Inside the largest flash crash in crypto market history
On the night between October 10 and 11, the cryptocurrency market experienced its largest flash crash in history, with the entire crypto heatmap turning red. Within 24 hours, liquidations reached $19-20 billion, and total market capitalization plunged by $370 billion. The crash was nine times larger than any previous event, with approximately 1.6 million accounts wiped out.
Bitcoin and Ethereum prices fell by 15-20% in a matter of minutes. The most famous altcoins, such as Solana, dropped by more than 30%, while smaller tokens lost between 50% and 90%, with some nearly reaching zero.
On Friday, October 10, at 20:50 UTC, Donald Trump announced on his Truth Social platform additional 100% tariffs on all Chinese imports starting November 1st. The move was a response to China’s tariffs on rare earth metal exports. As Trump’s announcement was made when traditional markets were already closed for the weekend, the crypto market absorbed the entire shock.
However, the tariff shock was only one of the factors. The crypto market was already overloaded with leverage, funding rates and open interest at all-time highs. Recent regulatory changes by the Trump administration allowed traders on centralized exchanges to have 10x leverage, with some able to operate with leverage up to 100x. During the crash, exchanges were forced to liquidate leveraged positions to prevent their own losses, thereby aggravating the crisis.
This process, known as Auto-deleveraging (ADL), is an exchange’s last resort. It activates when liquidation losses exceed the capacity of its insurance fund, and automatically reduces profitable positions on the opposite side to cover the deficit. During the extreme volatility of October 10-11, several platforms had to trigger ADL, closing profitable short positions to balance their books. In a zero-sum derivatives market, if longs go bankrupt and no new longs replace them, there isn’t enough money to pay all winning shorts.
But why were order books empty, forcing exchanges to close positions? The nature of the problem is structural: market makers — who provide market liquidity by ensuring that a trader wanting to buy or sell always has a counterparty — vanished from the market. The result was a vicious circle: shock, withdrawal of market makers, depletion of insurance funds, more liquidations, and further shock.
However, the initial shock was not purely external to the cryptocurrency system, such as the renewed U.S.-China trade tensions. It was also compounded by a targeted attack on oracles, applications that allow exchanges to have uniform cryptocurrency prices across platforms, at least to a certain degree. On-chain analysis suggests that an attack occurred, distorting price feeds and widening bid-ask spreads. As a result, market makers saw their margins on spreads drastically reduced, and, not obliged to remain active during periods of extreme volatility in the unregulated crypto market, withdrew liquidity, triggering the flash crash.
Below are displayed the Bitcoin charts on Kraken and Binance at the time of the crash. As you can see, the lows recorded were significantly different:
Same with Ethereum, with data taken from Bitstamp and Coinbase respectively:
On October 6, Binance announced a transition to oracle-based pricing, but implementation was delayed until October 14. During this period, the system relied excessively on Binance’s internal spot prices.
This flash crash was roughly 160 times larger than the largest previous oracle attack, but the methodology remained the same as in 2020: identify an oracle dependency from a manipulable source, calculate the cost of manipulation (approximately $60 million in tokens dumped), execute, and profit.
The incident once again highlighted the human factor as a source of fragility in the system. The crypto market must learn from these mistakes by limiting excessive leverage and developing a more secure price-feed system similar to that of traditional markets, so events like this will not undermine investor confidence in the future.