets, executed 34,012 trades totaling $5.4 billion. The buy-sell split: $2.61B in buys, $2.79B in sells. That leaves $180M in net short exposure built across two hours.
This is not a single whale selling. It is a coordinated, algorithmic operation using iceberg execution across dozens of wallets to build a massive short position without moving the market.
Why we call it an iceberg
An iceberg order splits a large position into many small, uniform trades. The goal is to hide the true size from other market participants. Three signals confirm the pattern.
Every cluster hits the exact same maximum trade size. Not $1.1M, not $1.2M — exactly $1,124,448. This is an algorithmic cap, not human execution.
Trades arrive in periodic bursts: 48–78 trades in a single minute, then 2–3 minutes of silence. Classic iceberg refill pattern. The algorithm waits for the order book to absorb before injecting the next batch.
The largest single trade is less than 1% of any entity's total volume. A $184M position built from 1,147 trades where no single trade exceeds $1.12M. That is algorithmic precision.
Why Hyperliquid?
Every trade is on Hyperliquid. Not a single trade on Binance, Bybit, or OKX. Hyperliquid is a decentralized perpetuals exchange where wallet addresses are visible on-chain. Ironically, the transparency that makes Glass tracking possible is the same reason these entities chose a single venue: cross-exchange execution would create detectable lead-lag patterns. By staying on one venue, they avoid our cross-venue TWAP sweep detector. But they cannot hide the on-chain wallet clustering.
The entities: who is selling
Glass clusters wallets into entities based on behavioral similarity, timing correlation, and trade pattern overlap. The operation involves at least 7 major clusters, each with a consistent net SHORT bias.
| Entity | SELL Vol | BUY Vol | Net SHORT | Trades |
|---|---|---|---|---|
| cl_0x19ac54 | $95.5M | $56.3M | -$39.3M | 1,890+ |
| cl_0x31dea2 | $111.3M | $81.3M | -$30.0M | 2,100+ |
| cl_0x39475d | $98.2M | $73.0M | -$25.2M | 1,900+ |
| cl_0x335f45 | $95.7M | $72.5M | -$23.2M | 1,800+ |
| cl_0xc6ac58 | $184.8M | $169.7M | -$15.1M | 2,162 |
| cl_0x34fb5e | $59.8M | $45.9M | -$13.9M | 1,100+ |
| cl_0x3bcae2 | $176.7M | $164.5M | -$12.2M | 2,122 |
| Top 7 Total | $822M | $663M | -$158.9M | 13,000+ |
Every single major cluster is net short. Not one exception. The smallest net short position among the top 7 is $12.2M. The largest is $39.3M. This is not a coincidence. It is a coordinated directional bet against BTC.
Inside a cluster: how they hide the size
Cluster cl_0xc6ac58 is the largest by raw volume ($354M in 2 hours). Inside it, Glass identified 15+ wallets. Each wallet trades both sides, creating the appearance of neutral market making. But the net position tells a different story.
| Wallet | Type | BUY | SELL | Net |
|---|---|---|---|---|
| 0x94d3...3814 | gambler | $79.5M | $41.2M | +$38.3M |
| 0xc6ac...e884 | liquidity | $29.1M | $41.5M | -$12.4M |
| 0xecb6...b00 | liquidity | $6.5M | $24.6M | -$18.1M |
| 0x3947...3b1 | liquidity | $11.3M | $20.5M | -$9.2M |
One wallet (0x94d3) is net long by $38M. The other three are net short by a combined $39.7M. The cluster looks balanced at a glance. It is not. The “long” wallet provides cover: it creates buy volume that masks the directional intent of the other wallets. This is textbook iceberg camouflage.
The alpha whale confirms the direction
One real alpha trader (0x621c...63ab, alpha score 0.38, dominant lifecycle) is active in this cluster. Position: $8.0M SHORT vs $5.2M LONG. Net: -$2.8M SHORT. When a behaviorally-verified dominant whale aligns with the iceberg direction, the signal gains conviction.
The execution timeline
The algorithm follows a pattern. Bursts of 48–78 trades in a single minute, injecting $6–12M into the order book. Then 2–3 minutes of silence while the book absorbs. Then another burst.
| Time (UTC) | Trades | BUY | SELL | Pattern |
|---|---|---|---|---|
| 14:13 | 78 | $6.2M | $6.2M | Burst |
| 14:12 | 6 | $0.2M | $0.2M | Quiet |
| 14:11 | 52 | $2.8M | $4.3M | Burst |
| 14:10 | 56 | $4.0M | $3.9M | Burst |
| 14:09 | 8 | $0.5M | $0.6M | Quiet |
| 14:08 | 11 | $0.5M | $0.2M | Quiet |
| 14:07 | 11 | $0.4M | $0.8M | Quiet |
The burst-silence-burst pattern is the signature of iceberg execution. A human trader does not stop and start with this regularity. An algorithm does. It fills a chunk, waits for the order book to refill, then injects again. The goal: absorb liquidity without tipping off other participants.
What this means for price
$180M in net short exposure does not exist in a vacuum. It represents a directional bet that BTC will move lower. The operators chose Hyperliquid because on-chain execution lets them bypass centralized exchange surveillance. But it also means the positions are visible to anyone who knows where to look.
If these clusters continue building short exposure in the next 4–8 hours, the $180M becomes a commitment. Once the position is large enough, the operator has incentive to push price into the liquidation zones below $62,625 to profit.
$180M in shorts is also $180M in potential forced buying. If price moves against them — above $65,500 — these positions begin to bleed. A squeeze would force rapid covering through the same iceberg mechanism, but in reverse.
Glass does not predict which scenario plays out. It identifies the participants, maps the positions, and monitors the flow. When $180M in short exposure moves, it will show up in the data before it shows up in price.
Real-Time Whale Tracking
Glass monitors 18,000+ whale wallets across Hyperliquid, Binance, Bybit, OKX, and 8 other venues. Cluster detection, iceberg identification, and alpha scoring update every 30 seconds.
Whale Intelligence Dashboard