IBIT ETF Absorbs 62% of the Money as GBTC Bleeds $65.5M in a Single Session

U.S. spot Bitcoin ETFs recorded $46.6 million in net outflows on September 8, ending a three-day inflow streak. Grayscale's GBTC lost $65.5 million while BlackRock's IBIT still took in $10.7 million.

That single session is the most informative datapoint in the current flow picture, and not because of its size. $46.6 million against a complex holding $103.34 billion in net assets is 0.045% — statistically nothing. What matters is the composition: the aggregate went negative while the dominant fund stayed positive, which means the headline number was set by Grayscale's structural redemptions rather than by any change in underlying demand.

The context is a September that started badly and then turned. Flows opened the month with $236.5 million of net outflows on September 1, reversed to $101 million of inflows on September 2, exploded to $730.8 million on September 3 — the strongest single session since January 14 — and added $174.6 million on September 4. Markets were closed Monday, September 7 for Labor Day. Then Tuesday's $46.6 million redemption.

Run the ledger and September stands at roughly $723.3 million of net inflows across five trading sessions.

Bitcoin trades at $79,390, up 1.20%, sitting 38.1% below its all-time high of $128,198.07 set October 6, 2025. IBIT closed Tuesday at $44.39, down $0.84 or 1.86%, and quoted $45.01 in pre-market Wednesday, up 1.40%.

The three-week window before September delivered approximately $3.8 billion — the strongest run of 2026 — with August totaling $3.52 billion and the final full week producing close to $1 billion.

The thesis running through this piece: the flow data shows genuine institutional accumulation into a 38% drawdown, and it also shows a market where one issuer determines the headline and one legacy vehicle determines whether that headline reads positive or negative. Both facts are true, and reading only one of them produces the wrong conclusion about what institutions are actually doing.

The September 16 Federal Reserve decision, with a 60% hike probability attached, is what the next three weeks of flow data will be reacting to.

September's Ledger: $723.3 Million Across Five Sessions

The month-to-date tape deserves laying out precisely, because the aggregate figure hides an enormously volatile sequence.

September 1: $236.5 million of net outflows. IBIT redeemed $201 million, FBTC lost $44 million, and Bitwise's BITB was the sole meaningful creation at $8 million.

September 2: $101 million of net inflows. The reversal came almost entirely through IBIT, with Grayscale offsetting through redemptions.

September 3: $730.8 million of net inflows — the strongest single session since January 14. IBIT absorbed roughly $454 million, ARK 21Shares' ARKB took approximately $137.7 million, and Fidelity's FBTC added $74.4 million. Several smaller funds saw inflows while VanEck and WisdomTree recorded modest outflows.

September 4: $174.6 million of net inflows, down 76% from the prior session.

September 7: markets closed for Labor Day.

September 8: $46.6 million of net outflows. GBTC shed $65.5 million while IBIT still took $10.7 million.

Net across the five sessions: approximately $723.3 million.

That figure is genuinely strong for a five-session window. Annualized at that pace it would represent roughly $36 billion of demand, well above what the complex has delivered in any recent year.

What it is not is steady. The range across those five days runs from -$236.5 million to +$730.8 million — a $967 million spread in a single week. A demand channel that swings by nearly $1 billion session to session is not a reliable signal about institutional conviction; it is a signal about a small number of large allocators making discrete decisions.

The September 3 print alone accounts for more than the entire month's net figure. Strip it out and the remaining four sessions total -$7.5 million. One day is the month.

That concentration is the reason the flow data should be read as a series rather than as a headline. A single allocation decision by one large holder can produce a week's worth of positive aggregate.

The historical seasonal context matters too. September has historically been an unlucky month for crypto markets and for Bitcoin specifically, and the flow data so far has run against that pattern.

The September 3 Print: $730.8 Million In One Day

The largest single-session inflow since January 14 deserves its own examination, because the composition tells you who was buying.

