NVDA Liquidation Map: Liquidation Levels by Leverage (Estimated)
Model est-1.0+hl-2026-09-27-25x · High leverage (≥25×) · positions opened in the last 3 days. Each bar is 0.8342 wide. Outside the chart: ≈ $0 of longs and $0 of shorts.
See how these levels built up over time on the NVDA liquidation heatmap.
This is a model estimate for all tracked exchanges, not real positions.
An estimated liquidation map for NVDA: liquidation levels by leverage, the modelled value of positions that would be liquidated at each price if it got there, longs below the current price and shorts above, from positions opened in the chosen window across Binance, Bybit, Hyperliquid, Bitget, Gate, HTX, MEXC, KuCoin, Aster, Lighter. Right now the nearest modelled NVDA long cluster is at $153.08 (-8.25%) and the nearest short cluster at $169.76 (+1.75%); about $262K of longs and $3.86M of shorts are modelled within ±5% of the $166.84 price. Looking for where liquidations may cluster over time? See the NVDA liquidation heatmap. Looking for liquidations that already happened? See NVDA liquidations.
The model places $6.06M of long and $7.22M of short liquidation levels for NVDA from positions opened in the last 3 days (high leverage (≥25×)). Within 10% of the 166.84 price: about $2.42M of longs below and $6.75M of shorts above. The densest modelled clusters are longs near 148.91 (-10.75%) and shorts near 170.60 (+2.25%).
Liquidation maps for other coins
The densest estimated long and short clusters, as tables with their distance and size: NVDA liquidation levels.
Modelled liquidations near the price
High leverage (≥25×) · last 3 days| Price move | Longs liquidated (price falls) | Shorts liquidated (price rises) |
|---|---|---|
| ±2% | $12K | $1.45M |
| ±5% | $262K | $3.86M |
| ±10% | $2.42M | $6.75M |
How this estimate is made
Model est-1.0+hl-2026-09-27-25xInputs are the open interest, prices and volume this site records from Binance, Bybit, Hyperliquid, Bitget, Gate, HTX, MEXC, KuCoin, Aster, Lighter. Every rise in open interest is booked as a new long and a new short at the period's volume-weighted price; falls in open interest close positions proportionally; positions also decay with a multi-day half-life; and a level is removed once the price trades through it (those are the "modelled liquidations"). Liquidation prices are spread over this distance-from-entry distribution, measured on real Hyperliquid positions and cut to positions at ≥25× leverage (the distribution table is on the NVDA liquidation heatmap).
Known limits: prices are last-trade bars averaged across exchanges, not mark prices; opens and closes inside one open-interest snapshot (5 minutes on Binance) cancel out; one distribution is used for every coin and both sides. USD values are "≈ modelled", best read as relative intensity. Full method and a daily accuracy scoreboard against real liquidations: liquidation heatmap methodology. Estimates may be materially wrong; not financial advice.
NVDA liquidation map FAQ
Is this NVDA liquidation map real data?
No, it is a model estimate. Exchanges do not publish where their users' positions would be liquidated, so every liquidation heatmap for Binance or Bybit (including the well-known ones) is a model. This one starts from real data: the open interest changes, prices and trading volume this site records from the exchanges' public APIs. Actual liquidations are usually lower, because most positions are closed long before they are liquidated.
How is it calculated?
Every rise in open interest is a new long and a new short, entered at that period's volume-weighted price. The model spreads their liquidation prices over a distance-to-liquidation distribution taken from a notional-weighted sample of Hyperliquid positions, where the distance was measured from the mark price; the model reuses it as an assumption about the distance from entry when it allocates new notional. It places 74.0% of modeled new notional within 50% of entry; the rest is assumed too far away or not liquidatable. Falls in open interest close positions proportionally, positions also decay with a 5-day half-life, and a level is removed once the price trades through it. The model version on the chart says which formula and distribution were used, and the methodology page describes the sample.
What do the ranges and the ≥25× view mean?
The range is the replay window: the map shows positions opened in the last 12 hours to 30 days. All leverage uses the whole distribution; High leverage (≥25×) keeps only positions liquidated within 4% of their entry, the short-term levels that get hit first.
How is this different from the Hyperliquid liquidation map?
Hyperliquid is an on-chain exchange, so its positions and their liquidation prices are public. The Hyperliquid liquidation map adds up the real positions of tracked wallets on Hyperliquid only (there is none for NVDA, which has no Hyperliquid perp). This page is an estimate for all tracked exchanges together, Binance, Bybit and Hyperliquid included. The two are not independent checks of each other: their scope differs (one venue's tracked wallets versus a multi-venue model), the tracked wallets are only a sample of Hyperliquid positions, the two are refreshed at different times, and this model's leverage distribution is itself measured on Hyperliquid positions. A level that shows on both is not shown to be more accurate, and neither predicts that the price will get there.
Can I trade on it?
Treat it as context, not a signal. Clusters show where forced closes may bunch up if the price gets there; they don't predict that it will. Nothing here is financial advice, estimates may be materially wrong, and derivatives trading can lose more than your deposit.