The Weekly Take: BlackRock Has Raised More Than Three Times Every Rival Combined

By Jason Jones, Editor in Chief

Spot Bitcoin ETFs took in $853.5 million last week and BlackRock’s IBIT accounted for 81% of it. Coverage treated that as the week’s caveat. Measured against the cumulative record, it is closer to the run rate.

Spot Bitcoin ETFs recorded net inflows on five consecutive trading days, Monday 3 August through Friday 7 August. The week totalled $853.5 million, the category’s strongest since mid-April.¹ BlackRock’s IBIT took $693.7 million of it, about 81 cents of every dollar.^1, 2 Spot Ether ETFs added $244.9 million over the same five sessions, a fifth straight positive week and their best since April.¹

Most coverage read the week as institutions returning. Several outlets noted the BlackRock share and treated it as the caveat — one fund doing the heavy lifting, a rebound narrower than the headline.²

That caveat is correct and it is filed in the wrong place. It is offered as an observation about an unusually lopsided week, which implies a normal week looks different. It doesn’t. Set the 81% against what this category has raised since launch and the week stops being lopsided. It becomes typical.

The cumulative record

Farside Investors, as of 12 August 2026, puts IBIT’s cumulative net inflow since launch at $61.157 billion and cumulative net inflows for the entire U.S. spot Bitcoin ETF category at $52.045 billion.³

Read carelessly, those two figures say something dramatic: IBIT has taken in roughly $9 billion more than the category has, so everything else combined must be net negative. That reading is arithmetically correct and substantively wrong, and it is worth walking through why, because the trap sits inside the two most widely quoted numbers in this market.

The gap is Grayscale. GBTC converted from a closed-end trust already holding tens of billions in Bitcoin, and has bled redemptions ever since on a 1.5% fee against competitors charging between 0.12% and 0.25%. Those redemptions count as category outflows. Farside puts GBTC’s cumulative net outflow since conversion at $27.513 billion.³

Put GBTC back, using Farside’s own figures throughout so the subtraction doesn’t mix trackers or dates:

$52.045bn category − $61.157bn IBIT + $27.513bn GBTC ≈ +$18.4 billion

The rest of the field, excluding IBIT and excluding GBTC, is net positive by roughly $18.4 billion since launch. Not negative. The apparent hole underneath the category is one legacy vehicle unwinding a position that predates the ETF structure entirely.

That figure is corroborated from outside Farside’s data. DWF Labs — a market maker rather than a research firm, and worth weighing accordingly — reported IBIT’s cumulative net inflow at $60.3 billion in early July and described it as 3.3 times every other fund excluding GBTC combined.⁴ Sixty point three divided by 3.3 is $18.3 billion, within $100 million of the figure derived above from a different tracker’s data five weeks later.

So the real finding is not that the competition is underwater. It is the ratio.

What 3.3x actually means

Roughly a dozen competing funds. More than two and a half years of trading. Fee structures deliberately undercutting BlackRock’s. And IBIT has taken in more than three times what all of them together have managed.

That is the context this week’s 81% belongs in. A fund running at a 3.3-to-1 cumulative advantage over its entire field taking 81% of a given week is not an anomaly requiring explanation. It is the structure asserting itself on a week when money happened to be moving. The genuinely newsworthy number would be a week where IBIT took 40%.

This also reframes what “institutional demand for Bitcoin ETFs” means as a measured quantity. Demand for the asset class and demand for one distribution channel are being reported as the same number. They may well be correlated, but they are not the same claim, and the flow headline cannot distinguish them.

Three totals, one week, all defensible

A second thing worth flagging, because it will keep happening.

Readers checking last week’s number will have encountered at least three: more than $750 million, $853.5 million, and about $1.1 billion. None is wrong.

The $750 million figure came from pieces filed on Friday 7 August, before the final session settled.⁵ The daily flows were $170.1 million, $211.5 million, $244.4 million, $128.8 million and $98.85 million.¹ The first four sum to $754.8 million — a real number, correctly reported, superseded within hours.

The $1.1 billion figure is Bitcoin plus Ether: $853.5 million and $244.9 million together.² Also correct, also a different question.

