ETF Flow Focus · Published

Bitcoin ETF Inflows: IBIT Slows to $578M From $1.73B

US spot bitcoin ETF inflows enter November resting on a single fund, and that fund is slowing. Over the trailing 20 sessions the category has shed a net -$808.5 million, and BlackRock's IBIT is the only product of twelve holding the number up, at +$578.3 million. That figure stood at +$1,727.3 million on October 19 and +$851.6 million on October 23 — a two-thirds decline in nine sessions, almost a third of it in the last five. Every other US spot bitcoin ETF combined sits at -$1,386.8 million across the same window, the weakest ex-IBIT reading since June 10.

Saturday's recap covered the completed October 26-30 week and the monthly review closed out October. This column looks forward: what the data says about the week now starting.

Where do bitcoin ETF inflows stand after a $643M Friday?

The category's most recent session, Friday October 30, was a net -$643.2 million. bryptoflows scores it at a z-score of -1.78 and the 1.7th percentile of all 720 sessions on record since January 2024 — a genuine tail event rather than an ordinary down day.

The internals were more one-sided than the headline. Gross inflows across all twelve products totaled exactly $0.0 million against -$643.2 million of gross outflows. Not a single fund reported a positive figure: zero of twelve positive, five negative, seven flat. A session with no gross inflow at all has happened 109 times in 720 sessions, or 15.1% of the time — uncommon rather than unprecedented, though October produced four of them, on the 7th, 12th, 16th and 30th.

The rolling windows, all current through the October 30 close:

Cumulative net flow into the category stands at $57,048.1 million, which is -$5,687.8 million below the all-time cumulative peak of $62,735.9 million — a drawdown of roughly 9.1%. October closed at -$515.9 million, the first net-negative calendar month since June, which also makes quarter-to-date -$515.9 million with November yet to print a session. Year-to-date 2026 is +$453.9 million across 209 sessions. That cushion is thin enough to matter: a five-session total of -$453.9 million or worse has occurred in 144 of 716 windows on record, or 20.1% of the time, so a single ordinary-bad week would put 2026 negative.

What usually happens after a session this bad

This is the most useful thing the flow history has to say about the week ahead, and it is not encouraging. Including Friday, the category has recorded 15 sessions of -$600 million or worse. Fourteen have five sessions of data after them:

Fourteen observations is a thin sample, and two things cut against reading it hard. Eight of the fifteen worst sessions fall in 2026 alone, so the pattern is weighted toward one year's conditions. And the most recent precedent broke it: October 22's -$635.6 million session was followed by five sessions worth +$302.9 million on net, one of the three positive cases.

Still, the balance of the record says that the category's worst sessions have tended to cluster rather than mark a bottom, and nothing in Friday's internals — no gross inflow whatsoever — argues that buyers stepped in as it happened.

The rolling windows improve this week on arithmetic alone

Last Monday's column made two mechanical projections and both landed, which establishes what this section is and is not. It put the ten-session base at +$301.2 million and the twenty-session at -$606.1 million. The week printed -$202.4 million; the ten-session window now reads +$98.8 million and the twenty-session -$808.5 million — exactly base plus the week.

The setup for this week is the reverse of last week's, and anyone reading a single rolling average as a demand signal will likely be misled in the opposite direction:

So the twenty-session number is set up to look materially better by Friday with no improvement in demand, and the ten-session worse. This is arithmetic, not forecast: it describes what the windows do if demand does nothing, which is the one outcome that never quite occurs.

Which issuers are driving bitcoin ETF outflows right now

The aggregate figure is everywhere by 9am. The issuer split is where the information is, and over the trailing 20 sessions it reads:

For the year, IBIT inflows of +$4,218.2 million stand against -$1,942.2 million from FBTC and -$2,794.1 million from GBTC, with Grayscale's mini trust (+$1,017.5 million) and MSBT (+$810.4 million) the only other products above half a billion on the positive side, against ARKB's -$669.3 million on the other. Life-to-date, the BlackRock bitcoin ETF has drawn +$66,310.4 million, which is 116.2% of the category's entire cumulative net flow — more, by itself, than the category has taken in, because GBTC has given back -$28,048.6 million on 67 positive days against 363 negative ones.

