Institutional Bitcoin Demand Recovers 11.5 Points to -3.4
Institutional bitcoin demand reads -3.4 on bryptoflows' composite institutional signal this morning, on data through Thursday, October 29. That is up 11.5 points from the -14.9 this column reported last Friday. The honest answer to whether institutions are buying bitcoin right now is that they restarted about two weeks ago and have not stopped — the last ten sessions took in +$659.6 million — but the model still declines to call a direction, because a single component is cancelling the whole of it.
All eight components computed this run. The missing-components list is empty and weight coverage is 1.00, so nothing was omitted and no weights were renormalised. This is the composite at its full intended construction, and therefore directly comparable to the -14.9, +28.9, +17.4 and +42.4 readings of the previous four Fridays.
Are institutions buying bitcoin right now?
Yes, in the ETF channel, and for four sessions running. No, in the only measure that compares that buying against the supply arriving to meet it.
Both statements come out of the same dataset. The model classifies -3.4 as neutral or mixed, the band it reserves for readings between -20 and +20 where it refuses to name a direction, with a confidence of 59.9 and component agreement of 0.60 — the eight inputs are only modestly of one mind. Four components are positive and four negative. The positives sum to +12.9 points and the negatives to -16.3, and the difference between them is the entire headline.
That is a different reading from a flat one. A composite at -3.4 because every component is quiet would mean no institutional bid at all; this one is at -3.4 because a real bid is outweighed by a single measurement.
What moved institutional bitcoin demand 11.5 points higher
The recovery is almost entirely the flow components catching a streak.
- Flow persistence scored +76.2 of a possible 100, contributing +7.6 points on a 0.10 weight. US spot bitcoin ETFs have now taken in money four sessions in a row.
- Flow breadth scored +33.3, contributing +4.7 on a 0.14 weight. On the October 29 session four of twelve products took in money, two lost money, and six reported none at all.
- Treasury accumulation scored +3.3, contributing +0.5 on a 0.14 weight.
- ETF flow momentum scored +0.7, contributing +0.2 on a 0.24 weight — the model's heaviest weight, producing almost nothing.
Against that, supply absorption at -60.8 took -12.2 points off a 0.20 weight, flow concentration at -36.0 took -2.2, and price confirmation at -32.9 took -2.0 on a 20-session bitcoin return of -6.16%. Treasury capacity scored exactly 0.0, for a reason worth its own section below.
One hurdle this column set has also been cleared. On Monday, this site's ETF flow piece put October $313.5 million underwater and calculated that the category needed better than +$313.5 million in the month's final five sessions to finish positive. It has delivered +$440.8 million week-to-date, putting October at +$127.3 million. Today's session, the month's last, is not yet in the data.
Absorption is the whole deficit: 5.9 hours of mined supply
Strip supply absorption out and this composite reads +8.8. It is not a rounding influence; it is the reading.
Over the twenty sessions through October 29, US spot bitcoin ETFs absorbed a net 110.8 BTC while miners created 9,000 BTC at the fixed 450 BTC per day issuance rate. That is an absorption ratio of 0.012x — the institutional ETF channel took down 1.2% of new supply across a month of trading, or in time rather than percentage, roughly 5.9 hours of global bitcoin issuance in twenty trading days.
That is a weak number and also a far better one than Monday's, when the supply absorption column reported the same window at -0.42x, with the funds handing back 3,750 BTC. The channel has gone from net supplier to barely-net-buyer in four sessions. The component scores it harshly anyway, because a ratio near zero sits far below its own history — and on that it is right. Absorbing 1.2% of issuance is not accumulation; it is the absence of distribution.
It is also the part most likely to be misread as bearish positioning — not a statement about selling, but that for a month the largest institutional access channel has not meaningfully competed with miners for supply.
Why the heaviest weight still reads zero on a positive month
The most interesting thing in this run is a measurement artefact, and it cuts in the market's favour.
ETF flow momentum carries the model's largest weight, 0.24, and reads off net flow over a trailing twenty sessions. That window sums to +$24.6 million against a reference deviation of $3,112.3 million — statistically zero, which is why the heaviest component contributed +0.2 points.
But the window is not uniformly flat. Split it in half:
- The older ten sessions, October 2 to October 15, net -$635.0 million.
- The recent ten sessions, October 16 to October 29, net +$659.6 million.
The window reads flat because those two halves very nearly cancel. And the recent half is the stronger of the two despite containing the worst session in the entire window — October 22's -$635.6 million, a bottom-2% day by this site's own distribution. The last ten sessions absorbed a bottom-2% outflow and still netted positive.
The roll-off that improves this score without new buying
That split has a forward-looking consequence that does not depend on anyone buying anything.
The October 2 to October 15 stretch — which includes October 7 at -$484.9 million, October 14 at -$261.4 million and October 8 at -$244.1 million — leaves the trailing window over the next ten sessions. If flows from here were exactly zero, the twenty-session input would converge on the recent half's +$659.6 million, roughly 27 times the +$24.6 million it reads today. The same mechanism works on supply absorption, which uses the same twenty-session window.
