Bitcoin Breaks Below $80,000 as Inflation Shock and AI Selloff Hit Crypto

Bitcoin fell below $80,000 on Thursday for the first time in a month, extending a selloff that followed Wednesday’s U.S. inflation report and spread alongside losses in AI stocks. At 9:54 a.m. EDT on October 15, CoinDesk reported a price of $79,950, down 3.5% over the preceding 24 hours. That is a time-stamped market snapshot, not a closing price.

The retreat was harsher in Ether. In a separate report published at 12:25 a.m. EDT, CoinDesk put Bitcoin’s rolling daily decline near 3% and Ether’s above 6%. The different observation times matter: those percentages should not be treated as simultaneous quotes from the later U.S. session.

Borrowed positions amplified the decline

CoinDesk, citing CoinGlass data checked during Asian morning trading, reported about $483 million in Bitcoin and Ether futures liquidations over 24 hours. Ether accounted for roughly $286 million, against $197 million for Bitcoin. Those figures describe positions forcibly closed, rather than an equivalent amount of investor cash disappearing.

Liquidations can turn an initial decline into additional selling. A leveraged buyer must maintain collateral; when that cushion becomes insufficient, the exchange can close the position. That selling does not require a fresh change in the trader’s view. It follows the financing rules of the position.

The imbalance is evidence of heavier forced closures in Ether. It is consistent with leverage amplifying the relative decline, although liquidation totals do not establish why the initial selling began or explain every difference between the two assets. They also do not show that Ether’s network suffered an operational failure.

ETF demand reversed again

Farside’s table shows $261.4 million of net withdrawals from U.S. spot Bitcoin ETFs on October 14, following $175.7 million of net inflows on October 13. Fidelity’s FBTC and BlackRock’s IBIT accounted for most of Wednesday’s withdrawals. The two-day net is therefore negative $85.7 million, calculated from those reported totals.

A daily ETF flow measures net creations and redemptions. It is not the same as exchange trading volume, and it cannot identify all the investors behind a transaction. Wednesday’s withdrawals show that this channel was not providing a net inflow at a moment when prices were under pressure.

A rolling daily move is not a calendar-day return

A 24-hour price change moves with the observation time. A figure recorded during Asian trading and one recorded during the U.S. morning can cover different starting and ending prices even on the same date. Comparing them as though they were synchronized can create a false picture of the relative selloff.

The same discipline applies to liquidation totals. Different data cutoffs, venues and contract coverage can produce different aggregates. A larger later total does not necessarily identify a fresh burst of forced selling unless the reporting windows are reconciled. The figures here support the presence of substantial leveraged closures within the stated period, with their coverage and timing kept separate from the spot-price observations.

Inflation started the repricing; falling yields complicate the explanation

CoinDesk’s Thursday morning market report linked the retreat to a broader reduction in risk exposure after the inflation surprise. It also documented falling Treasury yields and rising gold. That combination rules out a neat account in which cryptocurrency was simply tracking a continuously rising U.S. bond yield throughout Thursday.

Investors can first reassess the risk of tighter monetary policy and then seek government bonds as equities weaken. The order of events matters. A bond rally later in the session does not erase the earlier inflation shock, but neither should the earlier shock be used to describe every subsequent market move.

The last day’s evidence points to several pressures working together: a less comfortable policy outlook, a retreat in speculative assets, ETF withdrawals and forced futures closures. No single dataset proves the contribution of each. A durable stabilization would require more than reclaiming a round number; it would need buying that persists after the forced selling subsides.