IBIT took roughly $454 million — 62% of the total. ARKB took approximately $137.7 million, or 19%. FBTC added $74.4 million, or 10%. Those three funds account for 91% of the day. Several smaller products saw inflows while VanEck and WisdomTree recorded modest redemptions.

The timing aligns with a specific macro moment. The surge coincided with expectations for a less aggressive Federal Reserve, following commentary from a Fed governor indicating support for holding rates steady if the August CPI report showed continued progress on underlying inflation. Rate-hike odds collapsed from around 70% toward 48%, the dollar fell, yields declined, and Bitcoin rallied 4% in a session to reclaim $80,000 for the first time since late August.

That is a macro-driven allocation rather than a Bitcoin-specific one. Institutions positioned for looser policy across multiple assets simultaneously, and the Bitcoin ETF was one expression of it.

The more interesting evidence is what happened next. Friday's payrolls report landed at 162,000 jobs against a 56,000 forecast, hike odds snapped back toward 60%, and flows still printed $174.6 million positive on September 4. Continued inflows after stronger employment data suggest the story is broader than a single rate bet.

Some buyers appear willing to accumulate Bitcoin while policy rates remain high, inflation risk remains unresolved, and BTC trades below its 2026 peak. That is arguably the most important signal in the current dataset: institutional demand no longer seems to require perfect conditions.

The flow also matters because of what it says about market structure. Spot ETF creations require the fund to buy actual Bitcoin. $730.8 million of net creations at roughly $80,000 per coin means approximately 9,100 BTC removed from tradeable supply in a single day.

That is spot demand rather than futures leverage, which is a materially different quality of buying. Leverage unwinds violently. Spot allocations held in regulated wrappers do not.

The caution is that the buying was concentrated in fewer funds than a broad institutional wave would produce.

IBIT At $44.39 And $61.44 Billion In Net Assets

BlackRock's iShares Bitcoin Trust closed Tuesday at $44.39, down $0.84 or 1.86%, against a prior close of $45.23. The session ranged $43.91 to $44.71 on volume of 40,633,536 shares against a 47,044,646 average. Net asset value stood at $45.15. Pre-market Wednesday the fund quoted $45.01, up $0.62 or 1.40%.

Net assets total $61.44 billion at a 0.25% expense ratio.

The 52-week range runs $32.84 to $71.82. At $44.39 the fund sits 38.2% below its annual high and 35.2% above its low — tracking Bitcoin's own 38.1% drawdown from $128,198.07 almost exactly, which confirms the wrapper is doing its job.

The performance grid is where the tension lives. IBIT is up 20.63% over one month and 15.00% over six months. It is down 10.59% year to date and 30.31% over twelve months, with a five-day change of -0.63%.

A product that took in $3.52 billion in August alone and roughly $723.3 million so far in September is showing its holders a 30% twelve-month loss.

That gap between flow and return is the defining feature of the Bitcoin ETF complex in 2026. Money is arriving at the fastest pace of the year into a vehicle that has destroyed capital over the trailing year. Either the new money is early into a bottom, or it is late money buying a bounce inside a downtrend that began in October 2025.

The dip-buying interpretation has evidence behind it. Bitcoin had fallen from its 2026 highs and was trading in the upper-$70,000 range when much of the money entered, giving investors exposure at substantially lower prices than at the market's earlier peak.

The volume detail is neutral. 40.63 million shares against a 47.04 million average is 86% of normal turnover — a quiet session rather than distribution.

For scale against the peer complex: IBIT's $61.44 billion of net assets represents roughly 59% of the $103.34 billion held across all U.S. spot Bitcoin ETFs.

The Concentration Problem: One Fund Is The Market

Run the per-fund numbers across recent sessions and a single conclusion emerges: IBIT is not the largest participant in this market. It is effectively the market.