Three numbers, one week, mostly published without stating which basket or which window produced them. A reader has no way to reconcile them from the coverage alone. That is not a scandal, it is a routine labelling failure, and it costs nothing to fix: name the tracker, name the assets, name the window.

The same problem appears one layer down. Coverage described a five-day Bitcoin streak alongside a four-day Ether streak in the same week. The reason is that Ether funds posted an $11.9 million net redemption on Monday before turning positive Tuesday through Friday.¹ Without the daily series, the two streak lengths read as a contradiction. With it, they are two different windows that happen to share four days.

What actually moved the money

Two explanations are in circulation, and both are more interesting than the ones that led.

The first is macro. A weaker-than-expected U.S. jobs report during the week eased expectations of further Federal Reserve tightening, lifting risk assets generally. Bitcoin held around $64,000, Ether above $1,900.

The second is specific to crypto. Beginning 30 July, attackers drained self-custody wallets built on weakly generated seed phrases in what became known as the Coldcard exploit. TRM Labs put losses at roughly 1,816 BTC, about $116 million, across more than 5,200 addresses.⁶ Galaxy Research’s estimate moved as the theft was traced, from roughly $130 million on 4 August down to about $111 million by 8 August.⁷ Bloomberg Intelligence’s Eric Balchunas connected the ETF inflows to it directly: an ETF holder never touches a seed phrase, so the exploit is an argument for the wrapper.⁸

The Block named the weakness in that explanation, and it is a real one: Ether holders have no exposure to a Bitcoin-only hardware wallet vulnerability, yet Ether ETFs posted their own best week since April over the same five sessions.² If custody fear were driving flows, Bitcoin products should have pulled away from Ether products rather than moving with them.

We are not adjudicating this. We are pointing out that both explanations are testable against data that exists, and that the reflexive “institutions are back” framing is not.

What this does not show

The $18.4 billion is an upper bound, because of the Mini Trust. Grayscale runs a second vehicle, the Bitcoin Mini Trust (ticker BTC), spun off from GBTC in July 2024 at a 0.15% fee. Investors rotating from GBTC into it register as GBTC outflow and Mini Trust inflow. Our calculation removes the outflow side and keeps the inflow side, so some portion of the $18.4 billion is the same Grayscale client base moving to a cheaper share class rather than new money entering the category. The Mini Trust holds roughly $3.5 billion in assets, a large share of which came from the original spin-off rather than fresh subscriptions, so the contamination is likely on the order of one to two billion dollars — but we have not isolated it, and until we do, $18.4 billion is a ceiling rather than a measurement. Excluding both Grayscale vehicles would widen IBIT’s ratio, not narrow it.

We did not cross-check the trackers at a common date. SoSoValue’s cumulative figures as of 7 August ($61.175 billion for IBIT, $52.178 billion for the category) sit close to Farside’s as of 12 August, but five trading sessions separate the two readings and the inflow streak ended between them. Both Farside figures being lower is what intervening outflows would produce, so the proximity is consistent with tracker agreement without demonstrating it. We have not tested whether the two methodologies agree; we have observed that they are in the same neighbourhood. The independent check that does carry weight is DWF Labs’ ratio, which was derived from separate data. A third source would have helped: CoinGlass renders its flow table client-side and returned no populated figures when pulled directly, so it is absent from this reconciliation rather than contradicting it.

We have not built this from primary per-fund data. Every figure here is a tracker’s or a secondary report’s. TFTC publishes the full per-fund daily record since 11 January 2024 under an open licence,⁹ and rebuilding the cumulative series from it is the only way to remove tracker methodology as a variable and to isolate the Mini Trust properly. We intend to.

Even the fund count is unsettled. Sources put the number of U.S. spot Bitcoin ETFs at twelve or thirteen depending on whether recently launched products such as Morgan Stanley’s MSBT are included. We have used “roughly a dozen” rather than assert a figure we cannot verify — a small illustration of how little of this market’s basic accounting is standardised.

Flows are not holders. IBIT taking 81% of dollars says nothing about how many people those dollars came from. A wrapper with the widest wirehouse and RIA distribution can concentrate flow from a broad base. Concentration of vehicle and concentration of buyer are different questions and we measured only the first.

