
Upbit's hourly volume surged to $830M during the Aug. 22 flash crash, with XRP accounting for 32% of activity as liquidations hit $523M.
South Korean crypto exchange Upbit processed an estimated 1.15 trillion won, roughly $830 million, during one hour of volatile trading on Aug. 22.
The volume spike occurred around 05:00 UTC, when Bitcoin, XRP and other major cryptocurrencies recorded abrupt downward price wicks before recovering part of their losses. The resulting liquidations and rapid repositioning drove spot activity higher across South Korea’s largest exchanges.
Upbit’s rolling 24-hour volume reached approximately $3.818 billion during the volatility, according to figures published on X. Bithumb recorded about $1.954 billion, while Coinone processed approximately $172 million. These numbers were snapshots from rolling 24-hour periods and change continuously as older transactions leave the measurement window. They should not be treated as audited revenue or final daily figures from the exchanges.
The activity extended a recovery that began before the flash crash. The preceding Bitcoin rally lifted Upbit’s daily volume by 273% to approximately $1.84 billion on Aug. 21, based on CoinGecko data. That was Upbit’s highest reported daily volume since mid-March. Bithumb’s activity also climbed 132.9% to approximately $934.9 million during the earlier measurement period.
XRP accounted for 32.20% of Upbit’s measured volume during the surge, according to Upbit Datalab figures cited by market-data publications. TRUMP followed with 10.93%, while USDT represented 8.39%. Ether and Bitcoin accounted for 5.44% and 5.40%, respectively. The distribution shows that the increase was not driven solely by Bitcoin, even though its broader market rally helped restore trading interest.
XRP had risen strongly before the reversal. It gained more than 14% on Aug. 21 and briefly traded above important moving averages. The concentration also illustrates Upbit’s importance to XRP liquidity. Korean won-denominated XRP markets have regularly generated more turnover than the exchange’s Bitcoin pairs during periods of heightened retail activity.
The broader market move triggered approximately $523 million in liquidations within one hour, according to data attributed to CoinGlass. Long positions accounted for about $448 million, while short liquidations reached $74.76 million. Across the measured 24-hour period, liquidations reportedly approached $1.8 billion and affected more than 286,000 traders. The largest identified single liquidation was a BTC-USD position worth approximately $24.96 million on Hyperliquid.
Those figures suggest that forced position closures intensified the decline. When prices fall through liquidation thresholds, derivatives exchanges automatically close leveraged long positions. Those market sales can push prices lower and trigger another round of liquidations.
No verified exchange statement identified a technical fault at Upbit as the cause. Available evidence points to a broader market deleveraging event rather than an isolated Upbit malfunction. The precise order or trade that initiated the move has not been publicly established.
The sharp increase followed months of declining South Korean exchange activity. Trading across the country’s five won-based exchanges fell 54.6% during the first half of 2026 compared with the previous year. Upbit nevertheless strengthened its domestic position. Its share of trading across the five exchanges reached approximately 67.4% in July, concentrating much of the market’s returning liquidity on one platform.
The rebound does not yet establish a sustained recovery. A single volatile session can inflate exchange volume because the same assets may change hands repeatedly as traders enter, exit and hedge positions.
Whether Upbit’s activity stays elevated after the liquidation-driven turnover leaves the 24-hour window remains to be seen. Traders are watching XRP and other heavily traded assets for renewed price gaps, thinner order books or another increase in leveraged positions.
Prepared with AlphaScala editorial tooling from the source reporting linked above. Indexable analysis may include a cited Alpha Score value. Publishing checks screen each story before release. Educational coverage, not personalized advice.