September 3: IBIT $454 million of a $730.8 million total — 62%.
September 1: IBIT -$201 million of a -$236.5 million total — 85% of the redemption.
August 31: IBIT +$206 million of a +$216.7 million total — 95%.
August 27: IBIT +$278 million against a +$242.2 million total — more than 100%, with FBTC's -$84 million offsetting.
August 26: IBIT +$201 million of +$232.1 million — 87%.
August 25: IBIT +$284 million of +$314.4 million — 90%.
August 20: IBIT +$503 million of +$606.3 million — 83%.
August 21: IBIT +$239 million of +$307.5 million — 78%.

Across eight sessions, IBIT accounts for between 62% and 100% of the aggregate flow every single time.

That structure has two consequences that matter for anyone using flow data as a demand signal.

The first is that the "institutional demand" headline is really a statement about one issuer's client base. When BlackRock's advisory platforms and model portfolios allocate, flows are positive. When they rebalance, flows are negative. The other eleven funds are noise around that decision.

The second is concentration risk in the other direction. A single-day $201 million redemption from IBIT on September 1 dragged the entire complex negative. When that base turns, it turns all at once, and there is no offsetting issuer large enough to absorb it.

The exception that proves the pattern is August 28, when the complex printed -$201.8 million with ARKB leading redemptions at -$115 million, BITB at -$50 million and IBIT at only -$33 million. That was a genuine broad-based redemption day rather than a BlackRock event, and it is the rarest configuration in the dataset.

The smaller funds show up sporadically and in size. MINIBTC took $47 million on August 26 and $16 million on August 24. HODL moved between +$4 million and -$13 million. BTCO appeared once at +$4 million.

Twelve funds exist. Three matter. One decides.

GBTC's Structural Bleed

Grayscale's GBTC is the reason the aggregate flow number is systematically understated, and separating it out changes how the data reads.

On September 8, GBTC shed $65.5 million while the complex printed -$46.6 million. Without Grayscale, the day would have been positive by roughly $19 million. On August 27, GBTC lost $27 million against a +$242.2 million aggregate. On August 26, it lost $50 million against +$232.1 million. Across the recent period, GBTC recorded around $48 million of net outflows while newer products accumulated.

The mechanism is fee arbitrage rather than sentiment. GBTC converted from a closed-end trust into an ETF carrying a substantially higher expense ratio than the products launched alongside it. IBIT charges 0.25%. Holders sitting in a higher-fee vehicle with identical underlying exposure have a straightforward incentive to rotate, and they have been doing so continuously since conversion.

That flow is mechanical. It continues regardless of what Bitcoin does, regardless of what the Fed does, and regardless of whether the holder is bullish or bearish. A GBTC redemption is frequently an IBIT creation on the same day by the same beneficial owner.

The analytical consequence is that headline aggregate flow understates net new demand by whatever GBTC is bleeding on a given session. On September 8 that was the entire difference between a negative headline and a positive one.

The practical adjustment is to read the complex as two separate datasets: legacy conversion outflows on one side, and new-money creations across IBIT, FBTC, ARKB, BITB, MINIBTC and the rest on the other.

Under that framing, September 8 was not a demand reversal. It was a fee rotation session with IBIT still taking $10.7 million.

The caveat is that GBTC's bleed is finite. As assets migrate out, the drag on the headline shrinks, and future aggregate numbers become cleaner reflections of genuine demand. That transition has been running for over two years and is well advanced.

Everything else in the complex carries expense ratios in the same neighborhood as IBIT, which means the rotation trade is nearly exhausted.

August Delivered $3.52 Billion And A Three-Week Run

August was the strongest month of 2026 for Bitcoin ETF flows, and the daily record shows why.

The month totaled approximately $3.52 billion of net inflows. The three-week window spanning late August into early September delivered roughly $3.8 billion — described as the strongest three-week run of the year. The final full week before September produced approximately $987 million to $1.03 billion depending on the reporting window.

