Institutional Bitcoin Demand Falls 28.6 Points to -32.0
Institutional bitcoin demand reads -32.0 on bryptoflows' composite institutional signal this morning, on data through Thursday, November 5. That is 28.6 points below the -3.4 this column reported last Friday, and it is the first reading the model classifies as moderate distribution rather than neutral. The short answer to whether institutions are buying bitcoin right now is that they are not — the ETF channel has been a net seller across both windows the composite reads, and for the first time in the ten Fridays this column has published, the model is willing to put a negative direction on it.
One thing this reading is not is a partial one. All eight of eight components computed this run, the missing-component list is empty, and available weight coverage is 1.0, so nothing has been dropped and no weight renormalised. The score is the full composite on methodology institutional-support-v1.1, with a trailing window of 20 sessions.
Are institutions buying bitcoin right now?
No, on this measure, and the model's own confidence in saying so is middling rather than high.
The composite classifies -31.97 as moderate distribution, the band for readings at or below -20 but above the -60 it would call strong distribution. Confidence is 62.4 and component agreement 0.624 — the inputs are more of one mind than last week, when four were positive and four negative, but still not unanimous. Five components are negative, two are positive, and one reads exactly zero. The positives sum to +3.4 points and the negatives to -35.4.
That composition matters: a score of -32 with all eight components leaning the same way would be a stronger statement than this one, where -35.4 of negative weight meets almost nothing pushing back.
What took institutional bitcoin demand down 28.6 points
The decline decomposes exactly, and the largest movers are not the largest weights. Each component's contribution to the score, against the same component last Friday:
- Flow breadth went from +33.3 to -40.0, a swing of -10.3 points on a 0.14 weight. The largest single contributor to the fall.
- Flow persistence went from +76.2 to -24.5, a swing of -10.1 points on a 0.10 weight.
- Supply absorption went from -60.8 to -94.4, a swing of -6.7 points on a 0.20 weight.
- ETF flow momentum went from +0.7 to -25.6, a swing of -6.4 points on the model's heaviest 0.24 weight.
- Flow concentration went from -36.0 to +44.3, a swing of +4.9 points — the only component that improved.
- Treasury accumulation added +0.2, price confirmation cost -0.3, and treasury capacity was unchanged at zero.
Those swings sum to -28.6, which is the whole move from -3.4 to -32.0.
The honest reading of that table is that last week's recovery was mostly the two fastest-twitch components, and they have now given it all back and more. Flow persistence is the clearest case: it scored +76.2 last Friday because US spot bitcoin ETFs had taken in money four sessions running. It now reads -24.5 on a negative streak of one session. A single session's sign flipped a tenth of the model's weight by a hundred points. That is the component behaving as designed, but it is not evidence of a durable shift in conviction, and a reader who treated last week's +76.2 as one was misled in the other direction.
IBIT has stopped holding the category up
The genuinely structural change this week sits underneath the flow components rather than in them.
On Monday, this site's ETF flow piece reported that the category's trailing 20 sessions rested on a single fund: BlackRock's IBIT at +$578.3 million, with every other US spot product combined at -$1,386.8 million. Four sessions later IBIT's own 20-session total is -$206.9 million. Over the last five sessions it has shed -$1,040.2 million, which is 68.8% of the category's -$1,513.0 million for the week, and it is on a five-session negative run — its longest stretch of outflows in the current window. The fund with 58.3% of all gross ETF flow activity since January 2024 has gone from the only thing holding the number up to the largest single source of the decline.
This is why flow breadth fell so hard. Across the five sessions from October 30 to November 5, three of twelve products were net positive — BITB, BTCO and MSBT — seven were net negative, and two did not trade at all. On the November 5 session alone, zero of twelve products took in money, six lost money and six reported nothing.
Over the 20-session window, only two of twelve products are net positive: MSBT at +$20.7 million and BITB at +$5.2 million.
Supply absorption is still the single largest drag
Absorption remains the heaviest weight on the score in level terms, at -18.9 of the -32.0, about 53% of all negative contribution.
Over the 20 sessions through November 5, US spot bitcoin ETFs returned a net 13,397.2 BTC to the market while miners created 9,000 BTC at the fixed 450 BTC per day issuance rate. That is an absorption ratio of -1.49x: the channel handed back roughly 29.8 days of global issuance in twenty trading days. Strip absorption out of the composite entirely and the remaining seven components renormalise to -16.4 — still a distribution reading, which was not true last week, when removing absorption left the score positive at +8.8.
That is the key difference from last Friday: absorption was then the entire deficit, with everything else mildly constructive. It is now the largest piece of a deficit that exists without it.
Tuesday's supply absorption column read the same window one session earlier at -1.62x on 14,596 BTC returned. Today's -1.49x is marginally less bad, because November 4 contributed a positive session.
The roll-off now works against this score
Last Friday this column flagged a mechanism that would lift the composite without anyone buying anything: the weak older half of the 20-session window was about to roll out of it. That mechanism has inverted, and it is the most useful forward-looking thing in this run.
