ETF Flow Focus · Published

Three Issuers Supply 114% of Bitcoin ETF Inflows

Bitcoin ETF inflows now depend on three fund families

US spot bitcoin ETF inflows open this week resting almost entirely on three issuers. Across the trailing twenty reported sessions through Friday, October 2, the twelve US spot funds took in a net +$2,186.7M — but BlackRock, Fidelity and Morgan Stanley supplied +$2,500.8M of that between them, or 114.4% of the category's net, while the remaining eight issuers were a combined -$314.1M drag. The headline number is healthy. The composition behind it is narrower than it was a week ago, and much narrower than a month ago.

That is the state of spot bitcoin ETF flows as this week opens, which is a different question from how last week finished. The completed reporting week of September 28 through October 2 landed at +$82.9M, covered session by session in Saturday's recap. What follows is who is actually buying, how thin participation has become, and the arithmetic that will move the trailing numbers this week whether or not a single new dollar arrives.

Who is actually driving bitcoin ETF inflows right now

bryptoflows' attribution history rolls each issuer's net flow over a twenty-session window. Through the October 2 close, that window reads:

Four issuers positive, one flat, six negative. A month earlier the same window looked materially different: through the September 3 close it totalled +$3,370.9M with six issuers positive, and the eight outside the top three were a net +$159.4M contribution rather than a drag. Through September 25 the count was five positive and the tail had already flipped to -$259.5M. The aggregate has softened, and the share of it coming from outside the top three has gone from mildly additive to clearly subtractive.

One nuance cuts against the simple "concentration is rising" reading, and it is worth stating plainly. BlackRock's own share of the twenty-session net has actually fallen over that month, from 84.6% on September 3 to 69.4% on September 25 to 74.8% now. What rose is Fidelity's, from 8.3% to 30.4% to 28.4%. So the category has not become more dependent on one fund over the past month — it has become more dependent on a slightly larger handful, while everyone outside that handful has gone from contributing to leaking. Those are different stories, and the aggregate HHI readings conflate them.

BlackRock's Friday figure is still missing, and that is a first

The most unusual thing in Friday's data is not a number but the absence of one. In bryptoflows' 700 reported observations since January 2024, IBIT has gone unreported on 17 of them. Sixteen were full market blackouts — every one of the twelve products null and a session total of exactly zero, the shape a US market holiday produces. October 2, 2026 is the only observation on record where IBIT alone is unreported while other funds reported normally. As of this article's data pull, three calendar days later, the figure still has not arrived.

That matters for how Friday should be read. The +$31.7M session total bryptoflows reports for October 2 is Fidelity's +$29.3M plus Morgan Stanley's +$2.4M, with nine products at a reported zero and the largest fund in the category simply absent. bryptoflows never substitutes zero for an unreported value, so that total, the week's +$82.9M and the twenty-session +$2,186.7M all exclude whatever IBIT actually did — a number that could be positive or negative, and which across IBIT's 683 reported sessions has been negative 166 times. BlackRock's twenty-session figure above is computed on 19 observations, at 95% coverage, for the same reason.

None of this says anything about the BlackRock bitcoin ETF itself. It is a gap in the upstream flow source, not a signal. It is worth flagging only because a single missing fund changes the apparent breadth of a session dramatically: Friday's two-positive, zero-negative split across eleven reporting products, and the 0.860 product concentration reading that came with it, are substantially artefacts of who reported rather than who bought.

Is breadth widening or narrowing this week?

Narrowing, on every measure bryptoflows tracks, with the caveat above attached to the single-day readings.

At the daily level, Friday showed 2 positive products out of 11 reporting — a positive-breadth reading of 18.2% — against 12 known products, for 91.7% coverage. At the issuer level over twenty sessions, the positive count has walked down from six to five to four over the past month. And the flow-breadth component inside bryptoflows' institutional support signal is now the composite's single largest detractor, contributing -8.9 points on a normalized score of -63.6. The composite itself reads 31.8 with a regime of moderate accumulation and confidence of 46.0, down from the 42.4 that Friday's institutional read reported — ETF flow momentum is still scoring positive at 52.2 off that $2,186.7M window, but breadth is dragging the whole reading lower.

The 2026 year-to-date picture is the same shape with a longer lens. Category net flow for the year stands at +$1,104.2M. IBIT alone is +$3,481.7M. The other eleven products, taken together, are -$2,377.5M. Only five of twelve are net positive on the year at all.

Grayscale is 2026's biggest buyer and its biggest seller at once

The most under-reported line in the issuer data belongs to Grayscale, because its two products point in opposite directions and the issuer-level net hides both.