The session-by-session tape for the strongest stretch:

August 19: +$517.2 million. IBIT $285 million, ARKB $78 million, FBTC $62 million.
August 20: +$606.3 million. IBIT $503 million, FBTC $65 million, BITB $26 million, ARKB $12 million, BTCO $4 million.
August 21: +$307.5 million. IBIT $239 million, FBTC $30 million, MINIBTC $14 million, BITB $9 million, HODL $4 million.
August 24: +$337.6 million. IBIT $209 million, FBTC $105 million, MINIBTC $16 million.
August 25: +$314.4 million. IBIT $284 million, FBTC $15 million.
August 26: +$232.1 million. IBIT $201 million, MINIBTC $47 million, FBTC $26 million, GBTC -$50 million.
August 27: +$242.2 million. IBIT $278 million, ARKB $30 million, BITB $22 million, FBTC -$84 million, GBTC -$27 million.
August 28: -$201.8 million. ARKB -$115 million, BITB -$50 million, IBIT -$33 million, HODL -$13 million.
August 31: +$216.7 million. IBIT $206 million, MINIBTC $9 million, FBTC $7 million, HODL -$13 million.

Nine sessions, eight positive, one negative, netting roughly $2.57 billion.

Two features stand out. The August 20 print at $606.3 million was the second-largest of the period and IBIT took 83% of it. And August 28 broke the streak with a broad-based redemption led by ARKB rather than by BlackRock — the only session in the run where the concentration pattern inverted.

That August 28 session is a useful reminder that the smaller issuers can move the aggregate when they act together.

Volume Fell To $14.5 Billion — What That Means

Total Bitcoin ETF trading volume declined to approximately $14.5 billion from nearly $19 billion the previous week — a 24% drop.

That detail cuts against the bullish reading of the flow data and it deserves careful handling, because volume and flow measure different things.

Flow is net creations minus redemptions: the change in how much Bitcoin the funds hold. Volume is total shares traded on exchange: how much activity happens between existing holders without touching the underlying.

Rising net inflows alongside falling gross volume can mean two opposite things. It can indicate persistent quiet accumulation — a small number of large buyers stepping in without generating secondary-market churn, which is exactly what patient institutional allocation looks like. Or it can mean the market lacks broad participation and immediate momentum, with a handful of creations moving the net number while general interest evaporates.

The September 4 print supports the second reading. Flows fell 76% from $730.8 million to $174.6 million in one session, and the following trading day went negative. Momentum did not sustain.

The counterargument sits in the composition. The buying that did occur was concentrated in fewer funds — which is what happens when a small number of large allocators act rather than when retail participation broadens. That concentration is a quality signal about the buyer even as it is a weakness signal about breadth.

Applied to the current price level, the volume decline matters because Bitcoin is trading in a compressed range. The asset sits at $79,390 having been unable to hold above $80,966 on two separate attempts. Breaking a supply shelf that has rejected price twice requires participation, and $14.5 billion of weekly ETF volume against $19 billion the prior week is participation shrinking.

The historical pattern for this configuration is a range that persists until a catalyst forces resolution, rather than a breakout that builds gradually.

Friday's CPI print is the catalyst on the calendar.

$103.34 Billion Of Assets, 6.32% Of Bitcoin

Combined net assets across U.S. spot Bitcoin ETFs total $103.34 billion, equal to approximately 6.32% of Bitcoin's total market capitalization.

That percentage is the number worth holding onto, because it defines how much influence this flow data actually has.

Six percent is meaningful and it is not dominant. The marginal price of Bitcoin is still set on exchange by holders who own coins directly, not by fund creations. A $730.8 million inflow day represents roughly 0.045% of Bitcoin's market capitalization — real buying, but not enough to move a $1.25 trillion-plus asset by itself.

What the ETF complex does provide is a demand floor and a distribution channel. Coins acquired through creations are held in institutional custody and do not return to exchange order books on a daily basis. Every net creation permanently reduces tradeable float until a redemption reverses it.

The cumulative record frames the scale of what has been built. Net flow since launch stands at $51.8 billion across the twelve U.S. spot Bitcoin ETFs. That is two and a half years of accumulation that did not exist before January 2024.

