Institutional Bitcoin Demand Score Falls to -14.9
Institutional bitcoin demand reads -14.9 on bryptoflows' composite institutional signal this morning, on data through Thursday, October 22. That is down from -4.9 when this site's macro column read the same composite on Wednesday, and from +28.9 a week before that. The honest answer to whether institutions are buying bitcoin right now is that they are buying and selling in close to equal size, and the sellers have been marginally ahead for a month — not that the bid has disappeared.
All eight components computed this run. Nothing was omitted and nothing was renormalised away, so this reading carries the composite's full intended weight rather than a partial one dressed up as a whole.
Are institutions buying bitcoin right now?
The composite runs from -100 to +100 and lands at -14.9, which the model labels neutral or mixed. Confidence reads 71.1%, today identical to the model's agreement figure because weight coverage and component reliability are both 1.0 with all eight reporting. Agreement is one minus the weighted spread of component scores around the composite, so this one means those readings sit roughly 29 points apart — moderate dispersion, not consensus.
Read it as bid strength, not a verdict on price. A -14.9 says the institutional channels this site can measure returned slightly more bitcoin to the market than they absorbed over the trailing twenty sessions, with the balance tipping negative only in the last few days. It sits well above the -60s and -70s that individual components are printing, because the weights spread the damage.
Why institutional bitcoin demand scores -14.9 and not lower
The composite is a weighted sum. Each component scores -100 to +100 on its own scale, then contributes its score times its weight. Here is the full set, with this run's readings:
- Supply absorption — score -68.0, weight 20%, contributing -13.6. The dominant drag, and effectively the whole reading.
- Flow persistence — score -24.5, weight 10%, contributing -2.4. One negative session, no more.
- Price confirmation — score -18.9, weight 6%, contributing -1.1. Bitcoin is down 3.4% over the twenty sessions.
- ETF flow momentum — score -2.5, weight 24%, contributing -0.6. The heaviest weight in the model, and almost perfectly neutral.
- Flow breadth — score 0.0, weight 14%, contributing 0.0. An exact four-four split among funds that traded.
- Treasury capacity — score 0.0, weight 6%, contributing 0.0. Gated to zero from a raw 94.8, for the reason below.
- Flow concentration — score +24.3, weight 6%, contributing +1.5. Activity spread across funds rather than dominated by one.
- Treasury accumulation — score +9.9, weight 14%, contributing +1.4. Strategy is still buying.
Two components sit at exactly zero and one is within three points of it. Only three of the eight pull meaningfully downward, and the largest weight is idle — which is why the composite is mildly negative rather than deeply so.
Supply absorption supplies 13.6 of the 14.9 points
The absorption component asks how much bitcoin the US spot ETF complex took off the market against what miners created. Over the twenty sessions from September 25 to October 22, the funds net released 1,457 BTC while issuance added 9,000 BTC at 450 a day — an absorption ratio of -0.16x.
The component scores distance from 1.0x, where demand exactly matches new supply, not distance from zero. A reading of -0.16x is therefore 1.16 below its neutral midpoint, which is how a single component reaches -68.0.
What that window conceals is how violently two-sided it was. Twelve of the twenty sessions were net positive, adding +22,995 BTC between them. The other eight removed -24,452 BTC. The net of -1,457 BTC is the residue of two large opposing flows, not an absent buyer. ETF holdings now stand at 681,395 BTC, against 682,852 twenty sessions ago.
This week on its own reads the other way: Monday through Thursday absorbed +5,147 BTC against 1,800 issued, a ratio of 2.86x. Tuesday's supply absorption column flagged that the twenty-session figure was heading below water on flat flows. It went further.
Thursday's -$635.6 million was a bottom-2% session
One session did most of the damage. On October 22 the category saw -$635.6 million of net outflows — a z-score of -1.77 and the 1.7th percentile of all 714 trading observations since January 2024. In bitcoin terms it was -7,807 BTC, a -17.35x session against issuance.
The internals were worse than the headline. Gross inflows across every US spot product totalled $2.3 million. Gross outflows were -$637.9 million:
- IBIT: -$254.1 million
- FBTC: -$217.5 million
- ARKB: -$109.4 million
- GBTC: -$29.5 million
- BITB: -$27.4 million
- MSBT: +$2.3 million, the only positive fund in the category
- Six further products: flat
A single $2.3 million ticket at Morgan Stanley's fund was the entire buy side of a session — a reading that moves flow persistence and price confirmation at once. The ETF flow page carries the detail.
The heaviest-weighted component is doing almost nothing
ETF flow momentum carries 24% of the composite — more than any other input — and contributed -0.6 points. Its twenty-session net flow is -$92.6 million, measured against a reference deviation of $3,119.9 million across 695 historical windows. On that scale, -$92.6 million is statistical noise.
This cuts both ways, and it matters more than any other reading here. The model's biggest lever is sitting at neutral, so the composite is not being held down by sustained ETF selling. Monday's flow column described a category that had stalled rather than reversed, and the twenty-session figure still says exactly that. But an idle lever is also an unspent one: momentum has 24% of the weight available to move in either direction, and almost nothing priced into it.
The shorter windows show the churn the twenty-session figure nets out. Five-session flows are +$356.7 million, ten-session +$231.8 million, three-session -$87.6 million, sixty-session +$5,978.9 million.