Grayscale's newer BTC mini trust is +$1,043.2M year-to-date — the second-largest positive contributor in the category for 2026, ahead of Morgan Stanley's MSBT at +$785.2M and far ahead of every other non-BlackRock product. Legacy GBTC is -$2,641.5M over the same period, the category's largest source of bitcoin ETF outflows by a wide margin. Netted, Grayscale is -$1,598.3M on the year, which is why it shows up as a drag in the twenty-session table while quietly running one of the year's best-performing products.

GBTC's redemption pressure is also intermittent rather than steady, which makes it easy to misread in either direction. Twelve of its last twenty sessions printed exactly zero. Its most recent outflows were -$23.2M on September 28 and -$31.4M on October 1, following a stretch with nothing since September 16. Dormancy is GBTC's recent normal; it is not the same thing as the redemptions being over.

Morgan Stanley's MSBT remains the category's strangest record: 72 positive days and exactly one negative day in 124 reported sessions since April, for +$785.2M lifetime. It closed Friday on a two-session positive streak, one of only three live positive streaks in the category alongside IBIT and FBTC at one session each.

How does this run compare with the category's own history?

Less dramatic than the past three weeks have felt. Friday's +$31.7M carried a z-score of -0.17 and sat at the 47.9th percentile of all 700 sessions — almost exactly median. The five-session total of +$82.9M ranks above 285 of 696 comparable windows in the record, the 40.9th percentile; of those 696 windows, 436 were positive, so a small positive week is the category's most common outcome rather than a weak one.

What is genuinely unusual is the variance. The five-session window ending September 18 totalled +$6.1M, the smallest positive five-session window in the entire 696-window record. The window ending September 25 totalled +$2,385.8M, which ranks 43rd of 696. The category went from a near-record-low positive week to a top-7% week to an ordinary one inside fourteen sessions. Any narrative built on the pace of a single week over the past month was built on an unusually unstable base.

What would have to happen for this pattern to break

Two pieces of arithmetic will move the trailing numbers over the next five sessions regardless of what arrives, and they point in opposite directions.

The twenty-session window currently spans September 4 through October 2. The five oldest sessions in it total -$288.1M. If the coming week printed exactly zero every day, that window would rise to +$2,474.8M simply from dropping those sessions. The ten-session window is the mirror image: it currently spans September 21 through October 2 at +$2,468.7M, and the five sessions about to roll off it total +$2,385.8M. Five zero sessions would collapse it to +$82.9M.

So expect the medium-term reading to improve and the short-term reading to deteriorate this week on mechanics alone. Neither would be information about demand. The figures worth watching instead:

Price context has to come from outside bryptoflows, because the site's own candle feed is reporting unavailable in its source freshness check. Bitcoin was trading around $86,200 early Monday, roughly 32% below the $126,080 record set on October 6 last year, per a same-day report; separate coverage put the 2026 yearly open at $87,570 as overhead resistance after a brief visit above $87,000. Flow data establishes no causal link to price in either direction.

On the corporate side, Strategy's most recent 8-K, filed September 28, reported direction acquired — a purchase of 1,665 BTC for $142.7M, lifting holdings to 847,666 BTC — with the next filing due around today. Thursday's treasury piece covers the funding mix behind it, and the full filing history sits on the MSTR breakdown.

What this flow data cannot tell you

Net flow is creations minus redemptions in dollar terms. It cannot distinguish a smaller number of buyers from the same buyers writing smaller tickets, and it never identifies who is on the other side of a creation. A reported zero is a real zero, but a fund can print a flat net day with substantial offsetting creations and redemptions inside it — the data shows net, not churn.

Concentration readings in particular need care. A single-day HHI of 0.860 against a life-to-date 0.405 looks like demand consolidating, but on a session where one product did not report and nine printed zero, it mostly measures reporting coverage. The twenty-session issuer windows above are the more honest view of where bitcoin ETF inflows are actually coming from, and even those carry a 95% coverage flag on BlackRock.

One last caution on dates. Independent coverage corroborates bryptoflows' figures for the October 1 session — $102.7M category net with IBIT at $195.6M — but labels it October 2, a one-session difference that comes from the reporting lag rather than a disagreement in the data. Bitcoin ETF net flow this week will be published under at least two date conventions.

Flow and attribution figures throughout are read from bryptoflows' own read-only API, sourced from Farside Investors for ETF flows and SEC filings for Strategy, current through the October 2 close and the September 28 filing respectively. Bitcoin price figures are externally sourced, as noted. Track the live numbers on the ETF flow dashboard, the flow pulse overview and the institutional signal. Retrieved and published October 5, 2026.