So the heaviest-weighted and second-heaviest-weighted components in this model are both sitting on an input that improves mechanically as October's first half ages out, provided flows merely avoid getting worse. That is the most important thing about this reading, and it is also the clearest limit on it: a score that can rise on arithmetic alone is a score that was understating the present, not one that is predicting the future.
The caution is real too. Shorter windows are already cooling: net flow over the last three sessions is +$330.5 million and over five is +$302.9 million, so the five-session window is now smaller than the three-session one. Flow acceleration, a five-day average, reads -10.8, and Thursday's session was +$17.2 million, a 45th-percentile day. The streak is four sessions long, but thinning.
Breadth and persistence: four funds, one of them real
Flow concentration is the component that keeps the positives honest. Over the five sessions from October 23 to October 29, the concentration measure reads 0.706 against 0.083 for an even split across twelve products — about 8.5 times more concentrated than evenly distributed, or an effective field of roughly 1.4 funds doing $625.3 million of gross activity.
That is the qualification on both flow positives. The four-session streak that scored +76.2, and the four-of-twelve breadth that scored +33.3, are substantially one product: BlackRock's IBIT also carries a four-session positive streak, against one-session streaks at Fidelity's FBTC and Morgan Stanley's MSBT and two at Bitwise's BITB. Since inception IBIT has taken in $66.54 billion while Grayscale's GBTC has shed $28.00 billion. The dispersion is not new, but it means "institutional bitcoin buying" this week is closer to one allocator's book than a broad reallocation. Per-fund detail sits on the ETF flow dashboard.
The mNAV gate that costs the composite 5.7 points
Treasury capacity scored 0.0 this run, and the zero is deliberate rather than missing — an omitted component would have been renormalised away instead, not scored.
Strategy has roughly $44.03 billion of estimated financing capacity against a bitcoin position marked at $66.05 billion. On capacity alone the component would score 95.7, worth +5.7 points at its 0.06 weight, which would put this composite at +2.4 and positive. It scores zero because the model gates capacity on mNAV, and mNAV reads 0.861: with 423,834,222 shares at $134.19, the equity is capitalised at about $56.87 billion against that $66.05 billion of bitcoin. Issuing shares to buy bitcoin at a discount to the bitcoin you already hold destroys value for existing holders, so the model refuses to count capacity it does not believe will be used. Share-count parity with the treasury mark would need roughly $155.83, about 16.1% above the live mark the composite used this morning.
On the accumulation side, the four filings in the window total 851 BTC — 334 on October 5, 250 on October 13, 267 on October 19, and none on October 26 — leaving holdings at 848,517 BTC, a 0.10% growth contribution. No filing in this window was a sale. Thursday's treasury column covers that zero-purchase week and the preferred buyback behind it; the full position is on the MSTR treasury page.
What would change this reading next week
- Supply absorption turning positive in earnest. To lift the twenty-session ratio to 1.0x the ETF channel would have to absorb about 8,889 BTC more than it has, roughly $692 million at the $77,838.50 mark on the composite. Nothing in the current pace reaches that. A move from 0.012x toward 0.3x is the realistic version, and would still be worth several points on a 0.20 weight.
- The window roll-off landing. The largest single swing available is October's first half leaving the trailing twenty sessions, which lifts both heavy components without new inflows.
- A fifth, sixth and seventh session of inflows. Persistence is already scoring +76.2 and has limited room above that, but breaking the streak would cost most of its 7.6 points quickly.
- Breadth widening past one book. Four of twelve with a 0.706 concentration reading is a narrow bid. The same net total spread across six or eight funds would raise breadth and relieve concentration at once.
- Price confirmation stabilising. The twenty-session return is -6.16%. The Federal Reserve held at 3.75%–4.00% on October 28 with two officials dissenting for a hike, and CME FedWatch pricing cited after the decision put December hike odds at about 82.6% (CNBC, CNN). A tightening path is not a flow input, but it reaches this composite through price.
- An mNAV recovery above parity. That single gate is the difference between -3.4 and +2.4.
What this bitcoin demand indicator cannot see
The composite is a directional proxy for institutional bid strength. It is not a price forecast, it implies no mechanical link to future returns, and it is not investment advice.
Its blind spots are specific. It sees US spot ETF flows and Strategy's SEC filings; it does not see separately managed accounts, offshore vehicles, sovereign or pension allocations made outside those channels, structured notes, or bitcoin pledged as collateral. That last gap is topical: a RootstockLabs report published October 29 estimates about 1.5% of bitcoin supply is already deployed in lending and bitcoin-based finance and puts identifiable institutional holdings near 3.9 million BTC (Digital Today) — those are the report's estimates, not this site's measurements, and essentially none of it is visible to the eight components above.
It cannot distinguish a dollar of ETF inflow from a dollar of basis trade, which matters when concentration reads 1.4 effective funds. And its treasury components describe one company, so Strategy stands in for all corporate treasury demand — harmless while Strategy dominates the category, and not forever.
Every component's data is current: BTC ETF flows and prices through October 29 with no missing sessions, Strategy filings through October 26. The composite, component breakdown and live score sit on the institutional signal dashboard; the broader picture is on the overview. Last Friday's reading, at -14.9 with the same eight components, is here.