Why a four-four breadth split hides a one-fund tape
Flow breadth scored exactly 0.0 — over the five sessions from October 16 to 22, four funds were net positive, four net negative, and four flat. Counted that way it is a dead heat.
The dollar weighting is not close. IBIT took +$604.2 million across that window. The other three positive funds — Grayscale's BTC, EZBC and MSBT — managed +$17.8 million combined. One fund supplied 97% of the positive dollars, and the component cannot see that, because it counts funds rather than weighing them.
Flow concentration's +24.3 deserves the same scepticism. It measures the spread of gross activity, $2,094.9 million across twelve reporting products, with a Herfindahl index of 0.430 against an even-split baseline of 0.083. Low concentration scores positive on the reasoning that demand spread across many products is sturdier than one dominant fund. But on October 22 the funds making the tape look dispersed were mostly selling. Dispersed activity is not the same thing as dispersed demand, and this is the component where that distinction is lost.
Strategy keeps buying, but $44 billion of capacity scores zero
Treasury accumulation contributed +1.4, one of only two positive inputs. Across its last four filings Strategy acquired 2,516 BTC, lifting holdings to 848,517 BTC — growth of 0.30% of the position, against the 3% over four filings that the model treats as a strong accumulation quarter. All four were acquisitions; none was a sale. Thursday's treasury column covered the detail, including that the latest 267 BTC came almost entirely from selling new shares.
Treasury capacity is the more interesting zero. Strategy discloses $44.03 billion of undrawn financing against a bitcoin position worth $69.41 billion — a ratio of 0.63x that, taken alone, scores 94.8, the most bullish raw reading anywhere in the model. The composite counts it as zero.
Capacity is only a bid while drawing it adds bitcoin per share. With MSTR at a gross mNAV of 0.900, the equity trades below the value of the bitcoin behind it, so selling common hands new holders more bitcoin per dollar than existing holders own. The model scores that as a known nothing rather than a positive, and only restores full weight at 1.25x. mNAV last crossed parity on May 12, 2026 — this component has been switched off for roughly five months. The Strategy treasury page tracks it daily.
What this institutional demand indicator cannot see
The composite is a proxy built from disclosed flows, and its blind spots are real:
- It converts ETF dollar flows to bitcoin at each session's close: implied coins, not custodial counts.
- It sees US spot ETFs and one corporate treasury. OTC desks, offshore venues, direct custody mandates, futures basis and non-US vehicles are all outside it.
- Breadth counts funds without weighting them, so one fund can supply nearly all the demand and still read as balanced.
- Concentration reads gross activity, so active sellers improve the score.
- The mNAV gate applies to the whole $44.03 billion, though only $18.79 billion is the common ATM the dilution argument bites hardest on; the remaining $25.24 billion sits in preferred programmes.
- Strategy's share count is estimated, putting mNAV in a 0.897 to 0.911 band, and the net figure is omitted because senior claims are quarterly balance-sheet items a daily series would mostly interpolate.
- When a component cannot be computed it is dropped and the rest renormalised — never set to zero. That did not happen today, but it means readings from different runs are only strictly comparable when both are built on the same component set, and earlier published readings cannot be re-verified from the live endpoint.
How this compares with the past month
The trajectory is +28.9 in mid-October, -4.9 on Wednesday, -14.9 today — real deterioration, but concentrated in one component and one session rather than broad.
The longer windows are less dramatic. Month-to-date flows are -$175.6 million, sixty-session flows +$5,978.9 million, and cumulative net flows since launch $57,388 million — $5,347 million below the peak set earlier this cycle. The overview carries both series.
What would move institutional bitcoin demand next week
Specific, checkable thresholds:
- Momentum is the big lever. Holding the other seven components exactly where they are, the composite returns to zero if the twenty-session net flow reaches about +$2.58 billion, from -$92.6 million today. That is a $2.68 billion swing, or roughly $134 million per session sustained across the window. Nothing in the data says it is coming.
- Persistence is the cheap one. A four-session positive streak would lift that component from -24.5 to about +76 and add roughly 10 points to the composite on its own. Two consecutive positive sessions get about seven of them.
- The absorption drag gets worse before it gets better. On exactly flat flows from here, the twenty-session ratio deteriorates to about -0.69x over eight sessions, because the sessions rolling out of the window are mostly positive ones. October 7's -5,821 BTC does not drop out until the ninth session from here, around November 4. After that the window mechanically improves, all else equal.
- The capacity gate needs an 11% move. MSTR must rise 11.1% against bitcoin to reach parity and start opening the gate, and 38.9% to earn the component's full weight. At 1.125x it would add about 2.8 points.
- Another Thursday would hurt. A repeat of -7,807 BTC pushes absorption lower and flips persistence to a multi-session negative streak.
A bid-strength proxy, not a price forecast
The composite is a directional read on how hard institutions are bidding through the channels this site can observe. It is not investment advice, it does not forecast price, and there is no mechanical link between the score and what bitcoin does next — a -14.9 has no implied return attached to it. Its usefulness is that it decomposes: when the number moves you can see which component moved it, and whether that was one fund, one session, or a genuine change in the institutional bid.
Today it is one component and one session. The institutional signal dashboard updates the score and every component behind it as each session settles.