What is new this week is not that IBIT dominates but the rate of change. Its twenty-session reading sits at the 29.7th percentile of its 701 historical readings, the lowest since August 4, when it printed +$504.9 million. Last Monday this column argued the thing to watch was the level rather than the streak, since a two-day negative run is ordinary for a fund with 455 positive days against 174 negative. The level has since fallen by $273 million in five sessions.

IBIT gets no mechanical help this week: only +$1.1 million of its twenty-session total rolls off, leaving its base at +$577.2 million. For that figure to go negative by Friday the fund would have to shed more than $577.2 million over five sessions, which has happened in 48 of 716 windows, or 6.7% of the time. The mirror image is harsher: for the ex-IBIT complex to turn its twenty-session reading positive, the other eleven funds would need +$706.8 million collectively, achieved in 7.7% of windows. Neither extreme is the base case.

Breadth is narrow, but mostly because small funds are dormant

Over the last 20 sessions the median number of products posting a nonzero figure was three of twelve. IBIT traded a nonzero amount on all 20. bryptoflows records an unreported product as null and never as zero, so a 0.0 in the data is a genuine reported zero rather than missing data — and most of the zeros belong to funds that have been dormant for weeks, not to a buyer base that evaporated.

The honest reading is that narrow breadth here measures how few products are active, not how concentrated institutional demand has become. IBIT accounted for 53.6% of all gross flow magnitude over the trailing 20 sessions, below its 58.6% life-to-date gross share. Friday's single-session product HHI of 0.279 came in below the 0.414 life-to-date average precisely because the selling was spread across five funds rather than dominated by one.

November has no seasonal signal worth trading

Two Novembers are on record and they point in opposite directions, so anyone offering you a November seasonal for spot bitcoin ETF flows is working from a sample of two. November 2024 was +$6,465.3 million across 21 sessions — the best calendar month in the category's history. November 2025 was -$3,466.7 million across 19 sessions — the third worst, behind June 2026 and February 2025. The first five sessions of each were +$1,282.9 million and -$1,208.4 million respectively.

That is the whole seasonal record, and it says nothing. The useful forward anchors are the mechanical ones above, plus the fact that this week carries no Federal Reserve decision: third-party Fed calendars retrieved this run put the next FOMC meeting on December 8-9, 2026, dates the Fed itself publishes as tentative. Whatever moves flows over the next five sessions will not be a rate announcement.

bryptoflows' composite institutional signal reads -32.2 on data through October 30, a regime it labels moderate distribution, with all eight components live and agreement at 64.4%. That is down from the -22.1 this column reported a week ago on data through October 23 — the deterioration is driven by the twenty-session flow reading, Friday's zero-positive breadth, and a twenty-session price return of -5.75%. On the corporate side, Strategy's most recent 8-K, filed October 26, reported no bitcoin purchases and no sales, leaving holdings at 848,517 BTC, and nothing new has landed since; Thursday's treasury column owns that subject.

What would have to happen for the pattern to break

Three things would mark a genuine change rather than noise, in rough order of how much they would tell you:

And the cleanest falsifiable marker: if the twenty-session window is still negative on Friday, demand genuinely deteriorated, because the arithmetic was handing it a $678.9 million improvement for free.

What this flow data cannot tell you

Daily net creations and redemptions net every participant together and do not distinguish motive. A redemption may be a basis trade unwinding, an advisor rebalancing, or a holder leaving; nothing above identifies who sold. The data also says nothing about bitcoin held outside US spot ETFs — offshore vehicles, futures, direct custody, or corporate treasuries.

The conditional statistic deserves the heaviest caveat: fourteen prior observations is a thin base, they are not independent of each other, and the most recent one went the other way.

Track these numbers live on the flow pulse overview and ETF flow dashboards, and Strategy's treasury history on the MSTR breakdown. All flow, attribution and concentration figures are read directly from bryptoflows' own read-only API, sourced from Farside Investors and current through the October 30 close; Strategy detail is from SEC filings through the October 26 8-K. Federal Reserve meeting dates are from third-party calendars retrieved on publication. Retrieved and published November 2, 2026.