IBIT's $61.44 billion of net assets against the $103.34 billion complex total means BlackRock holds 59% of the industry's assets. Net assets exceed cumulative net flow at the complex level because appreciation on early-acquired coins is embedded in the asset figure.

The comparison against other digital assets puts it in perspective. Ethereum's spot ETF complex is a fraction of this size, with its largest fund holding $8.76 billion. XRP's seven funds hold roughly $2 billion combined with $1.68 billion of cumulative inflows. Solana's complex has attracted $1.35 billion cumulatively.

Bitcoin's ETF infrastructure is roughly seven times Ethereum's and fifty times Solana's, against market capitalization ratios of five-to-one and twenty-to-one respectively. Institutional allocators overweight Bitcoin relative to its share of the digital asset market, consistently, across every measurement window.

That preference is structural and it has not shown any sign of shifting.

The Base Rate: 54% Of 2026 Sessions Have Been Negative

The single most useful context for interpreting any individual flow print is the historical base rate, and it is not what most commentary assumes.

Across the 666 trading sessions from January 11, 2024 through August 14, 2026, net flows were negative on 266 of them — roughly 40% of all sessions.

The share has shifted materially each year. Negative sessions accounted for 31% of trading days in 2024, 40% in 2025, and 54% so far in 2026.

More than half of 2026's sessions have printed net outflows. That reframes the current data completely: a $46.6 million redemption on September 8 is not an anomaly requiring explanation. It is the modal outcome for this year.

And yet cumulative net flow remains positive at $51.8 billion. The complex has grown despite the majority of individual sessions being negative, because the positive days are larger than the negative ones. The $730.8 million on September 3 and the $606.3 million on August 20 outweigh dozens of small redemption days.

That asymmetry is the actual signature of institutional accumulation: infrequent large buying punctuating persistent small selling.

The streak records provide the tail context. The longest outflow streak ran 13 sessions from May 15 to June 3, 2026, shedding $4.37 billion. The longest inflow streak ran 19 sessions from May 13 to June 7, 2024.

The May-June 2026 episode is the relevant comparison for anyone worried about the current data. Thirteen consecutive negative sessions and $4.37 billion of redemptions happened four months ago, and the complex has since recovered to add $3.52 billion in a single month.

A three-day inflow streak ending is not comparable to a thirteen-session, $4.37 billion drawdown.

The rising share of negative sessions in 2026 — 54% against 31% in 2024 — does indicate maturation. The novelty premium that drove near-continuous creations in the first year has gone. What remains is a normal two-way institutional market where allocators both buy and rebalance.

Dip-Buying Or Catch-Up? Reading Intent From Flow

The flow data cannot identify who is buying, and that limitation should be stated plainly before drawing conclusions from it.

Spot Bitcoin ETFs are used by hedge funds, registered investment advisers, family offices, retirement accounts, institutions and ordinary brokerage customers. Daily flow data does not identify the ultimate buyer. The word "institutional" gets attached to every ETF inflow headline and it is an assumption rather than a disclosure.

What can be inferred is scale and concentration. A $454 million single-day creation in one fund is not retail. The size and the concentration of recent inflows point toward meaningful professional allocation.

Two competing readings have been offered for the recent buying.

The first is dip-buying. Bitcoin had fallen from its 2026 highs and was trading in the upper-$70,000 range when much of the money entered, giving investors exposure at substantially lower prices than during the earlier peak. That is the straightforward interpretation and it fits the price context: an asset 38.1% below its all-time high with a functioning institutional wrapper is exactly what a long-horizon allocator buys.

The second is a catch-up trade. The activity has been described as renewed institutional exposure after Bitcoin had lagged some other risk assets. Under that reading, allocators who were underweight are rebuilding position sizes rather than expressing a fresh view.

