How Crypto Liquidations Work: What a $19 Billion Cascade Actually Looks Like

Crypto liquidations are the forced closure of leveraged positions when a trader's collateral can no longer cover their losses, and understanding the mechanism is the clearest way to make sense of a crash like the one on October 10, 2025. Most explanations of this topic are written for people who want to trade. This one is written for people who simply want to understand what happened.

If you have watched a crypto crash unfold and seen headlines about billions of dollars being "liquidated" in a matter of hours, you have probably wondered what that number actually means and where the money went. The honest answer is that most coverage explains the drama without explaining the machinery. This article fixes that. It walks through how a single liquidation works, how one liquidation can set off thousands more, and how those mechanics produced the October 2025 crash that erased a record amount of leveraged positions in a single day.

None of this is trading advice. The goal here is comprehension, not tactics. By the end you should be able to read a crash headline and understand the structure underneath it, which is a more durable skill than any single market call.

What a crypto liquidation actually is

To see why this happens, start with leverage. Leverage lets a trader control a position larger than the money they put down. If someone wants 10,000 dollars of Bitcoin exposure but only deposits 1,000 dollars, they can use 10x leverage to control the full amount. That deposit is called margin. It works like a safety buffer for the borrowed exposure.

The buffer is thin, and that is the whole point of the risk. With 10x leverage, a price move of roughly 10 percent against the position is enough to wipe out the deposit. The exchange does not wait for the collateral to reach zero. It sets a liquidation price, a level at which the position is automatically closed to make sure losses do not exceed the deposit. The higher the leverage, the closer that liquidation price sits to the entry price, which leaves very little room for normal market movement before a forced closure happens. Exchange documentation from major venues describes this same mechanism consistently, because it is the basic control that keeps a leveraged account from going into negative equity.

Liquidation can be partial or total. A partial liquidation closes only enough of the position to bring the account back above the maintenance requirement. A total liquidation closes the whole thing. Better-designed systems try partial, laddered closures first, because closing a position in pieces is less destabilising than dumping the entire thing onto the market in one go.

How one liquidation becomes a cascade

A single liquidation is a private event between one trader and one exchange. A liquidation cascade is what happens when many of them fire at once and start feeding each other. This is the mechanism behind almost every violent crypto crash, and it is the part most explanations skip.

Two safety mechanisms sit underneath this loop. The first is the insurance fund, a pool of capital the exchange keeps to absorb losses when a position closes at a worse price than its collateral can cover. The second only appears when the insurance fund is overwhelmed. It is called auto-deleveraging, and it is the least understood part of the system. When liquidations cannot find buyers and the insurance fund runs dry, the exchange force-closes the profitable positions of winning traders to cover the shortfall from the insolvent ones. Auto-deleveraging is controversial because it penalises traders who were positioned correctly, but it is the mechanism that lets venues offer round-the-clock leverage without a guaranteed counterparty on the other side of every trade.

Understanding this loop is not academic. It is the difference between reading a crash as random chaos and reading it as a predictable consequence of how leverage and liquidity interact. When you understand the cascade, you also understand why the same kind of event keeps recurring in different forms. The instruments and the venues change. The leverage-plus-thin-liquidity mechanism does not. If you want to see how the surrounding plumbing works, our explainer on how crypto exchanges work and why they sometimes fail covers the custody and counterparty side of the same system.

The amplifiers: cross-margin and collateral pricing

A basic cascade only needs leverage and a falling price. The worst events add two amplifiers that turn a sharp correction into a systemic one.

The first is cross-margin. In a cross-margin account, all of a trader's positions share one collateral pool. That is efficient when markets are calm. It is dangerous when one asset moves violently, because a loss in that single asset can drain the collateral supporting completely unrelated positions. A bad moment in one corner of the portfolio can force-liquidate the rest of it.

The second amplifier is how an exchange prices the collateral itself. Every leveraged position is valued against a reference price, and where that price comes from matters enormously. If an exchange marks collateral using its own internal order book rather than a robust, multi-venue price feed, then a local price dislocation on that one venue becomes the accounting truth for every account using that asset as collateral. A price that is only real on one exchange gets treated as real everywhere inside that exchange. This is exactly the failure that turned October 10 from a bad day into a record one.

