Hyperliquid whales: what our position tracking shows

Hyperliquid is unusual among derivatives venues: every position of every account is public. Anyone can read a wallet's open positions, their size, leverage and liquidation price. We track a large set of these wallets and keep their positions current, which turns a public but scattered dataset into one picture. This post shows what that picture looked like at our snapshot (the biggest positions, the leverage behind them and how close they were to liquidation) and how to use the same data for a wallet of your own.

What we track

At the snapshot we held the open positions of 9,298 wallets: 41,763 positions across 295 markets, worth $12.6B in total.

The book was close to balanced. Long positions added up to $6.1B and short positions to $6.5B, so longs made up 48% of the tracked size. 1,719 positions were larger than one million dollars, and 192 were larger than ten million.

A position here is one wallet's position in one market. A wallet with a long in BTC and a short in ETH counts as two positions, and the size is the notional value at the mark price, not the margin behind it.

The largest positions

These were the ten largest tracked positions at the snapshot:

A few things stand out. The two largest positions belonged to the same wallet, 0x5b5d…c060: a $249M short in ETH and a $211M short in BTC, both at 5x. Their liquidation prices were far away: the price would have had to rise +57.5% and +68.8% respectively. Large size does not have to mean large risk: these positions carried a lot of margin.

Further down the list the picture changes. The fifth-largest position, a $114M ETH long at 10x, sat −7.4% from its liquidation price. The seventh, a $105M ETH long, was −7.1% away. Positions like these are the ones that appear as bright bands on a liquidation heatmap: one wallet alone can make up a visible cluster.

Concentration was moderate overall. The ten largest positions made up 11.4% of all tracked size: large individual wallets, but not a book dominated by a handful of them.

How much leverage

Leverage is where tracked positions differ most from each other:

The median position used 5x leverage, but the median dollar sat in positions with 10x. In other words, larger positions tended to use more leverage than small ones. The tails show the same thing: positions above 20x were 6% of all positions but held 16% of the size. Positions at up to 2x were 7% of the count and only 2% of the size.

Margin mode matters as much as the number. 20% of positions used isolated margin, where the collateral is limited to that one position; the rest used cross margin, where all cross positions of the account share one pool of collateral.¹ In cross margin the leverage setting of a single position says little on its own: a profit on one position supports the others, and a loss on one drags them all closer to liquidation.

How to read a whale position

Three numbers describe a large position better than its size alone:

• Size is the notional value at the mark price. It says how much the position gains or loses per percent of price movement, not how much money is at stake.
• Leverage is size divided by the margin behind it. Two positions of the same size can carry very different amounts of margin.
• Distance to liquidation combines both: how far the mark price has to move before the position falls below its maintenance margin. It is the number that decides whether a position matters for the next hours.

When a large position does reach its liquidation price, Hyperliquid first tries to close it with market orders on the order book; if that fails, a backstop vault takes the position over.² For a very large position this is where the order book's depth matters: the closing orders can move the price, which is why a single whale near its liquidation price shows up so clearly on a heatmap.

Size and distance together also show why the largest positions are not automatically the most fragile ones. In the table above, the four largest positions all used 5x and sat far from their liquidation prices; the closest were the ETH longs further down the list, at 10x and 9x.

What the data does not show

Public positions are a strong dataset, but not a complete one:

• Tracked wallets only. We track a large set of active wallets, not every account on Hyperliquid. Totals are sums of tracked positions, not the venue's total open interest.
• Positions, not people. A wallet address is not an identity. One person or fund can use many wallets, and a short on Hyperliquid can be the hedge of a holding somewhere else: on another venue, in a vault or in spot. We show short addresses, and a name only where a public source gives one, such as the name an account owner chose for the official leaderboard.
• Leverage is a setting. The figure is the leverage Hyperliquid reports for the position. In cross margin the effective leverage of the whole account can be lower, because other collateral backs it.
• A moment in time. Positions change every minute. The tables above are frozen at their snapshot; the live pages are not.

Using it for your own positions

The same data that shows the largest wallets can watch a single one. Liquidation guard is read-only: paste a Hyperliquid address and it shows how close each of its positions is to liquidation, which one liquidates first, and can send an alert before the price gets there. No wallet connection and no keys are needed, because the positions are public.

For the bigger picture, Whales lists the largest positions and position changes as they happen, and every coin has a whales tab, for example BTC. Traders ranks wallets by their results on the official leaderboard.

The numbers in this post are frozen at the snapshot time shown on each block. For the current positions, open Guard or the whale pages.