Split the current window in half and the halves have swapped places since last week:
- The older ten sessions, October 9 to October 22, net +$231.8 million.
- The recent ten sessions, October 23 to November 5, net -$1,210.1 million.
The older, positive half is what leaves. If flows from here were exactly zero, the 20-session input would not improve — it would converge from -$978.3 million toward -$1,210.1 million over the next ten sessions as October 19's +$526.7 million and October 21's +$524.8 million drop out of the window. Over a five-session horizon the effect is smaller and mildly favourable, with -$124.9 million rolling off and -$853.4 million carrying, but past that the arithmetic turns against the score.
Supply absorption reads the same 20-session window, so the mechanism applies to the model's second-heaviest weight too. Standing still does not fix this reading the way it would have fixed last week's.
Why Strategy's buying barely registers, and the gate that costs 5.8 points
The two treasury components are the composite's only non-flow positives, and one of them is switched off.
Treasury accumulation scores +5.2, contributing +0.7 on a 0.14 weight. Four Strategy filings fall in the window and across them the company acquired 1,328 BTC; the October 26 filing reported no change, and there were no disposals. Holdings stand at 849,327 BTC. The component scores the buying against the size of the existing position, and 1,328 BTC is 0.16% growth, which is why real purchases produce a near-zero reading. Last week's treasury column and the MSTR dashboard carry that subject properly.
Treasury capacity is the more interesting zero. Ungated, it would score 96.0 — Strategy has roughly $43.87 billion of identified financing capacity against a $64.23 billion bitcoin position, a ratio of 0.68x. But v1.1 gates that capacity on mNAV, and at 0.854x the equity trades below the value of the bitcoin it holds. Issuing common to buy bitcoin at a discount to NAV reduces the bitcoin backing each existing share, so the model counts undrawn capacity for nothing at or below parity and ramps it to full weight at a 25% premium. The gate costs the composite 5.8 points this week. Capacity that cannot be drawn accretively is not support, and the model declining to credit it is correct rather than merely conservative.
What this bitcoin demand indicator cannot see
Candour about the limits is most of what makes a composite like this worth reading.
- It is a directional proxy for institutional bid strength, not a price forecast and not investment advice: nothing here implies a mechanical link between the score and future price.
- Five of the eight components read the same US spot ETF flow series through different lenses — momentum, breadth, persistence and concentration all derive from it, and absorption converts it to coins. When that series turns, the composite moves more than eight independent inputs would suggest: this week one dataset turning took four components down together.
- It sees only the channels it measures. Direct custody, separately managed accounts, non-US vehicles, options and perpetual positioning, and any allocator buying spot without a fund wrapper are all invisible to it. A reading of -32 means the measured channels are distributing, not that every institution is.
- The 60-session flow window is still +$4,133.7 million and sits at the 51st percentile of its own history — almost exactly median. The composite's 20-session lens is deliberately shorter than that, so it registers a turn the quarter-long view has not yet confirmed.
- Readings are only strictly comparable within a methodology version. The mNAV gate and the windowed breadth and concentration measures arrived with v1.1 on October 16, so the comparable run is -14.9, -3.4 and today's -32.0; the higher September and early-October figures were computed on v1.0.
- Independent flow trackers reported the same -$430.2 million November 5 session and IBIT's five-session run, corroborating the inputs. This column found no verifiable change in institutional access, mandate or regulation this week to account for the move: the selling is in the data without a documented cause attached.
What would change this reading next week
Concrete thresholds, each on its own component, in rough order of how quickly it could move:
- Flow persistence is the cheapest to flip: it reads a one-session negative streak, so a single positive session reverses its sign, and three or four in a row would restore most of the +7.6 it contributed last Friday.
- Flow breadth needs participation, not size: six or more of the products that actually trade turning positive over a five-session window moves it above zero, whatever the dollar total.
- ETF flow momentum needs about +$978 million of net inflow to bring the 20-session window back to zero, and more than that as the positive older sessions roll out.
- Supply absorption needs the channel to stop returning coins. Offsetting the window's 13,397.2 BTC deficit would take on the order of $1.0 billion of net buying at the current mark.
- Treasury capacity needs Strategy's mNAV back above 1.0x, which from 0.854x means roughly a 17% re-rating of the equity against bitcoin. That is worth up to 5.8 points, the largest block of score currently switched off.
- Flow concentration is already the composite's best input at +44.3, on a window HHI of 0.339 against an even-split baseline of 0.083. It has little room left to help.
The one to watch is breadth rather than the headline: a single large creation at IBIT would move momentum and persistence without saying anything about how widely the bid is spread, where six funds buying modestly would. Track it on the institutional signal dashboard, with the component readings alongside the flow and treasury data on the overview.
Saturday's weekly recap covers the week's flows in full. This column's job is the composite, and its message this Friday is narrower than the number looks: the measured institutional bid has gone, the model has stopped calling it neutral, and the window mechanics no longer do the recovering for it.