The strongest evidence for genuine conviction is not that the funds received money. It is that they continued receiving money after the macro case weakened. The September 3 surge coincided with expectations for a less aggressive Fed. Then payrolls came in at 162,000 against a 56,000 forecast, hike odds ran back to 60%, and September 4 still printed $174.6 million positive.

Buyers willing to accumulate while policy rates remain high, inflation risk remains unresolved, and Bitcoin trades below its 2026 peak represent a different quality of demand from buyers positioning for a rate cut.

Institutional demand does not appear to require perfect conditions anymore. That is the most consequential shift in this dataset.

Macro: CPI Friday, Fed Tuesday, And The Flow Function

Bitcoin ETF flows have become a derivative of the Federal Reserve probability, and the relationship has been close to mechanical for three weeks.

Futures price a 60% chance the Fed raises the funds rate 25 basis points from 3.75% at the September 15-16 meeting. The 10-year Treasury yields 4.8130% at a three-year high. Brent crude cleared $100.865, up 3.01%, feeding the inflation expectation that drives the hike odds.

Trace the sequence. Hike odds near 70% in late August produced strong but uneven flows. Dovish Fed commentary collapsed odds toward 48% and produced $730.8 million on September 3. Payrolls at 162,000 sent odds back toward 60% and flows fell 76% the next session and turned negative the one after.

That is a near one-to-one mapping between rate expectations and creation activity.

The August consumer price index releases Friday, September 11 at 8:30 a.m. ET, with headline expected to hold at 3.40% and core forecast at 2.4%. Producer prices land Thursday with headline forecast at 5.3%. The FOMC decision follows September 16.

The transmission runs through real yields. Bitcoin pays no coupon, and the ETF wrapper adds a 0.25% expense ratio on top of that. At a 4.8130% 10-year with inflation at 3.40%, the real cost of holding a zero-yield asset in a fee-bearing wrapper is at its cycle high.

The offsetting force is fiscal. Total federal debt has passed $40 trillion, the annual deficit runs above $2 trillion, and the Treasury is running buyback operations of at least $4 billion in 10- and 20-year notes to cap long yields while separately buying yen to keep Japan from liquidating its $1.1 trillion Treasury position.

Gold at $4,403.01 is up 1.08%. Silver at $66.168 is up 0.63%. Bitcoin at $79,390 is up 1.20%. Three non-yielding assets bid on a day when the S&P 500 is down 0.38%.

That is the debasement trade, and it is what the ETF flows are increasingly expressing.

What The Flow Data Says About Price

Connecting the flow record to the price chart produces a clearer picture than either does alone.

Bitcoin trades at $79,390, having failed twice at the $80,100 to $80,966 supply band — once on August 27-28 and again on September 3, when it printed $80,966 and settled near $80,680 on the largest volume bar since August 25.

Overlay the flow data on those two attempts. The August 27-28 push came with $242.2 million of inflows followed immediately by a $201.8 million redemption. The September 3 push came with $730.8 million followed by $174.6 million and then -$46.6 million.

Both rallies into the same band were met by flow reversals within one to three sessions. That is not coincidence. It suggests a cohort of holders using ETF redemptions to take profit at $80,000, which is precisely the mechanism that creates a supply shelf.

The downside picture is the mirror image. Bitcoin has held above $77,165 through the entire period, including the September 1 session when $236.5 million left the funds. Redemptions of that size failing to break the floor indicates spot demand outside the ETF complex absorbing the supply.

That combination — ETF selling capping rallies at $80,966, non-ETF demand defending $77,165 — describes the range precisely and explains why it has held for two weeks.

The resolution requires one of two things. Either ETF flows have to sustain above the $400 million daily level for several consecutive sessions to overwhelm the $80,000 supply, or a macro catalyst has to move the whole complex at once.

The three-week $3.8 billion run averaged roughly $250 million per session. That was enough to lift Bitcoin from the low $70,000s to $80,966 and not enough to hold it there.

