Bitcoin's best August since 2017 ran on $2.75bn of forced short covering

Divide the record day by the record week and you get the least flattering number of the month: $2.75bn of short liquidations in a single session against $1.92bn of US spot ETF inflows across the entire rally week is a ratio of 1.43, so the forced closing of bearish bets moved about 43% more capital in a day than the regulated wrappers pension funds and wealth managers can actually buy moved in five ^{[3][4][8][14]}. Spread that week across five trading sessions and the ETF bid averages roughly $384m a day, which makes the cascade worth something close to seven sessions of institutional buying, compressed into one ^[15].

The shorts were there for a reason, and that reason is the part worth keeping. If bitcoin finished August down roughly 28-33% on the year after gaining 25-30% during the month, it entered August down somewhere between 42% and 48% year to date ^[1][10][16], and a chart like that recruits sellers the way a bull market recruits buyers. The 200-day moving average near $69,000 sat only about 8-11% above the early-August lows of $62,000-$64,000 ^[1][5][12], close enough that a modest bounce tripped the systematic strategies watching that line, and the systematic bid then ran into other people's stops ^[5]. Liquidity expectations tied to the US Treasury buyback results, per Crypto Briefing, gave the discretionary money its excuse ^[6].

The recurring number is the smaller one. Annualise $1.92bn a week and you get roughly $100bn a year ^[13], which nobody should annualise, but the exercise frames the gap: a 25-30% month against a since-2013 August median of about negative 7% is a 32-37 point swing away from seasonality ^[1][9][18], and of the two drivers Crypto Briefing names, only one can print that twice.

The 2017 comparison carries its own arithmetic. That August's roughly 65% gain did precede a run from about $4,700 to nearly $20,000 by December ^[2][7], but that market had no ETFs and much thinner institutional participation, with regulators mostly confused rather than actively engaged ^[11]; the vehicle that delivered $1.92bn last week is also the vehicle through which redemptions clear.

This is probably wrong in one specific way, but the view: August was a positioning unwind that borrowed its magnitude from the shorts, and at $79,000 the asset still needs 59% to reclaim the $126,000 it printed earlier in 2025 ^[1][10][17], which $384m a day does not buy in a hurry. The counter-thesis has the better tape, or rather the more durable one, since nearly $2bn into regulated products in a week is allocation rather than leverage, and allocation compounds where a squeeze does not. What would settle it is dull: two consecutive weeks of inflows near that $1.92bn pace with no liquidation event behind them, plus a clean break of the $81,000 level Crypto Briefing calls the first real test ^[19], and the mechanism read is wrong. One caveat on my own figures. The source reports ranges wide enough that my end-of-July drawdown spans 42% to 48% ^[16], so treat the direction as the finding and the decimals as decoration.