The streak is over. U.S. spot Bitcoin ETFs recorded a net outflow of $225.1 million on Thursday, July 23, per settled Farside Investors data — ending seven consecutive sessions of inflows that had accumulated $999.3 million between July 14 and July 22, the complex's longest positive run since late April. BlackRock's IBIT accounted for $202.5 million of the redemptions — roughly 90% of the day's total, per CryptoSlate. This update settles the numbers, grades the marker we set on it, and lays out the tells for next week. Background reading: Friday's news report and our guide to reading flow data.
The settled arithmetic
| Date (2026) | Total (US$m) | IBIT | Note |
|---|---|---|---|
| Jul 14 | +181.1 | +138.9 | Streak day 1 |
| Jul 15 | +107.7 | +80.8 | — |
| Jul 16 | +79.1 | +33.4 | FBTC +30.7 |
| Jul 17 | +132.3 | +136.5 | FBTC −4.2 |
| Jul 20 | +226.8 | +116.5 | Streak peak; ARKB +72.7 |
| Jul 21 | +203.2 | +163.9 | — |
| Jul 22 | +69.1 | +38.8 | Streak day 7 (final) |
| Jul 23 | −225.1 | −202.5 | Streak broken |
US spot Bitcoin ETF daily net flows, settled data. Source: Farside Investors, farside.co.uk/btc, retrieved July 25, 2026.
The other Thursday cells: FBTC −$5.6M, BITB −$7.0M, ARKB −$4.3M, EZBC −$5.6M, BTCW −$5.1M, with Morgan Stanley's MSBT the only fund positive at +$5.0M. The outflow erased roughly 22.5% of the streak's accumulation in a single session, per CryptoSlate. Zoom out further and the streak itself looks smaller: those seven green days recovered only about 15% of June's outflows, per Yahoo Finance. Cumulative since launch: IBIT +$60.6B, GBTC −$27.4B, complex-wide +$51.7B, per Farside. Friday's (Jul 24) cell had not settled at publication — consistent with the one-day lag we documented last week.
Marker W4: graded, failed
On July 24 we set marker W4: the inflow streak survives Fed week with no single day worse than −$100 million. It failed before Fed week began — Thursday's −$225.1M broke both conditions at once. We grade it failed rather than quietly re-scoping the window to start Monday; the point of pre-registered markers is that they cost something when wrong. What we got wrong was not the flows' direction so much as the exogenous input: the marker implicitly assumed the week's risk would arrive via the Fed, and it arrived two days early via the Red Sea. Thursday was the day Brent crude settled above $100 after Houthi strikes on two Saudi tankers, per CNBC, the 10-year Treasury yield hit an 18-month high near 4.7%, and rate markets repriced Wednesday's FOMC from a ~89% hold toward 64/36, per CBS News. Institutional de-risking through the most liquid crypto wrapper on the shelf is exactly what that macro tape predicts — the redemptions were the symptom, not the story.
What didn't happen
Three absences worth logging, because they bound how bearish Thursday's cell actually is. First, no GBTC bleed: Grayscale's legacy fund — the complex's reflexive seller in every prior stress episode — sat at zero, and its low-fee Mini Trust wasn't a source of redemptions either. Second, no cascade: the outflow did not produce the derivatives liquidation spiral that −$200M days triggered in June; Bitcoin's drop stopped at the familiar $63,000–$64,000 shelf rather than slicing through it. Third, no full-complex exit: seven of twelve funds printed zero rather than negative, and MSBT stayed green — this was one desk-driven redemption through the biggest wrapper, not a synchronized institutional exodus. A genuine regime change would look different on all three counts.
The July 13 precedent — and the tells for next week
One outflow day is a data point, not a regime. The nearest precedent is barely two weeks old: on July 13, the complex bled −$424.7M in a single session — nearly double Thursday — and then flipped straight into the seven-day streak the very next day. Single-day capitulation cells in this tape have marked local turning points as often as trend starts. The distinction to watch: clustering. In June's grind lower, −$100M+ days came in bunches; the streak's whole significance was that they stopped. So the tells, in order: (1) Friday's cell when it settles — a small negative or flat print keeps this a one-day event, a second −$200M day starts a pattern; (2) Monday and Tuesday, the pre-FOMC sessions — de-risking ahead of a 64/36 meeting would be unsurprising, which paradoxically makes any inflow on those days a strong signal; (3) Wednesday's cell itself, settled Thursday, as the verdict on the decision. We'll grade all of it, alongside markers W1, W3 and W5, in next week's coverage — the full Fed scenario map is in today's analysis.
The wider frame hasn't changed: flows now explain an estimated 45% of weekly Bitcoin price variance, the wrapper concentrates decision-making in a handful of institutional desks — IBIT alone was 90% of Thursday's move — and those desks answer to the bond market before they answer to any halving-cycle model. Until the oil chart calms down, the flow table is a derivative of the tanker map. What we own when we own these products, and who actually holds the coins, is covered in our ETF custody guide.
Disclaimer: This article is for informational purposes only and does not constitute investment, legal or tax advice. Cryptocurrency markets are highly volatile and you can lose some or all of your capital. Nothing here is a recommendation to buy or sell any asset. Always do your own research and consult a qualified professional before making investment decisions.