At $103.34 billion of assets representing 6.32% of Bitcoin's market capitalization, the ETF complex is now large enough to be the marginal buyer at the margin and too small to set price on its own.

Bitcoin ETF Flows Forecast: Scenarios And Probabilities

The executable framework for the next three weeks.

Base case at 47% probability: flows stay volatile and net modestly positive, running between -$100 million and +$300 million on most sessions with occasional outliers in either direction. GBTC continues bleeding roughly $30 million to $60 million per session, keeping headline aggregates suppressed. IBIT continues taking 60% to 90% of net creations. September finishes with $800 million to $1.5 billion of net inflows — below August's $3.52 billion but comfortably positive. Bitcoin holds $77,165 to $82,178.

Bull case at 31% probability: core CPI prints at or below 2.4% Friday, hike odds collapse toward 48%, and flows repeat the September 3 pattern with multiple sessions above $400 million. A three-to-five session streak at that level would represent $1.5 billion to $2 billion of concentrated buying, enough to clear the $80,966 shelf. September totals $2.5 billion or more, and Bitcoin tests $82,178 and then the $85,000 to $86,000 supply zone.

Bear case at 22% probability: producer prices accelerate Thursday, core CPI runs 2.7% or above Friday, the September 16 hike becomes near-certain, and the complex enters a sustained outflow sequence. The precedent exists and is recent: 13 consecutive negative sessions from May 15 to June 3, 2026, shedding $4.37 billion. A repeat at half that scale would take Bitcoin through $77,165 toward the $75,000 to $76,500 demand zone.

The distribution skews positive because the base rate matters. Cumulative net flow stands at $51.8 billion positive despite 40% of all sessions being negative and 54% of 2026 sessions being negative. The positive days are systematically larger than the negative ones, and that asymmetry has held across two and a half years.

The metric to watch is not the daily headline. It is whether IBIT prints two or more consecutive creation days above $200 million. That configuration preceded both the August 19-27 run and the September 3 spike.

Verdict: Real Demand, One Buyer, And Friday Decides

The Bitcoin ETF complex has taken roughly $723.3 million across five September sessions and $3.52 billion in August, and it recorded $46.6 million of net outflows on September 8.

The constructive case is documented. September 3 delivered $730.8 million, the strongest single session since January 14, with IBIT absorbing $454 million, ARKB $137.7 million and FBTC $74.4 million. The three-week run into early September totaled approximately $3.8 billion, the strongest of 2026. Cumulative net flow since launch stands at $51.8 billion positive across twelve funds holding $103.34 billion, equal to 6.32% of Bitcoin's market capitalization. Buying continued after the payrolls report pushed hike odds back toward 60%, indicating demand that no longer requires a dovish Fed. And the money arrived while Bitcoin traded 38.1% below its $128,198.07 October 2025 high, which is what accumulation into weakness looks like.

The cautious case is equally documented. IBIT accounts for between 62% and 100% of aggregate flow on nearly every session, which means the entire institutional demand signal for a $1.25 trillion-plus asset reflects one issuer's client base — and that base redeemed $201 million in a single day on September 1. ETF trading volume fell 24% to $14.5 billion from nearly $19 billion. Strip the September 3 print and the remaining four sessions of the month net -$7.5 million. Negative sessions now account for 54% of 2026 trading days against 31% in 2024. GBTC's structural bleed continues at $48 million to $65 million per session. And a 13-session, $4.37 billion outflow streak happened four months ago, in May and June of this year.

The verdict is real demand routed through a single channel. The flow data shows genuine professional allocation into a drawdown, arriving in large discrete blocks rather than in a steady stream, and concentrated to a degree that makes any single session a poor signal.

Watch for two consecutive IBIT creation days above $200 million — that is the configuration that preceded both of this cycle's rallies. Watch $77,165 on the underlying, which has held through every redemption day including the $236.5 million on September 1.

Friday's 8:30 a.m. CPI print is what the next three weeks of flow data will be reacting to.

That's TradingNEWS