Flows are not price. We have not compared this week’s inflow against Bitcoin’s roughly $1.27 trillion market capitalisation or its daily spot turnover, which is the comparison that would say whether $853 million should be expected to move price at all. Without it, any claim about what these flows did to the market is unsupported, including the negative version.

One week is one week. Spot Bitcoin ETFs recorded heavy net outflows across the first half of 2026 and remain in the red year to date. Reported figures for the size of that deficit vary between roughly $4.5 billion and $5.4 billion depending on source and window, which we have not reconciled and are not relying on.

What would change our mind

• A per-fund daily series rebuilt from primary data showing the ex-IBIT, ex-GBTC total materially different from roughly $18.4 billion — or showing the Mini Trust’s contribution large enough to move the ratio meaningfully.
• Any sustained stretch, more than a single week, where IBIT’s share of category inflows falls below roughly half. That would be the first real evidence of a competitive second tier rather than a single-vehicle market.
• GBTC redemptions ceasing entirely. Its outflows have slowed considerably from their 2024 peak, and once that unwind is exhausted the category’s headline cumulative figure will jump without a dollar of new demand arriving. Anyone reading that jump as institutional conviction will be making the same category error in the opposite direction.
• Bitcoin ETF flow tracked against Ether ETF flow through the weeks after the Coldcard exploit. If custody flight is driving this, the two should diverge. If they keep moving together, it is macro.

Where this leaves us

The streak was real: five sessions, all positive, $853.5 million, correctly reported by everyone who reported it. What went unexamined was the denominator. One fund has raised more than three times what its entire field has raised put together, over the life of the product class, and the week’s 81% is that ratio showing up in a five-day window rather than departing from it.

The interesting question about this market is not whether institutions are coming back. It is whether anyone other than BlackRock is ever going to be more than a rounding error in the answer.

This is commentary and opinion — our read on the news, not financial advice or a recommendation to buy, sell, or hold any asset.

Sources

Figures pulled 13 August 2026 unless otherwise dated.

1. SoSoValue — U.S. spot Bitcoin ETF daily and weekly net flows, 3–7 August 2026; issuer-level breakdown; spot Ether ETF weekly and daily net flows for the same period. Cumulative figures as of 7 August 2026, cited in the limitations section only.
2. The Block, 8 August 2026 — combined Bitcoin and Ether ETF weekly total; IBIT share of category inflows; commentary on the cross-asset limits of the Coldcard explanation.
3. Farside Investors — cumulative net flow since inception for IBIT ($61.157 billion), GBTC (−$27.513 billion), and the U.S. spot Bitcoin ETF category ($52.045 billion), all as of 12 August 2026. All three inputs to the ex-GBTC calculation are taken from this single source and date.
4. DWF Labs, reported 2 July 2026 — IBIT cumulative net flows since launch ($60.3 billion) and stated ratio of 3.3x to all other funds excluding GBTC. DWF Labs is a market maker; the figure is cited as an independent data point, not as third-party research.
5. Contemporaneous flow reporting carrying the partial four-session total, filed 7 August 2026 before the week’s final session settled — cited as an example of window labelling, not as a source for the weekly figure.
6. TRM Labs — Coldcard exploit loss estimate, approximately 1,816 BTC (~$116 million) across more than 5,200 addresses, beginning 30 July 2026.
7. Galaxy Research — Coldcard exploit loss estimates, 4 August 2026 (~$130 million, provisional) and 8 August 2026 (~$111 million, revised).
8. Bloomberg Intelligence (Eric Balchunas) — commentary linking ETF inflows to the Coldcard exploit, 7 August 2026.
9. TFTC Bitcoin ETF Flow Tracker — per-fund daily flow record for all U.S. spot Bitcoin ETFs since 11 January 2024, published under CC BY 4.0.

Note on method: weekly flow figures are SoSoValue’s, measured Monday 3 August through Friday 7 August inclusive. The ex-GBTC cumulative calculation uses Farside Investors’ figures as of 12 August 2026 for all three inputs, so no dates or trackers are mixed within the subtraction. The resulting $18.4 billion is an upper bound: it excludes GBTC’s outflows while retaining Grayscale Mini Trust inflows, an asymmetry we have flagged but not yet quantified. Where sources disagree, we have said so rather than choosing.