October 10, 2025: a cascade in real numbers

The crash began on the evening of Friday, October 10, 2025, when a sudden escalation in United States and China trade tensions triggered a broad risk-off move across global markets. Because crypto trades 24 hours a day with no circuit breakers, and because it sat near record leverage heading into a weekend when liquidity is naturally thin, it absorbed the shock with nothing to slow the fall. What followed was the largest single-day liquidation event in crypto's history.

The figures come from derivatives data tracked by groups like CoinDesk Research and intraday analysis from Amberdata, and they tell a clear structural story. As the analysis from FTI Consulting put it, this was leverage meeting liquidity. Both amplifiers were present. Exchange infrastructure also buckled under the load, with several major venues reporting outages and delays that left traders unable to add collateral or close positions, which turned survivable situations into total liquidations according to reporting from CoinShares.

There is a second part of the October story that almost everyone got wrong, and it matters for understanding the difference between a market mechanism and an asset failure.

What the USDe headlines got wrong

During the cascade, a synthetic dollar token called USDe briefly printed as low as 0.65 dollars on one exchange, while holding its roughly one-dollar value on every other major venue at the same time. Many headlines compressed this into "a stablecoin collapsed." That framing is inaccurate, and the distinction is worth getting right.

Here is the mechanism. The exchange in question priced certain collateral, including USDe, using its own internal order book rather than a multi-venue price feed. When thin local liquidity and a large sell order dragged the local price down, that depressed mark was applied to every cross-margined account holding the asset as collateral. Positions that were perfectly solvent under accurate pricing were force-liquidated against a price that was only real inside one venue. The asset did not break. The pricing did. USDe is a synthetic dollar, backed by a hedging strategy rather than cash reserves, which is a different design from a fiat-backed stablecoin and worth understanding on its own terms. Our explainer on how stablecoin payments move money covers where different stablecoin models hold up and where they get stuck.

This is why the cross-margin and collateral-pricing amplifiers are not footnotes. They are the reason a contained sell-off became a record liquidation event.

The part the fear coverage leaves out

Crash coverage tends to treat liquidations as a permanent state of emergency. The data says otherwise. Across all of 2025, forced liquidations in crypto derivatives totaled an estimated 150 billion dollars or more, but the vast majority of that was routine daily activity, typically tens to low hundreds of millions of dollars per day. That flow is closer to a maintenance function of a leveraged market than a crisis. Cascades like October 10 are the rare, dangerous tail, not the normal state of the market.

This reframing matters because it points at the real cause. The October crash was not evidence that "crypto is broken." It was evidence of what concentrated leverage does when a shock arrives and liquidity is thin. Bitcoin and Ethereum's underlying networks kept running normally throughout. Spot holders who used no leverage were not force-sold. The damage was concentrated almost entirely in leveraged derivatives positions. Leverage, not crypto, produced the wipeout.

One of the most common ways people get hurt in events like this is using leverage without understanding the liquidation risk attached to it. It is an easy mistake to make, because the interfaces make leverage feel like a simple multiplier on gains and rarely make the liquidation price obvious before the trade. The cost of that gap only becomes clear during a cascade, when positions close faster than anyone can react. Understanding the mechanism before approaching leveraged products, rather than after, is the difference between an informed decision and a forced one.

Where this fits in a wider understanding of risk

Reading a crash correctly is a risk-literacy skill, and it connects to a few others. Knowing how liquidations cascade helps you interpret on-chain and market data without panicking, and it pairs naturally with knowing how to evaluate the venues and assets involved. Blockready's Trading module covers leverage, liquidation, and market psychology as distinct topics, and the DeFi module covers stablecoins, lending protocols, and oracles, because the October event sat exactly where those subjects overlap. Treating them as separate building blocks is what makes a cascade legible rather than frightening.

From a curriculum-design perspective, our view is that market-structure events like this should be taught as mechanism first and headline second. The fear-led version of crypto education trains people to react to crashes emotionally. The mechanism-led version trains them to recognise the structure underneath, which is what actually reduces costly decisions. A reader who understands why longs get wiped out in a cascade, why cross-margin spreads risk, and why a single-venue price can lie, is far better equipped to assess any future crash than one who has only memorised that "crypto is volatile." That is the difference between being informed and being capable, and it is the standard we hold our own material to. If you want a structured way to evaluate the assets and platforms involved before any of this becomes relevant to your own money, our DYOR checklist and our guide to the most common crypto mistakes are good next steps.

Frequently Asked Questions